From ???@??? Mon Oct 21 02:20:06 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id CAA28887; Mon, 21 Oct 1996 02:13:32 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id CAA12588; Mon, 21 Oct 1996 02:13:11 -0700
Message-Id: <199610210913.CAA12588@lists1.best.com>
From: dannys@stowlake.com (Danny Swarzman)
Subject: What are the rules?
Date: Mon, 21 Oct 1996 02:16:28 -0800
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
A lot of discussion about the advantages of various pairing systems over
rec.games.go doesn't specify the exact rules. For example the term 'McMahon
style or Swiss-McMahon is frequently used.
>From the Ranka yearbook 1986, page 95, in the discussion about Terry
Benson, it says:
"He also introduced the use of the McMahon system (invented by Americans
Lee McMahon and Bob Ryder of Bell Labs but modified by the Europeans) to
the U.S. Champonship tournaments."
Did any of these people write a description? Do we know how to contact
Benson, MacMahon or Ryder?
I've tried to guess the meaning from context. Here's what I think:
1. At the beginning of a tournament each player is assigned a band or
McMahon score based on the player's rating, usually the rating times some
constant. The constant could be 1.
2. An attempt is made to pair players nearest in McMahon scores.
3. Each time a player wins a game, the player's score goes up by 1.
I used MacTD for a handicap tournament. It seemed to do as above. One
player complained that there is another rule that we need to observe:
4. Handicaps are assigned as the difference in McMahon scores.
Was the player correct in saying 4. is a necessary part of the McMahon
system? MacTD based the handicaps on ratings.
If there were no handicaps than the system would resemble a form of Swiss
tournament used by the United States Chess Foundation, known as
'Accelerated Form: Rank-Point Pairing' which is explained in "The Chess
Competitor's Handbook" by Bozidar M. Kazic. The method is credited to
Philip G. Haley.
What are the systems in practice? Does anyone have the rules of any
tournament written? What are the alternatives to McMahon called?
Does the McMahon system specify the method of determining prizes? Are the
SOS and SODOS tie-breaking methods something used in another format and
ported to McMahon?
I would appreciate any information you have on these subject.
- Danny Swarzman
From ???@??? Mon Oct 21 10:59:56 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id IAA27166; Mon, 21 Oct 1996 08:43:23 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id IAA29479; Mon, 21 Oct 1996 08:41:19 -0700
Message-Id: <199610211541.IAA29479@lists1.best.com>
From: jasiek@berlin.snafu.de (Robert Jasiek)
Subject: Re: What are the rules?
Date: Mon, 21 Oct 1996 17:41:08 +0200 (MET DST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>- Danny Swarzman
> [asks about McMahon, Swiss, etc.]
>From my experience of having visited many European tournaments
and seen different programs I can give some useful information
about McMahon (MM).
McMahon features:
- tournament with an announced number of R rounds (generally fixed)
- the average player is expected to participate in more than R/2
- according to announcement or check of present players before first
round a minimal valid rank (e. g. 25kyu) is set equal to a MM-value of
zero, Z
- according to announcement or check of present players before first
round a maximal valid rank, called MM-bar, B, is set for distinguishing
all players who should have a realistic chance to win the tournament;
e. g. 3dan; the MM-bar should not exclude too many players from the
top and not include too many players what could lead to awful
opponents' scores
- all players of rank that equals Z or is lower get Z as their value
- each rank greater than Z gets a MM-value that increases by 1 each
rank until B
- all ranks over B become B
- each player gets his proper starting MM-value before the first round
- the Swiss system is a special case of MM with B=Z=0 and useful for
tournaments with full handicaps or for some championships that
cannot be played as round-robin (e. g. WAGC or a local blitz
tournament)
- after each round for each player his MM-value is increased by 1 or
0.5 or 0 if he wins, ties / misses, loses
- before the first round pairing parameters are fixed: in decreasing
order of importance MM-value, second value, third value
- useful second and third values for not too small tournaments are
SOS, SODOS
- SOS = sum of opponents' scores, where score = current MM-value
- SODOS = sum of defeated opponents' scores
- further criteria may be: B/W choice, not same opponent again, not
same city
- each critirion may have a corresponding probability value
- not all criteria for all players can be fit
- an algorithm (in a program) chooses a proper pairing for all players
- in tournaments up to 30 persons hand draw might be enough
MM result:
- before the tournament criteria for determining a winner are set
- useful are in decreasing order: points, second value, third value
- points = win and tie points
- second value, third value: as used in pairing algorithm, e.g.
SOS, SODOS
- for a single player the element of luck of draw is rather high as
to values of decreasing importance; it may be wise to ignore a
third value and have more than one winner or second place getter
- it may be a good idea to spend prices on players under B with a
sufficiently high points value
Handicaps:
- normally none
- maybe only below some rank used
- often reduced handicap as to MM-value difference, e. g. difference
minus 2
Algorithms:
- a pairing needs to be fair, i.e. especially to avoid one player to be
paired too often to players with greatly differing MM-values
- not all players can be paired optimally in all rounds
- Christoph Gerlach (Germany) has proved in his diploma (which
can be obtained from him as a manuscript) that a global optimum
for all players and all rounds exists and can be detected by a
proper algorithm
Program:
- I strongly recommand to use "MacMahon" by Christoph Gerlach.
It implements an algorithm for a global optimum of a pairing, has
many criteria options and a documentation. It runs with windows.
A 386+ is recommanded, at least 8MB are very useful for quick
pairings (a few minutes for a few hundred people). I do not know
whether it is free or currently has a tax of about 0.5DM per player.
I do not know whether it is available in the net . His address:
Christoph Gerlach, Lange Laube 24, D - 30159 Hannover
(Germany). The program has been used for many German and
several European tournaments including EGC.
SuperMM:
- in a championship that needs to determine the true champion
a MM tournament might have a super MM group if too many
strong players participate and cannot be distinguished in too few
rounds
- super MM is B+1 and donated too a seeded part of all B players
- seeding is always questionable, but sometimes more than 10
rounds or so is not enough with many strong players
-- Robert Jasiek jasiek@berlin.snafu.de
From ???@??? Mon Oct 21 11:00:01 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id JAA11733; Mon, 21 Oct 1996 09:05:34 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id JAA01570; Mon, 21 Oct 1996 09:04:05 -0700
Message-Id: <199610211604.JAA01570@lists1.best.com>
From: evan behre
Subject: Re: What are the rules?
Date: Mon, 21 Oct 1996 12:03:10 -0400 (EDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
On Mon, 21 Oct 1996, Danny Swarzman wrote:
> A lot of discussion about the advantages of various pairing systems over
> rec.games.go doesn't specify the exact rules. For example the term 'McMahon
> style or Swiss-McMahon is frequently used.
>
> From the Ranka yearbook 1986, page 95, in the discussion about Terry
> Benson, it says:
>
> "He also introduced the use of the McMahon system (invented by Americans
> Lee McMahon and Bob Ryder of Bell Labs but modified by the Europeans) to
> the U.S. Champonship tournaments."
>
> Did any of these people write a description? Do we know how to contact
> Benson, MacMahon or Ryder?
dont know about a written description off hand, but we can research this.
i know how to contact terry benson: terrybenson@delphi.com. i have
copied him on this message. terry benson heads the aga rules committe,
btw. i dont know about lee macmahon, but i believe bob ryder has passed
away.
>
> I've tried to guess the meaning from context. Here's what I think:
>
> 1. At the beginning of a tournament each player is assigned a band or
> McMahon score based on the player's rating, usually the rating times some
> constant. The constant could be 1.
>
> 2. An attempt is made to pair players nearest in McMahon scores.
>
> 3. Each time a player wins a game, the player's score goes up by 1.
>
> I used MacTD for a handicap tournament. It seemed to do as above. One
> player complained that there is another rule that we need to observe:
>
> 4. Handicaps are assigned as the difference in McMahon scores.
>
> Was the player correct in saying 4. is a necessary part of the McMahon
> system? MacTD based the handicaps on ratings.
>
i believe to handicap based on macmahon scores is the original way. if
mactd is handicapping based on ratings, this is probably jared roach's
interpretation and consitant with his philosophy that ratings should not
change during the course of a tournament. note: this is different from the
accelrat or original swiss-macmahon approach.
> If there were no handicaps than the system would resemble a form of Swiss
> tournament used by the United States Chess Foundation, known as
> 'Accelerated Form: Rank-Point Pairing' which is explained in "The Chess
> Competitor's Handbook" by Bozidar M. Kazic. The method is credited to
> Philip G. Haley.
>
> What are the systems in practice? Does anyone have the rules of any
> tournament written? What are the alternatives to McMahon called?
>
tournament directors have various ways of assigning pairings. besides
macmahon (swiss-macmahon), accelrat, and mactd, there are the old
classical ways of swiss, round robin, single elimination, double
elimination, and various methods done by hand by the td.
> Does the McMahon system specify the method of determining prizes? Are the
> SOS and SODOS tie-breaking methods something used in another format and
> ported to McMahon?
>
good questions.
> I would appreciate any information you have on these subject.
>
> - Danny Swarzman
>
>
>
evan behre (evanb@clark.net) of dayton, maryland, usa
i stand and salute the migrating geese.
From ???@??? Mon Oct 21 14:22:03 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id OAA08313; Mon, 21 Oct 1996 14:16:04 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id OAA04278; Mon, 21 Oct 1996 14:01:47 -0700
Message-Id: <199610212101.OAA04278@lists1.best.com>
From: Jared Roach
Subject: MacTD is not an algorithm! And other comments...
Date: Mon, 21 Oct 1996 14:01:38 -0700 (PDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
(btw, apostrophes all gone :< during pasting to mainframe)
I just wanted to clarify a couple of points concerning MacTD.
There seems to be a misconception that MacTD is an algorithm. MacTD is a
program. The misconception is understandable since all the other TD
programs implement only one algorithm (to my knowledge) so the program and
algorithm can be referred to by the same name. Not so with MacTD. MacTD
uses the algorithm of the human TDs choice. So when referring to MacTDs
algorithms, it would be best if we referred to, for example, "MacTDs
implementation of the Swiss algorithm", rather than the "MacTD algorithm."
Ill post MacTDs internal documentation on the algorithms at the
end of this email message.
But first, a few more comments.
MacTD does not have an implementation of Accelrat. I just havent
had time to program it. Its one of the more complex and computationally
intensive algorithms out there. Its not that I have anything in
particular against it. For example, I feel that it is superior to certain
types of McMahon tournaments, which I have implemented in MacTD.
MacTDs handicapping algorithms are completely divorced from its
pairing algorithms. The three automated handicapping algorithms are all
based on a players recorded rank at the time the algorithm is executed.
This would almost always be their entering rank. ****!!!!!But it is also
easy to assign whatever handicap the TD desires to any particular
game!!!!!***. Because of the immense variety of handicapping methods, I
have not programmed them all into automated buttons or menu items, but I
probably will for the more popular ones at some time. Its pretty easy for
me to do. Handicapping based on current McMahon ranks makes a lot of
sense. Not only that, but I should point out that I do not refer to
specific philosophies as being possessed by me. Rather, I consider myself
an objective observer capable of listing the advantages and disadvantages
of any philosophy, most of which were almost certainly originally proposed
by someone else. So handicapping a McMahon tournament based on starting
rank is not "consistent with Jared Roachs philosophy," at least not
according to Jared Roach, because Jared Roach does not have a specific
philosophy on that issue. I would tend to argue that choosing a McMahon
pairing system for a tournament is most appropriate for a situation where
ranks are changing during the tournament (or the measured value of a rank
is changing). In this case, since the rank is changing, it would seem to
be more consistent to handicap based on the current rank (i.e. band) than
on the starting rank. But I think it would also be OK to run a McMahon
tournament and handicap based on the starting rank. I still dont think
it makes any sense to underhandicap. Ive seen reasonable arguments in
favor of no handicaps and reasonable arguments in favor of traditional
handicaps, but never in favor of underhandicaps.
MacTD can implement just about any McMahon system imaginable. It
can, for example, implement the system that Robert Jasiek described in his
email. Danny mentioned a tournament in which he used MacTD to run a
McMahon tournament. I believe the version he used was v0.50; the current
version is v0.99j (or v1.0 for optimists); v0.50 was more limited in
flexibility than the current version. Visit my web site to get the latest
version/documentation/information on MacTD.
MacTD is also very good at implementing ad hoc systems used by
tournament directors. This was what MacTD was originally designed to do.
It has several very useful routines for helping TDs do this. For example,
it has an algorithm that checks the pairings for a given round to make
sure no one has the same opponent twice or receives a second bye. This is
very useful for people who do pairings by hand.
OK, here are MacTDs algorithm descriptions (it starts with a few
definitions). This is pasted from the program:
DEFINITIONS
Score: The number of wins a player has. A bye may be counted as
zero, one-half, or one win, depending on the tournament director's whim.
A drop is always zero.
Band: A function based on a player's rank and score. It is of the
form function(rank)+score. The tournament director defines the function.
The most common function is a linear scale with each rank separated by one
unit. If the slope of this function is negative at any point, some of the
algorithms may behave oddly.
Available: A player may not be available for pairing with a
particular player for several reasons: the two may already have played in
a previous round, the tournament director has forbidden the two to play
(perhaps because they are from the same club) for the round in question,
the player is already paired, the player has been granted a bye, or the
player has dropped from the tournament.
Note that the algorithms only pair players who are not already paired
(because of the execution of a previous algorithm or because the TD has
dictated one or more pairs), and who have not been given a "drop" or
"bye." An exception to this is the Desperation algorithm.
EQUALITY
Round 1
a) find the highest ranked unpaired player
b) find the rank of the next highest ranked available player (if
none, give bye)
d) repeat steps a-b until all players are paired
Subsequent Rounds
a) if an odd number of unpaired players, give a bye to the most
deserving player
b) pair all players who already have had byes in previous rounds
--Match to closest score (choose higher if no equal
available)
--Then to rank (choose lower if no equal available)
--Then to SOS (choose higher if no equal available)
--Then to SODOS (choose higher if no equal available)
c) make a list of all unpaired players with the highest score
d) apply the Round 1 Algorithm to that group except that instead
of giving a bye, pair with the next available highest scoring player
(highest ranking if tied, then SOS, then SODOS, then random if still
tied), if still no available players, give bye
e) repeat c-d until list is empty
f) repeat steps c-e until all players are paired
Note: Round 1 is a special case of the Subsequent Round Algorithm
where all players have the same score (zero), except for steps a-b, which
keep a player from getting two byes (obviously unnecessary in Round 1).
SWISS
Round 1
a) find the highest ranked unpaired player
b) make a list (sorted by rank) of all available players not
forbidden to play that player (if none, give bye)
c) pair with the player halfway down that list
d) repeat steps a-c until all players are paired
Subsequent Rounds
a) special case of Swiss-McMahon, with everyone in the same
starting band (i.e. a constant band function)
EQUALITY-MCMAHON
a) if an odd number of unpaired players, give a bye to the most
deserving player (unless round1)
b) find the unpaired player with the highest score in the
highest band
c) pair with the player with the next highest score in the same
band
d) if none in same band, pair with player with closest score in
next band down
e) if no available players at all, give bye
f) repeat b-e until all players are paired
SWISS-MCMAHON
a) if an odd number of unpaired players, give a bye to the most
deserving player (unless round1)
b) find the unpaired player with the highest rank in the highest
band
c) make a list (sorted by rank) of all available players not
forbidden to play that player (if none, give bye) who are in the same band
d) pair with the player halfway down that list
e) if no players of the same band, pair with the highest ranked
player of the next band down
f) if no availbale players at all, give bye
g) repeat b-f until all players are paired
SEQUENTIAL-MCMAHON
a) if an odd number of unpaired players, give a bye to the most
deserving player (unless round1)
b) find the unpaired player with the highest rank in the highest
band
c) pair with the player with the next highest rank in the same
band
d) if none in same band, pair with player with the highest rank
in next band down
e) if no available players at all, give bye
f) repeat b-e until all players are paired
ROUND-ROBIN
This will make round-robin tournaments when implemented during the
pairings for Round 1. It works for up to eight players in a section. If
there are more than that, it will make a round-robin for the first eight
(or 2 or 4 or 6 depending on what you tell it to do), then an independent
round-robin for the next batch of players in the section and so on.
*eight player limit is temporary and will be removed in v1.0
final*
DESPERATION
This generally should not be used unless another pairing algorithm
has already been implemented and has failed to pair all the players (you
will know this if more than one player has a bye). Players must have an
entry of "bye" for the round in question or Desperation will ignore them.
So if you want them ignored, make sure they have "drop" or some other
entry in their pairings for the active round. Desperation works by
searching all possible pairings for the players with byes. A pairing is
not possible if it requires two players to play each other for a second
time, or if it assigns a bye to a player who has more than the minimum
number of byes (usually zero). If Desperation does not succeed it will
unpair a pair of already paired players. It finds the player with the
lowest score and unpairs them and their opponent (regardless of the
opponent's score). It then looks through all possible pairings. If it
fails, it repeats the process until it succeeds or it has examined every
possible pairing. It will find the "best" set of pairings if there is
more than one possible set among a given group of players with byes. Even
so, there may just be one possible set of pairings among, say, four
players with byes late in the tournament. This can cause "strange"
pairings, say between a 5 dan who has lost all their games and a 27 kyu
who has also lost all their games. If the tournament director does not
like this kind of thing, they should intervene in an appropriate manner,
perhaps by forcing certain pairings. Or, don't put players in the same
section that you don't want playing together. If you have a convincing
alternative algorithm that avoids this kind of problem, tell me what it is
- in detail! There are none that leap out and bite me in the nose. It
should not be surprising that in my opinion the Desperation algorithm is
the best "arbitrary" algorithm. It is also my opinion that personal
intervention by the TD based on personalities and circumstances is often a
good idea.
One more thing. If you don't like the pairings Desperation gives
you, you can enable it to forbid certain pairs. This is done with a radio
button on the Options card. Actually, the default for this radio button
is to forbid the pairs you have dictated. So if you don't like the
pairings Desperation gives you and you want to cheat the arbitrary hand of
justice, revert to your saved file just before you ran Desperation (you
did save, didn't you?), then (not before, or you will revert to the
previous settings) forbid whatever pairs you don't like using the "Forbid
Pairs" command in the Micromanagement card. Then run Desperation again.
You can keep doing this until you get what you like or the pairings become
mathematically impossible - whichever comes first.
---------------------------------------------------
"Better to play Go than to do nothing..."
Jared Roach, AGA 2D
Seattle, WA
roach@u.washington.edu
http://weber.u.washington.edu/~roach/
From ???@??? Mon Oct 21 15:55:29 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id PAA28258; Mon, 21 Oct 1996 15:49:37 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id PAA15842; Mon, 21 Oct 1996 15:43:27 -0700
Message-Id: <199610212243.PAA15842@lists1.best.com>
From: dannys@stowlake.com (Danny Swarzman)
Subject: Re: MacTD is not an algorithm! And other comments...
Date: Mon, 21 Oct 1996 15:44:00 -0800
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
Jared Roach wrote:
> I just wanted to clarify a couple of points concerning MacTD.
>There seems to be a misconception that MacTD is an algorithm. MacTD is a
>program. The misconception is understandable since all the other TD
>programs implement only one algorithm (to my knowledge) so the program and
>algorithm can be referred to by the same name. Not so with MacTD. MacTD
>uses the algorithm of the human TDs choice. So when referring to MacTDs
...
>email. Danny mentioned a tournament in which he used MacTD to run a
>McMahon tournament. I believe the version he used was v0.50; the current
>version is v0.99j (or v1.0 for optimists); v0.50 was more limited in
>flexibility than the current version. Visit my web site to get the latest
>version/documentation/information on MacTD.
I want to respond to just some of the points raised in Jared's posting here.
Jared is correct to say that I used v0.50. I had some problems with it. I
haven't had a chance yet to look at his latest version but I have looked at
his list of changes. It looks like most or all of the problems I had have
been fixed. I would like to offer some opinions for those considering using
MacTD.
MacTD is, so far as I know, the only TD program that runs on a Mac. The
majority of computer users in our club are Mac users. If I am lucky enough
to get someone else in our club to direct some of our tournaments, any
software used MUST run on a Mac. It is also, so far as I know, the only TD
program for which source code is distributed.
As Jared says, MacTD can be used with any of several pairing systems.
MacTD is a HyperCard stack. That is both an advantage and a disadvantage.
It is an advantage because it can run on any Mac. It is a disadvantage
because HyperScipt is arcane, lacks beauty and hasn't Apple has a bad
attitude about improving the product. It also cannot be ported to another
platform. It can be easily modified and that is certainally an advantage.
Another disadvantage of using MacTD is that you need to save the entire
stack for each round of the tournament because the program does not produce
files with intermediate results. I don't know if this has changed in the
most recent version. The problem is related to the way things are done in
HyperCard - HyperCard is designed to keep all information in one file, or
stack.
My final comment is about using software for directing a tournament: you
need to experiment with the program. You must find someone who has already
run tournaments to help. In real tournaments people show up late, disappear
at unexpected times. Sometimes a player will report results incorrectly at
first and then change. Players may play with a handicap or color different
from what the td posted. Although you use 1/2 points in the komi or other
devices to avoid ties, there are situations for which there is no better
way to arbitrate. When you do the dry run, you need to experiment with all
these types of events. You need to learn to make adjustments quickly and
easily.
I want to encourage Jared to continue his efforts with MacTD and his
contributions to these discussions.
Danny
From ???@??? Mon Oct 21 17:28:25 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id RAA24803; Mon, 21 Oct 1996 17:22:44 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id RAA27705; Mon, 21 Oct 1996 17:07:36 -0700
Message-Id: <199610220007.RAA27705@lists1.best.com>
From: jasiek@berlin.snafu.de (Robert Jasiek)
Subject: Re: MacTD is not an algorithm! And other comments...
Date: Tue, 22 Oct 1996 02:07:30 +0200 (MET DST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>[...]DESPERATION
> This generally should not be used unless another pairing algorithm
>has already been implemented and has failed to pair all the players[...]
> This can cause "strange"
>pairings, say between a 5 dan who has lost all their games and a 27 kyu
>who has also lost all their games. If the tournament director does not
>like this kind of thing, they should intervene in an appropriate manner,
>perhaps by forcing certain pairings. Or, don't put players in the same
>section that you don't want playing together. If you have a convincing
>alternative algorithm that avoids this kind of problem, tell me what it is
>- in detail!
>[Jared Roach about MacTD]
Desperation does not need to appear at all and strange pairings are
totally superfluous. Christoph Gerlach's algorithm always gives the best
of all theoretically possible pairings from a global view (for all players the
sum of MM-value differences is minimized). Since the algorithm chooses
the global best, it cannot give any bad pair. Even a pair with a difference
of two is rare. And performance is fast.
I had a look on his diploma text, but I have none here and the algorithm
(at least its proof) requires a little graph and matching theory. So you
should get the paper for an implementation from Gerlach.
[begin insertion from another letter by Danny]
My final comment is about using software for directing a tournament: you
need to experiment with the program. You must find someone who has already
run tournaments to help. In real tournaments people show up late, disappear
at unexpected times. Sometimes a player will report results incorrectly at
first and then change. Players may play with a handicap or color different
from what the td posted. Although you use 1/2 points in the komi or other
devices to avoid ties, there are situations for which there is no better
way to arbitrate. When you do the dry run, you need to experiment with all
these types of events. You need to learn to make adjustments quickly and
easily.
Danny
[end insertion]
Such exception handlings and others are treated well by Gerlach's program.
-- Robert Jasiek jasiek@berlin.snafu.de
From ???@??? Mon Oct 21 18:22:18 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id SAA07555; Mon, 21 Oct 1996 18:15:51 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id RAA04494; Mon, 21 Oct 1996 17:57:44 -0700
Message-Id: <199610220057.RAA04494@lists1.best.com>
From: evan behre
Subject: Re: MacTD is not an algorithm! And other comments...
Date: Mon, 21 Oct 1996 20:57:40 -0400 (EDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
On Mon, 21 Oct 1996, Danny Swarzman wrote:
[snip]
>
> I want to encourage Jared to continue his efforts with MacTD and his
> contributions to these discussions.
>
> Danny
>
i enthusiastically agree. jared has obviously thought hard, and worked
hard to contribute to the business of go tournament direction. although he
and i may at times disagree, or misunderstand one another, we do so on
civil terms, with respect and admiration.
evan behre (evanb@clark.net) of dayton, maryland, usa
i stand and salute the migrating geese.
From ???@??? Mon Oct 21 22:45:57 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id TAA07376; Mon, 21 Oct 1996 19:47:49 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id TAA21068; Mon, 21 Oct 1996 19:46:33 -0700
Message-Id: <199610220246.TAA21068@lists1.best.com>
From: Jared Roach
Subject: Saving MacTD
Date: Mon, 21 Oct 1996 19:46:30 -0700 (PDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
On Mon, 21 Oct 1996, Danny Swarzman wrote:
> Another disadvantage of using MacTD is that you need to save the entire
> stack for each round of the tournament because the program does not produce
> files with intermediate results. I don't know if this has changed in the
> most recent version. The problem is related to the way things are done in
> HyperCard - HyperCard is designed to keep all information in one file, or
> stack.
Saving and loading tournament data is simple in v0.99. It goes
into a pretty small text file which can even be edited with a word
processor in the unlikely event that would become necessary. Individual
components such as the registration list or the data from a single section
can also be imported and exported as comma delimited tables.
>
> My final comment is about using software for directing a tournament: you
> need to experiment with the program.
Please! Read the documentation and run a mock tournament
practicing all the "bad" things that Danny mentioned. If you need it
(your dime) I'll even give phone support.
------------------------------------------------------------------
Jared C. Roach
Department of Molecular Biotechnology
Health Sciences Building, Room K354
University of Washington
Box 357730
Seattle, WA 98195
phone (206) 616-4536
FAX (206) 685-7301
roach@u.washington.edu
http://weber.u.washington.edu/~roach/
From ???@??? Mon Oct 21 22:46:06 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id UAA11116; Mon, 21 Oct 1996 20:40:50 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id UAA28427; Mon, 21 Oct 1996 20:29:32 -0700
Message-Id: <199610220329.UAA28427@lists1.best.com>
From: Jared Roach
Subject: Re: MacTD is not an algorithm! And other comments...
Date: Mon, 21 Oct 1996 20:29:27 -0700 (PDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
On Tue, 22 Oct 1996, Robert Jasiek wrote:
> Desperation does not need to appear at all and strange pairings are
> totally superfluous. Christoph Gerlach's algorithm always gives the best
> of all theoretically possible pairings from a global view (for all players the
> sum of MM-value differences is minimized). Since the algorithm chooses
> the global best, it cannot give any bad pair. Even a pair with a difference
> of two is rare. And performance is fast.
There are an awful lot of extreme adjectives here, such as: best,
totally, always, and "at all." I certainly would benefit greatly from
seeing Christoph's work. I would guess that the previous paragraph is
correct given that one assumes that "best" implies that "sum of MM-value
differences is minimized." None of MacTD's algorithms attempts this, so
it is not surprising that MacTD uses Desperation. Desperation is
"necessary" or at least "very desirable" given what MacTD optimizes. And
I also can guarantee that MacTD's Desperation algorithm gives the *best*
possible pairings in *all* cases given that MacTD is giving overiding
consideration to the pairings for those who are arguably the most
important players: those with the highest ranks and the best records or
those who have been previously slighted by being given a bye.
It would not surprise me if the "optimum" minimal McMahon score
sum algorithm could produce a last round pairing that did not pair the two
sole undefeated players in an open section that included both 5 and 6
dans. Depending on the implementation, it might not even pair them if
both were 6-dans. I'd have to look at the algorithm. Other anomalies
are almost certainly possible. Given that different people have different
value systems, it is unlikely (heck, I'll go out on a limb and say
impossible) that any pairing algorithm can be proven "best." It's a
subjective argument. For example, why not use least-squares instead of
sum? Who can say which is better?
That said, I certainly have thought about searching all possible
pairings and assigning a value to them, then mininizing the value
function. I am tickled pink that someone else has already done this,
particularly if they've written theory on it. I had not even considered
using minimizing band score differences as the sole criterion, but I had
thought about using it as a weighted contribution to the total value of
the optimization function. The optimization function might include, for
example, a penalty for pairing players from the same club (or brothers or
whatever). The various weights on the different parameters could be set
by the TD according to their personal philosophy. For example, if the TD
felt that players from the same club should have no preferential
treatment, they could set that parameter weight to zero.
The problem I have with spending a lot of time programming such an
algorithm is that I'm not sure anyone would use it. I think there is an
attractiveness to simple algorithms that are easy to understand by the
participants in the tournament (and the TD, too).
> I had a look on his diploma text, but I have none here and the algorithm
> (at least its proof) requires a little graph and matching theory. So you
> should get the paper for an implementation from Gerlach.
If anyone has his email address I'd be grateful. Also, can anyone
verify whether or not his program is available without cost?
Thanks,
Jared
From ???@??? Tue Oct 22 11:27:23 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id FAA24050; Tue, 22 Oct 1996 05:39:02 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id FAA17012; Tue, 22 Oct 1996 05:38:38 -0700
Message-Id: <199610221238.FAA17012@lists1.best.com>
From: Bill Hyman
Subject: Tournament and Rating Information
Date: Tue, 22 Oct 1996 08:34:35 -0400 (EDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
Well, I'm not a tournament director (yet), but am fascinated with rating
systems and tournament algorithms in other games (i.e. bridge). Where
can I go to find written information regarding tournament seeding
methods and rating systems? I am also interested in getting my hands on
some software to see it in action.
Private e-mail requested.
--
Bill Hyman
bhyman@soundview.com
From ???@??? Tue Oct 22 11:27:26 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id GAA03570; Tue, 22 Oct 1996 06:08:50 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id GAA20536; Tue, 22 Oct 1996 06:08:03 -0700
Message-Id: <199610221308.GAA20536@lists1.best.com>
From: evan behre
Subject: Re: What are the rules? (fwd)
Date: Tue, 22 Oct 1996 09:07:59 -0400 (EDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
attached is a message from terry benson re: mcmahon, etc.
evan behre (evanb@clark.net) of dayton, maryland, usa
i stand and salute the migrating geese.
---------- Forwarded message ----------
Date: Tue, 22 Oct 1996 01:09:29 -0500 (EST)
From: TERRYBENSON@delphi.com
To: evanb@clark.net
Subject: Re: What are the rules?
Evan,
You did a good job in answer Danny Swarzman's questions. There is a written
description of the McMahon - from the British Go Association! who picked it
up, modified it, and used it more than we did until I went to the European
Go congress in '76 and brought back their variant.
Ryder and MacMahon's original was a double knockout with a seeded Swiss
initial pairing system. They ran "thin" tournaments and wanted to avoid the
slaughter rounds typical of a normal swiss (top plays bottom, etc.) Thus the
bands. But they would continue the contest until all but one player had two
loses. The British used it as we do. To create many mini contests within a
larger tournament but to allow movement up and down as determined by results.
/vb
The McMahon is designed however as an EVEN game system. A handicap tournament
- properly run - makes all the players equal! there's no need for bands.
Whether you figure out the proper handicap from the ranks directly or assign
McMahon bands and then determine the handicap is irrelevant. They produce the
same handicap. Whether you adjust the handicap each round is something else -
it's a big jump and probably not justified by a one game result. And adjusting
the handicap is not part of the McMahon (altho the idea is similar).
he McMahon system does not specify the tie-break method. (Remember they
used it as part of a double knockout.) SOS, SODOS, head to head are the most
common methods.
I believe Lee McMahon died some years ago.
The AGA Tournament Handbook (which Ken Koester was the last to work on and
complete - to the degree that anything like that is ever complete) has a
description of a number of systems including the McMahon.
Terry
From ???@??? Tue Oct 22 11:27:30 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id GAA05458; Tue, 22 Oct 1996 06:12:56 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id GAA21004; Tue, 22 Oct 1996 06:12:20 -0700
Message-Id: <199610221312.GAA21004@lists1.best.com>
From: evan behre
Subject: Re: Tournament and Rating Information
Date: Tue, 22 Oct 1996 09:12:16 -0400 (EDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
On Tue, 22 Oct 1996, Bill Hyman wrote:
> Well, I'm not a tournament director (yet), but am fascinated with rating
> systems and tournament algorithms in other games (i.e. bridge). Where
> can I go to find written information regarding tournament seeding
> methods and rating systems? I am also interested in getting my hands on
> some software to see it in action.
>
> Private e-mail requested.
>
> --
> Bill Hyman
> bhyman@soundview.com
>
>
>
apparently there are several documents available from the aga: aga
tournament guide, and aga tournament handbook. there are also various
pairing software available from the authors: mactd/jared roach,
acellrat/paul matthews, gotour/sam zimmerman. acellrat is also a ratings
calculator. gotour impliments a version of mcmahon pairing. mactd is not
an algorthym.-)
evan behre (evanb@clark.net) of dayton, maryland, usa
i stand and salute the migrating geese.
From ???@??? Tue Oct 22 11:27:33 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id HAA21886; Tue, 22 Oct 1996 07:37:06 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id HAA03089; Tue, 22 Oct 1996 07:35:22 -0700
Message-Id: <199610221435.HAA03089@lists1.best.com>
From: jasiek@berlin.snafu.de (Robert Jasiek)
Subject: Re: MacTD is not an algorithm! And other comments...
Date: Tue, 22 Oct 1996 16:35:15 +0200 (MET DST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
> There are an awful lot of extreme adjectives here, such as: best,
>totally, always, and "at all." I certainly would benefit greatly from
>seeing Christoph's work. I would guess that the previous paragraph is
>correct given that one assumes that "best" implies that "sum of MM-value
>differences is minimized."
Very sorry for my awfully grammatically effective use of this word class:)
Half the number of adjectives would have been enough. Semantically
their use is appropriate due to Gerlach's proof.
> None of MacTD's algorithms attempts this, so
>it is not surprising that MacTD uses Desperation. Desperation is
>"necessary" or at least "very desirable" given what MacTD optimizes. And
>I also can guarantee that MacTD's Desperation algorithm gives the *best*
>possible pairings in *all* cases given that MacTD is giving overiding
>consideration to the pairings for those who are arguably the most
>important players: those with the highest ranks and the best records or
>those who have been previously slighted by being given a bye.
Without a mathematically given minimum a use of some desperation
routine is natural. - According to your description MacTD seems to work
well in many cases. (Before I was a little shocked by the caricatural 5d-27k
example.)
> It would not surprise me if the "optimum" minimal McMahon score
>sum algorithm could produce a last round pairing that did not pair the two
>sole undefeated players in an open section that included both 5 and 6
>dans. Depending on the implementation, it might not even pair them if
>both were 6-dans.
This is possible. But Gerlach's program allows the TD to set desired
pairings or to repair.
> I'd have to look at the algorithm. Other anomalies
>are almost certainly possible. Given that different people have different
>value systems, it is unlikely (heck, I'll go out on a limb and say
>impossible) that any pairing algorithm can be proven "best." It's a
>subjective argument. For example, why not use least-squares instead of
>sum? Who can say which is better?
Sure, a choice of aims is subjective. Only after a choice of presuppositions
an algotithm can become best.
> That said, I certainly have thought about searching all possible
>pairings and assigning a value to them, then mininizing the value
>function.
This is the first measure of the algorithm: evaluating all pairs as to a value
function. Then, of course, brute force is impossible and mathematical
knowledge comes into play. This is not trivial, so there's no need for...
> I am tickled pink that someone else has already done this,
>particularly if they've written theory on it. I had not even considered
>using minimizing band score differences as the sole criterion,
To keep things simple for non-specialists I restricted the description to the
core.
> but I had
>thought about using it as a weighted contribution to the total value of
>the optimization function. The optimization function might include, for
>example, a penalty for pairing players from the same club (or brothers or
>whatever). The various weights on the different parameters could be set
>by the TD according to their personal philosophy. For example, if the TD
>felt that players from the same club should have no preferential
>treatment, they could set that parameter weight to zero.
Yes, this is very useful and part of an extended algorithm. Given a pair of
two players. One analyses MM-value difference, second value, third value,
same opponent, colour choice, same city, etc. Before the first round each of
the values gets an importance value, e.g. MM-value 99, second value 70,
not same player 99, not same city 20, etc. as is deemed proper by the TD .
> The problem I have with spending a lot of time programming such an
>algorithm is that I'm not sure anyone would use it. I think there is an
>attractiveness to simple algorithms that are easy to understand by the
>participants in the tournament (and the TD, too).
In Europe with all programs much criticism arose, because it happened too
often that single players got opponents that were too strong or too weak. (To
fit certain rating systems this was even desired.) Now the situation has
improved. (My personal impression is that the pairings are naturally and fairly
created.) If a player is still not satisfied, then a TD has the global
optimum as
a convincing argument. Not many algorithms are understood by all; an
exception may be round-robin.
> If anyone has his email address I'd be grateful. Also, can anyone
>verify whether or not his program is available without cost?
As far as I know currently Gerlach does not have email. A complete German
manuscript of his diploma can be obtained from him for something like 15DM.
I will ask him by snail mail if he might put an english summary and his
program on the net [if not yet there] and if there is some kind of fee. Hope, it
helps a little.
-- Robert Jasiek jasiek@berlin.snafu.de
From ???@??? Tue Oct 22 12:05:29 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id LAA02592; Tue, 22 Oct 1996 11:59:26 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id LAA06692; Tue, 22 Oct 1996 11:50:54 -0700
Message-Id: <199610221850.LAA06692@lists1.best.com>
From: dannys@stowlake.com (Danny Swarzman)
Subject: Re: MacTD is not an algorithm! And other comments...
Date: Tue, 22 Oct 1996 11:54:59 -0800
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
Robert Jasiek wrote:
...
>I will ask him by snail mail if he might put an english summary and his
>program on the net [if not yet there] and if there is some kind of fee. ...
If Gerlach is distributing his program at no cost and it's not on the net
and he wants it to be on the net, I would be willing to put it up for ftp.
- Danny
From ???@??? Tue Oct 22 19:45:32 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id OAA26139; Tue, 22 Oct 1996 14:13:56 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id OAA23755; Tue, 22 Oct 1996 14:01:28 -0700
Message-Id: <199610222101.OAA23755@lists1.best.com>
From: jasiek@berlin.snafu.de (Robert Jasiek)
Subject: Re: MacTD is not an algorithm! And other comments...
Date: Tue, 22 Oct 1996 23:01:23 +0200 (MET DST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>Robert Jasiek wrote:
>...
>>I will ask him by snail mail if he might put an english summary and his
>>program on the net [if not yet there] and if there is some kind of fee. ...
>If Gerlach is distributing his program at no cost and it's not on the net
>and he wants it to be on the net, I would be willing to put it up for ftp.
>
>- Danny
>
Thank you, this might turn out to be very helpful! So let's await his reply
and watch moving the snail mail.
-- Robert Jasiek jasiek@berlin.snafu.de
From ???@??? Wed Oct 23 10:18:33 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id KAA01834; Wed, 23 Oct 1996 10:12:46 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id KAA01704; Wed, 23 Oct 1996 10:08:19 -0700
Message-Id: <199610231708.KAA01704@lists1.best.com>
From: Jared Roach
Subject: Tournament Participant Survey Results
Date: Wed, 23 Oct 1996 10:08:12 -0700 (PDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
We ran a participant survey at the 1996 Seattle Korea Times Tournament.
75 people played at least one game in this tournament making it the
largest tournament ever held in Seattle (except US Congress tournaments).
We received 22 completed surveys. The questions were:
1) Do you prefer handicaps or no handicaps?
(when playing a player of a different rank)
2) Do you prefer 4 or 5 rounds?
3) Which pairing method do you prefer?
_McMahon
_Accelrat
_Pair players with equal number of wins
_Don't care (or don't know)
4) Do you prefer traditional or incremental handicapping?
(incremental uses reverse komi for half a rank difference)
5) Do you prefer trophies or cash prizes?
6) What time do you prefer that a tournament start?
The results are tabulated below:
Handicaps No Handicaps Don't Care
5-6 Dan 1 4 2
1-4 Dan 5 0 2
Kyu 4 2 1
Unknown Rank 1 0 0
Totals: 11 6 5
4 rounds 5 rounds Don't Care
5-6 Dan 3 1 3
1-4 Dan 2 2 3
Kyu 4 1 2
Unknown Rank 1 0 0
Totals: 10 4 8
McMahon Accelrat Equal Don't Care
5-6 Dan 1 0 2 4
1-4 Dan 1 0 3 3
Kyu 0 0 0 7
Unknown Rank 0 0 0 1
Totals: 2 0 5 15
Traditional Incremental Don't Care
5-6 Dan 6 0 1
1-4 Dan 1 1 5
Kyu 2 1 4
Unknown Rank 0 0 0
Totals: 9 2 10
Trophies Cash Don't Care
5-6 Dan 0 3 4
1-4 Dan 1 3 3
Kyu 4 0 3
Unknown Rank 0 1 0
Totals: 5 7 10
9am 10am 11am Don't Care
5-6 Dan 3 4 0 0
1-4 Dan 1 5 1 0
Kyu 1 3.5 2.5 0
Unknown Rank 0 1 0 0
Totals: 5 13.5 3.5 0
From ???@??? Wed Oct 23 13:10:52 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id LAA21724; Wed, 23 Oct 1996 11:35:47 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id LAA11531; Wed, 23 Oct 1996 11:27:38 -0700
Message-Id: <199610231827.LAA11531@lists1.best.com>
From: dannys@stowlake.com (Danny Swarzman)
Subject: Re: Tournament Participant Survey Results
Date: Wed, 23 Oct 1996 11:27:52 -0800
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
Jared Roach wrote:
>We ran a participant survey at the 1996 Seattle Korea Times Tournament.
>75 people played at least one game in this tournament making it the
>largest tournament ever held in Seattle (except US Congress tournaments).
>We received 22 completed surveys. The questions were:
22 responses out of 75 players is a good result.
It looks like the only question about which there is a clear consensus is
the time to start. I think that we can conclude that is likely that most go
players don't want to start at 9:00AM. The related question about the
number of rounds gave results that surprised me.
That leads me to wonder about the best way to schedule a tournament. Can
you post the schedule for the Seattle Korea Times Tournament?
The last tournament I directed was 6 rounds in two days, 3 each. We used
Ing clocks with 45 minutes + 5 reserve periods of 20 seconds. Most rounds
finished within two hours. Only one game ran longer.
I would like to learn more about the one-day tournament idea. I haven't
tried it.
- Danny
From ???@??? Wed Oct 23 13:11:00 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id NAA00991; Wed, 23 Oct 1996 13:00:39 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id MAA22027; Wed, 23 Oct 1996 12:52:40 -0700
Message-Id: <199610231952.MAA22027@lists1.best.com>
From: evan behre
Subject: Re: Tournament Participant Survey Results
Date: Wed, 23 Oct 1996 15:52:27 -0400 (EDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
On Wed, 23 Oct 1996, Danny Swarzman wrote:
> Jared Roach wrote:
> >We ran a participant survey at the 1996 Seattle Korea Times Tournament.
jared, this is a great idea. good work. i am currently collecting
survey information about a poll i sent out to my mostly local list of go
players who like to attend tournaments in the mid-atlantic region. i asked
many of the same questions you did. i will share the questions and the
compilation of answers when a little more time has elapsed to make sure
all those who want to have a chance to respond. so far, i have about 18 or
20 replies out of a total of about 150 on the list.
>
> I would like to learn more about the one-day tournament idea. I haven't
> tried it.
>
> - Danny
>
>
>
in my survey, i asked if players prefered one or two day tournaments. the
preliminary trend seems to be that most players prefer one day
tournaments. it takes less of a time commitment. they have more time to do
chores, be with family, etc. i still think that 2 day tournaments have
their place. more likely as larger scale events, held annualy, and spaced
about a quarter away from the previous 2 day tournament. the 1 day
tournaments tend to be a little less of a big deal, and spaced about on
monthly intervals.
evan behre (evanb@clark.net) of dayton, maryland, usa
i stand and salute the migrating geese.
From ???@??? Wed Oct 23 14:44:14 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id OAA10822; Wed, 23 Oct 1996 14:37:52 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id OAA05310; Wed, 23 Oct 1996 14:34:32 -0700
Message-Id: <199610232134.OAA05310@lists1.best.com>
From: Jared Roach
Subject: One or Two-Day tournaments
Date: Wed, 23 Oct 1996 14:34:17 -0700 (PDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
In Seattle, two-day tournaments have dramatically fewer
participants than one day tournaments. We knew the answer, so we didn't
ask the question on our survey. One thing to note is thast few people
travel long enough distances to Seattle tournaments to warrant staying
overnight. The only tournament in Seattle to have been two days
historically (other than the US Congresses) has been the Cherry Blossom.
For several years we ran a one-day tournament on Saturday and a lightning
tournament on Sunday morning (we have to vacate the Seattle Center by
early afternnon). Particiaption at the lightning tournaments tended to be
very low. Last year we ran a two-day tournament. Particiapation was low,
but now on Saturday as well as Sunday.
If we are wise next year, maybe we will run a one-day tournament
on Saturday and then just man a public relations booth on Sunday.
-------------------------------------------------------------------------
"If you take all four corners, you have either won the game, or lost it."
Jared Roach, AGA 2D
Seattle, WA
roach@u.washington.edu
http://weber.u.washington.edu/~roach/
From ???@??? Wed Oct 23 20:05:23 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id PAA20109; Wed, 23 Oct 1996 15:55:09 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id PAA13664; Wed, 23 Oct 1996 15:34:00 -0700
Message-Id: <199610232234.PAA13664@lists1.best.com>
From: jasiek@berlin.snafu.de (Robert Jasiek)
Subject: Re: Tournament Participant Survey Results
Date: Thu, 24 Oct 1996 00:33:55 +0200 (MET DST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>We ran a participant survey at the 1996 Seattle Korea Times Tournament.
>75 people played at least one game in this tournament making it the
>largest tournament ever held in Seattle (except US Congress tournaments).
>We received 22 completed surveys. The questions were:
In Europe things are a little different. A statistician might have added a
question about time limits. Unfortunately, I have no inqiries to present, only
some guesses.
Time limit wishes would have been between 3minutes and infinite with a
peak at 60minutes.
Different types of tournament players exist: Regular and non-regular
(definition: more than 5 / at most 5 tournament visits per year). Regular
players more likely prefer 5+ rounds and 60+ minutes.
Typical prizes are: Cash for top players (maybe additional trophies), books
or material for others. This is good for "semi-professionalism" and for
further improvement of book-winners.
Different types of tournaments exist: regional, national, international with
player numbers of about 25 / 30-120 / 100-500(mostly 150). Most players
prefer playing (about) equal opponents. Only in regional tournaments
handicaps are necessary. The more important a tournament is the more
rounds it has, the longer its time limit is, the higher the prizes are. Most
German national tournaments are 2 days and typically visited by
students.
A TD should predict the type of a tournament and the number of players.
Then he can choose reasonable features: rounds, time limits, possible
handicap, entry fee, prizes, pairing system, starting time,...
-- Robert Jasiek jasiek@berlin.snafu.de
From ???@??? Wed Oct 23 20:05:27 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id QAA11796; Wed, 23 Oct 1996 16:16:16 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id QAA17225; Wed, 23 Oct 1996 16:01:58 -0700
Message-Id: <199610232301.QAA17225@lists1.best.com>
From: evan behre
Subject: Re: Tournament Participant Survey Results
Date: Wed, 23 Oct 1996 19:01:47 -0400 (EDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
here is the survey i sent to my local/regional players. the results are
still coming in, so i will wait a bit before reporting on them. i
apologize for the length.
evan behre (evanb@clark.net) of dayton, maryland, usa
i forgot to wind my body clock.....
------------[cut here]------------
go tournament poll
this poll will ask questions about what type of go tournament you prefer.
your responses to the questions will be appreciated and help to build a
consensus about optimal tournament design.
1. frequency
if you had the time (under ideal circumstances), how often would you
attend a go tournament? (e.g. once a month, every week, once every quarter
or 3 months, once a year, never, etc.)
how many tournaments did you attend in the last year (12 month period) ?
(please list them. e.g. md open, u.s. go congress, etc.)
do you think that there are not enough, just right, or too many go
tournaments happening in your area?
other thoughts on frequency?
2. duration
do you prefer two day tournaments, or one day tournaments, and why?
if you attend one day tournaments, do you prefer saturdays or sundays?
is there a day that you will not attend tournaments (e.g. for religious
reasons, or other scheduling difficulty)?
for a one day, paired tournament, do you prefer 3 rounds or 4 rounds, or
more (or less)?
other thoughts on duration?
3. handicapping
if playing an opponent 1 or 2 rating points (ranks) different from you
(either up or down) do you prefer to play even games or handicap games?
if you prefer handicap games with opponents of different rating from you,
do you prefer to play handicap games with stronger, or weaker opponents,
or do you prefer to play even games with players of similar rating.
would you object to playing either a slightly stronger or slightly weaker
opponent with some handicap, but slightly less than the number of stones
equal to the difference in rating (e.g. let's say a no komi game with a
player 2 "stones" stronger or weaker)?
please describe any other thoughts or feelings about how you prefer to
play handicap, underhandicapped, or even games.
other thoughts on handicapping?
4. pairing
do you prefer to be paired by the tournament director, or do you prefer
self-paired tournaments?
do you have a favorite pairing system? (e.g. accelrat, macmahon, swiss,
self-paired, other)?
if in a paired tournament, do you feel it is more important to play with
other players nearest to you in rating, or do you prefer to be paired up
to stronger players after winning a round(s)?
when travelling to out-of-town go tournaments with fellow local players do
you prefer not to be paired with them at the tournament? (e.g. "we came
all this way just to play each other like we do every week at go club!")
other thoughts on pairing?
5. prizes
do you prefer to play interesting games with challenging opponents (and
possibly lose), or do you prefer to win most of your games (esp. against
weaker players) and win a prize?
what kind of prizes to you prefer? (e.g. cash, trophies, medals, plaques,
certificates, etc.)
would you play in a tournament that does not offer substantial prizes?
(i.e. that was primarily for ratings).
if you only play in tournaments that offer prizes, how much do you expect
for first place? for second place? etc. ?
other thoughts on prizes?
6. rules
do you prefer ing rules or aga rules?
even if you prefer aga rules, would you play in an ing rule tournament if
it meant the chance to winner a larger prize?
are you aware that in aga rules, that white must play the last stone in a
tournament game, or pass a stone (i.e. a point) to black?
what do you think about this rule?
other thoughts on rules?
7. time control
what time control do you prefer?
are there any short or long limits to time control that you do not like?
(e.g. "i dont like fast overtime periods; faster than 5 stones in 20
min.", or "i dont like slow games.")
do you own your own clock?
would you consider purchasing a clock for your own use and bring it to
tournaments?
were you aware that the aga is try to get out of the clock business? (i.e.
that you shouldn't necessarily expect to be provided a clock by the
tournament director.)
did you know that in chess tournaments players are expected to bring their
own clocks?
other thoughts on time control?
8. location
what state do you live in?
how far are you willing to drive for a go tournament? (in hours or miles)
what kinds of tournaments are you willing to drive farther for?
in what geographical range do you attend go tournaments?
other thoughts on location?
9. entry fees
have you ever not entered a tournament because the entry fee was too high?
what kind of entry fees are you used to paying, and for how many rounds,
and what level of prize funds?
would you consider paying more of an entry fee to play in a nice venue,
with a larger prize fund?
would you prefer to play in tournaments that dont charge very much, and
also dont offer much of a prize?
other thoughts on entry fees?
10. children in go tournaments
do you think children should be encouraged to play in go tournaments?
if so, what can you do to bring more children to go tournaments?
should they play only with other children or be seeded into the adult
field according to their playing strength?
do you think children should be allowed to play for free or at discount
rates?
do you think children should be given cash awards at go tournaments?
if not, what would be a good substitute?
if so, should they be eligible for the same level of prizes the
adults are eligible for (or more, or less)?
11. any other thoughts about other aspects of go tournaments that you
would like to add?
thank you for your time in replying to this poll. it is my hope that you
will be better served as a result of your answers here.
From ???@??? Wed Oct 23 23:43:39 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id WAA22740; Wed, 23 Oct 1996 22:32:20 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id WAA22741; Wed, 23 Oct 1996 22:25:15 -0700
Message-Id: <199610240525.WAA22741@lists1.best.com>
From: bull@netcom.com (michael bull)
Subject: Re: Tournament Participant Survey Results
Date: Wed, 23 Oct 1996 22:25:10 -0700 (PDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Length: 781
Content-Transfer-Encoding: 7bit
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>the Swiss system is a special case of MM with B=Z=0 and useful for
>tournaments with full handicaps or for some championships that
>cannot be played as round-robin (e. g. WAGC or a local blitz
>tournament)
Isn't the MacMahon a special case of the Swiss? (the most commonly
used system?)
When Jared refers to 'pairing players with equal numbers of wins', isn't this
the garden variety swiss?
In my experience, players with good records feel that they have earned the
right to play other contenders,(and get their tie breaking points) and that
playing in the decisive games of the tournament is as important as the prize.
A system that does not bring contenders face to face to decide the outcome
of the tournament is flawed. IMHO.
'Where there is light there is shade'
MIke
From ???@??? Thu Oct 24 11:14:27 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id EAA11322; Thu, 24 Oct 1996 04:25:52 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id EAA18756; Thu, 24 Oct 1996 04:24:35 -0700
Message-Id: <199610241124.EAA18756@lists1.best.com>
From: jasiek@berlin.snafu.de (Robert Jasiek)
Subject: Re: Tournament Participant Survey Results
Date: Thu, 24 Oct 1996 13:24:31 +0200 (MET DST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>>the Swiss system is a special case of MM with B=Z=0
>Isn't the MacMahon a special case of the Swiss? (the most commonly
>used system?)
By deduction from a general B value to B=0 one considers Swiss as a
special case of MM, by induction from Swiss to an extension of a B value
one considers MM as a special case of Swiss. To state the main point:
They are closely related.
In Europe local tournaments tend to have Swiss, tournaments attracting
players of more than one city MM, national championships Swiss or
round-robin.
>In my experience, players with good records feel that they have earned the
>right to play other contenders,(and get their tie breaking points) and that
If not a TD ambitiously sets all games, then this is achieved by
first-second-third pairing values like MM-SOS-SODOS or
points-SODOS-SOSOS-SOSOSOSOS or whatever...
>playing in the decisive games of the tournament is as important as the prize.
Yes, nothing is more important than the fun:)
-- Robert Jasiek jasiek@berlin.snafu.de
From ???@??? Thu Oct 24 11:44:10 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id LAA19962; Thu, 24 Oct 1996 11:38:16 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id LAA22850; Thu, 24 Oct 1996 11:37:10 -0700
Message-Id: <199610241837.LAA22850@lists1.best.com>
From: dannys@stowlake.com (Danny Swarzman)
Subject: Putting mail list contributions on our web site
Date: Thu, 24 Oct 1996 11:40:02 -0800
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
I would like to put selections from this list on my web page:
http://www.stowlake.com/tournament
Would anyone object if I were to do that?
I am also wondering about digests of the list. There is a command to
generate digests. I don't know if the command causes digests to be sent to
members of the list or just to me. I was thinking of generating these once
a month. (The word 'digest' is used in a pecular way here. These 'digests'
contain the entire transcript of the contributions.)
I would like to make these digests available for ftp or directly on the web
page. There may be reformatting to put the entire thing directly on the web
page. In that case it wouldn't be practical to put in on the page. We'll
experiment with that. There should be no problem with ftp.
Reactions?
Danny
From ???@??? Thu Oct 24 11:50:40 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id LAA25667; Thu, 24 Oct 1996 11:44:56 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id LAA24005; Thu, 24 Oct 1996 11:43:15 -0700
Message-Id: <199610241843.LAA24005@lists1.best.com>
From: "Robert J. Barber"
Subject: subscription
Date: Thu, 24 Oct 96 13:41:53 CDT
BestServHost: lists.best.com
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
I've tried a couple of times to subscribe, but without positive
results. bob.
From ???@??? Thu Oct 24 11:59:20 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id LAA26901; Thu, 24 Oct 1996 11:46:36 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id LAA24028; Thu, 24 Oct 1996 11:43:35 -0700
Message-Id: <199610241843.LAA24028@lists1.best.com>
From: Jared Roach
Subject: Re: Tournament Participant Survey Results
Date: Thu, 24 Oct 1996 11:43:20 -0700 (PDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
On Wed, 23 Oct 1996, michael bull wrote:
> When Jared refers to 'pairing players with equal numbers of wins', isn't this
> the garden variety swiss?
Yes, but there are other methods besides Swiss that pair players
with equal number of wins. I wanted to have as few choices as possible on
my survey (to make it simple), so I grouped all these "equal score"
methods together.
In general, Swiss assumes that everyone has a seeded rank. It
dominates the chess world, and perhaps works better there than in Go. As
I recall, it was developed for chess tournaments. Within a score band, in
Swiss tournaments, the top half plays the bottom half. In Go, this can
result in large rank differences between opponents. The main imrovement
of Swiss-McMahon to Go tournaments (IMHO) is to dramatically reduce these
large rank differences.
MacTD has implementations of both the Swiss and Swiss-McMahon
algorithms. It also has an "Equality" algorithm, which is my personal
favorite. This pairs players with equal numbers of wins to each other,
but rather than pairing the top half to the bottom half, it pairs each
player to the closest opponent in rank. This seems to produce good games
and good tournaments.
> A system that does not bring contenders face to face to decide the outcome
> of the tournament is flawed. IMHO.
This is a value I share. Emphasizing this value forces one to
give low ratings to Accelrat and most McMahon systems.
One option is to use MacTD's implementation of "Equality-McMahon."
This pairs players within McMahon bands with other players of the same
score in the same band. This means that players who start out at the same
rank will have an opportunity to face others of the same starting rank in
decisive games. This, I think, would work well in the US Open,
where prizes are distributed for each rank, but there are other compelling
reasons to run a McMahon-like system. Of course, it might not work
well for a section that consisted of, say, 1-4 dans, where a 1 dan who
went 4-0 in the first four rounds might not in round 5 face a 4-dan who
went 4-0.
---------------------------------------------------
"Better to play Go than to do nothing..."
Jared Roach, AGA 2D
Seattle, WA
roach@u.washington.edu
http://weber.u.washington.edu/~roach/
From ???@??? Thu Oct 24 17:43:13 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id NAA03592; Thu, 24 Oct 1996 13:43:19 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id NAA10711; Thu, 24 Oct 1996 13:40:27 -0700
Message-Id: <199610242040.NAA10711@lists1.best.com>
From: jasiek@berlin.snafu.de (Robert Jasiek)
Subject: Re: Putting mail list contributions on our web site
Date: Thu, 24 Oct 1996 22:40:23 +0200 (MET DST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>I would like to put selections from this list on my web page:
>
>http://www.stowlake.com/tournament
>
>Would anyone object if I were to do that?
>
The mailing list serves publicity, so why object to your good work?
-- Robert Jasiek jasiek@berlin.snafu.de
From ???@??? Fri Oct 25 10:08:10 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id IAA05001; Fri, 25 Oct 1996 08:51:10 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id IAA23178; Fri, 25 Oct 1996 08:40:55 -0700
Message-Id: <199610251540.IAA23178@lists1.best.com>
From: evan behre
Subject: welcome new members
Date: Fri, 25 Oct 1996 11:40:45 -0400 (EDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
welcome to new subscribers to this list, including phil straus, aga
president.
danny, is there a way that we can see who are the subscribers? i.e. see a
list of subscribers?
evan behre (evanb@clark.net) of dayton, maryland, usa
i forgot to wind my body clock.....
From ???@??? Fri Oct 25 11:09:11 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id LAA18188; Fri, 25 Oct 1996 11:03:05 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id KAA13892; Fri, 25 Oct 1996 10:56:13 -0700
Message-Id: <199610251756.KAA13892@lists1.best.com>
From: dannys@stowlake.com (Danny Swarzman)
Subject: Re: welcome new members
Date: Fri, 25 Oct 1996 10:58:42 -0800
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
Evan Behre wrote:
>danny, is there a way that we can see who are the subscribers? i.e. see a
>list of subscribers?
>
No, yes. I sent a moderator command getlist and got this:
Command: getlist
BEGIN GETLIST td
+d...... sboddy@freenet.victoria.bc.ca
+d...... cdlin@amex.com
+d...... Cnchapman@aol.com
+d...... dannys@best.com
+d...... steve.plate@boeing.com
+d...... 71022.1406@CompuServe.COM
+d...... 70451.2663@CompuServe.COM
+d...... pmp@ghgcorp.com
+d...... angus@halcyon.com
+d...... bull@netcom.com
+d...... bhyman@soundview.com
+d...... dannys@stowlake.com
+d...... veg@teleport.com
+d...... jasiek@berlin.snafu.de
+d...... U50329@UICVM.UIC.EDU
+d...... roach@u.washington.edu
+d...... sloane@UKANS.EDU
+d...... evanb@clark.net
+d...... bfelice@earthlink.net
+d...... barryp@atlas.otago.ac.nz
+d...... go@seattlego.org
+d...... niklas.mellin@sto.fdata.se
+d...... pstraus@holonet.net
END GETLIST td
I don't know of an easy way to generate the 'real' names, that are sent as
comments in address fields.
I plan to generate a digest in a few days. That will tell you more about
who the contributors are.
- Danny
From ???@??? Fri Oct 25 13:33:46 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id NAA03138; Fri, 25 Oct 1996 13:26:57 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id NAA02467; Fri, 25 Oct 1996 13:19:37 -0700
Message-Id: <199610252019.NAA02467@lists1.best.com>
From: evan behre
Subject: poll results (long)
Date: Fri, 25 Oct 1996 16:19:33 -0400 (EDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
sorry about the length of the attached. if you are not interested, just
dont read it and delete it. if you have the interest and patience, i hope
you find it useful.
evan behre (evanb@clark.net) of dayton, maryland, usa
i forgot to wind my body clock.....
------------[cut here]------------
results of a go tournament poll taken in october, 1996.
pollster: evan behre
who was polled? a group of mostly mid-atlantic go players who attend go
tournaments and whose email address i have. about 150 in all. a few from
outside the mid-atlantic region.
respondents: about 25 (25/150 = ~17%). not all respondents answered all
questions.
the purpose of the poll was to try to get a feel for what the consensus
was for various tournament factors, and to try to plan future tournaments
based on the results.
1. frequency
if you had the time (under ideal circumstances), how often would you
attend a go tournament? (e.g. once a month, every week, once every quarter
or 3 months, once a year, never, etc.)
1/week or 1/2 weeks: 1
1/month: 12
1/2 months: 3
1/quarter: 9
1/year: 1
the consensus seemed to be between 1/month and 1/quarter.
how many tournaments did you attend in the last year (12 month period) ?
(please list them. e.g. md open, u.s. go congress, etc.)
23 replied to this ranging from 0 to 10. average was 4. not surprisingly,
those who attended more than 4 tournaments per year tended to prefer
monthly or semi-monthly tournaments in the above question.
do you think that there are not enough, just right, or too many go
tournaments happening in your area?
about right: 9
not enough: 7 <- 3 of these were from fringe areas, nc, nj/nyc
too many: 1
no opinion: 2
the consensus seems to be that there are about the right number of
tournaments in the region, with perhaps room for 1 or 2 more during the
year. some local areas on the outer edges of the mid-atlantic region are
in more desparate need of more tournaments. specifically, nyc and nc.
other thoughts on frequency?
here are some examples:
like many small local, low cost tournaments.
each club should have quarterly tournaments of their own, then cooperate
with other clubs to put on larger tournaments.
too many tournaments may dilute resources and player interest.
attendance seems better now that there are more tournaments than before,
why not put on more?
2. duration
do you prefer two day tournaments, or one day tournaments, and why?
prefer 1 day tournaments: 19
both: 3
prefer 2 day tournaments: 3
the consensus clearly prefers 1 day tournaments. most folks cited schedule
conflicts (work, family, etc.) as limiting there ability to participate in
2 day touranments. 2 day tournaments are still seen as important, but
perhaps these tournaments should be on a quarterly, or semi-annual basis.
(e.g. princeton, md open, mid-atlantic). if there were no 1 day
tournaments, a whole lot of otherwise potential go tournament players
would be left without an opportunity.
if you attend one day tournaments, do you prefer saturdays or sundays?
no preference: 8
prefer sat: 8
prefer sun: 1
is there a day that you will not attend tournaments (e.g. for religious
reasons, or other scheduling difficulty)?
mostly no.
no: 10
sunday: 2
others include: wife's birthday, wedding anneversay, family vacations,
piano lessons.
for a one day, paired tournament, do you prefer 3 rounds or 4 rounds, or
more (or less)?
consensus is 4 rounds:
more than 4: 4
4 rounds: 14
3 rounds: 4
no preference: 3
other thoughts on duration?
examples include:
5 round, 2-day, or self-paired 1 day.
need more 7 round tournaments.
keep on schedule, end on time.
3. handicapping
if playing an opponent 1 or 2 rating points (ranks) different from you
(either up or down) do you prefer to play even games or handicap games?
most folks prefer handicapping.
handicap: 16
even: 3
no opinion: 1
if you prefer handicap games with opponents of different rating from you,
do you prefer to play handicap games with stronger, or weaker opponents,
or do you prefer to play even games with players of similar rating.
both or doesnt matter: 9
even vs similar: 10
would you object to playing either a slightly stronger or slightly weaker
opponent with some handicap, but slightly less than the number of stones
equal to the difference in rating (e.g. let's say a no komi game with a
player 2 "stones" stronger or weaker)?
dont object: 12
would object: 7
please describe any other thoughts or feelings about how you prefer to
play handicap, underhandicapped, or even games.
examples include:
prefer reverse komi to 2 stones.
take handicap and ratings seriously, handicap according to difference.
do not change rating during tournament.
play for fun, not for prize or glory, so prefer handicap -> 50% outcome.
prefer even games.
dont believe reverse komi is worth the trouble.
full handicapping!
dont like "enter as you please" ratings, prefer enter at official ratings.
prefer even games or up to 2 stones, dont like high handicap in
tournament.
tdlist should be updated monthly, ratings updated monthly.
4. pairing
do you prefer to be paired by the tournament director, or do you prefer
self-paired tournaments?
most prefered to be paired by td.
prefer td pairing: 16
self-paired: 2
no preference: 4
do you have a favorite pairing system? (e.g. accelrat, macmahon, swiss,
self-paired, other)?
not a big issue. slight preference for accelrat.
self-paired: 3
macmahon: 3
accelrat: 7
no pref: 12
if in a paired tournament, do you feel it is more important to play with
other players nearest to you in rating, or do you prefer to be paired up
to stronger players after winning a round(s)?
not a clear consensus. slight preference to similar rating.
either/both: 5
similar: 8
paired up: 6
when travelling to out-of-town go tournaments with fellow local players do
you prefer not to be paired with them at the tournament? (e.g. "we came
all this way just to play each other like we do every week at go club!")
consensus is to play those not usually played.
prefer not to be paired with local players: 16
doesnt matter: 5
other thoughts on pairing?
examples include:
prefer to have rating adjusted during tournament and be paired
accordingly.
dont adjust rating during tournament.
dont like self paired.
like self paired.
winners should be paired up, losers should be paired down.
players should play similar, stronger, and weaker at tournament.
5. prizes
do you prefer to play interesting games with challenging opponents (and
possibly lose), or do you prefer to win most of your games (esp. against
weaker players) and win a prize?
clear consensus prefering challenging games over winning.
prefer challenging games: 18
prefer to win: 1
prefer to win challenging games: 4
no pref: 2
what kind of prizes to you prefer? (e.g. cash, trophies, medals, plaques,
certificates, etc.)
the main theme seemed to be cash vs non-cash prizes.
prefer cash: 8
cash + trophies: 3
non-cash: 10
no pref: 3
besides the above listed examples, i also got go books, fans, stones,
posters, clocks, gift certificates, and a lot of players dont particularly
care about prizes, some even abhor them.
would you play in a tournament that does not offer substantial prizes?
(i.e. that was primarily for ratings).
strong consensus here
yes: 17 (mostly enthusiastically)
maybe: 1
if you only play in tournaments that offer prizes, how much do you expect
for first place? for second place? etc. ?
100-300 for 1st, 50-120 for 2nd, 25-50 for 3rd
prizes should help to pay for tournament and travel expenses (entry fees,
meals, hotel, travel, etc)
other thoughts on prizes?
examples include:
prizes are necessary to attract the strong players, but i dont play for
prizes, so it doesnt really matter.
high prize money for amateur tournaments tend to corrupt the game.
should offer significant prizes to top players only.
chess players regret loss of civility at chess tournaments due to big
prizes.
i like prizes.
cash goes well with anything.
books for kyu players.
6. rules
do you prefer ing rules or aga rules?
aga: 12
ing: 3
neither, prefer japanese rules: 1
no pref: 8
even if you prefer aga rules, would you play in an ing rule tournament if
it meant the chance to winner a larger prize?
yes: 11
no: 2
no pref: 4
are you aware that in aga rules, that white must play the last stone in a
tournament game, or pass a stone (i.e. a point) to black?
yes: 18
no: 3
what do you think about this rule?
maybe correct mathematically, but is a nuisance: 1
like it, or dont mind it: 10
dont like it: 4
confused: 3
no opinion: 2
other thoughts on rules?
dont mind ing rules, dont like ing equipment.
7. time control
what time control do you prefer?
regular times of 30 min to 90 min, mostly in the 45 to 60 range.
overtimes vary. 20/5, 20/10, 25/15 canadian overtime, or 30 sec per move
japanese byo-yomi, or 20 sec per move ing overtime, or sudden death.
consensus seemed to be about 45 min regular time with some reasonable
overtime.
are there any short or long limits to time control that you do not like?
(e.g. "i dont like fast overtime periods; faster than 5 stones in 20
min.", or "i dont like slow games.")
no real consensus.
prefer fast overtime to move game along: 2
do not like fast overtime: 1
prefer fast games, dont like slow games: 3
dont like fast games, prefer slow games: 3
do you own your own clock?
yes: 11
no: 11
would you consider purchasing a clock for your own use and bring it to
tournaments?
yes (would, will, or did): 13
no: 4
were you aware that the aga is try to get out of the clock business? (i.e.
that you shouldn't necessarily expect to be provided a clock by the
tournament director.)
yes: 4 (yet one of these pointed out that ing clocks are a necessary
use of ing equipment grant and will be aga clocks provided to tournaments)
no: 15 (one of these objected, "bad idea")
did you know that in chess tournaments players are expected to bring their
own clocks?
yes: 7 (but several didnt care and resisted comparisons to chess)
no: 12
other thoughts on time control?
prefer standardized time control (same for different tournaments) rather
than each td doing his/her own thing.
suggestion of a speed/blitz tournament.
like clocks to be used for paired tournaments.
8. location
what state do you live in?
md: 9
pa: 5
va: 4
nc: 3
nj: 2
il: 1
how far are you willing to drive for a go tournament? (in hours or miles)
most willing to drive an hour or so. some longer who live on the fringe.
1 hr: 5
1-2 hrs: 3
2 hrs: 5
3-4 hrs: 4
5 hours: 2 (nc -> md)
100 miles: 2
200 miles: 1
what kinds of tournaments are you willing to drive farther for?
some folks are willing to drive farther for 2-day tournaments, congress,
workshops, bigger prizes, to visit evan's barn. some folks are not willing
to drive far for any tournament. no real consensus here.
in what geographical range do you attend go tournaments?
mostly mid-atlantic. one continental us and canada.
md: 14
va: 11
dc: 9
pa: 6
nj: 4
nc: 4
de: 3
ny: 1
other thoughts on location?
example include:
prefer big cities to farm towns.
prefer suburbs to downtown.
like to see more big tournaments in nyc.
business class hotels near major highways for easy access.
like picnic games at evan's barn.
9. entry fees
have you ever not entered a tournament because the entry fee was too high?
no: 18
yes: 3 (one of those was a chess t)
what kind of entry fees are you used to paying, and for how many rounds,
and what level of prize funds?
varied. consensus seemed to be $5 per round. many liked $10, 4-5 rounds,
small prizes.
$25 for 2 days, 5 rnds
$10 for 1 day, 4 rnd
$5 for 1 day, self-paired
$10 to $20 for 4 rounds, less than $25
would you consider paying more of an entry fee to play in a nice venue,
with a larger prize fund?
yes: 3
no: 4
depends: 4
would pay more for nice venue, but not larger prize fund: 7
would you prefer to play in tournaments that dont charge very much, and
also dont offer much of a prize?
yes: 14
no: 3 (small affairs, matches not good)
indiff: 2
other thoughts on entry fees?
examples include:
entry fees not a problem.
maybe offer discounts to students, children, unemployed, etc.
10. children in go tournaments
do you think children should be encouraged to play in go tournaments?
very clear consensus to encourage children.
yes: 21
no: 2
if so, what can you do to bring more children to go tournaments?
not sure, give prizes, teach go in schools.
should they play only with other children or be seeded into the adult
field according to their playing strength?
according to strength: 15
depends: 3
there was some sentiment for a mix of seeding strong children in with
adults, as well as having children only, or scholastic tournaments or
sections within tournaments.
do you think children should be allowed to play for free or at discount
rates?
yes: 17
no: 1
do you think children should be given cash awards at go tournaments?
no: 4
yes: 6
if not, what would be a good substitute?
books, certificates, kid oriented, t-shirts, nintendo, go
related prizes.
if so, should they be eligible for the same level of prizes the
adults are eligible for (or more, or less)?
same level: 8 (if competing with adults)
less: 2 (esp if discounted entry)
11. any other thoughts about other aspects of go tournaments that you
would like to add?
examples include:
tournaments should be free of technicalities, high level of courtesy,
relaxed atmosphere.
would be nice to see an up to date, on-line tournament calendar.
i go to tournaments to measure my performance, prizes not important to me,
but location is.
like the idea of a side tournament of 9x9 games between rounds.
like nice venues, and refreshments.
-----------------------------------------------------------------------------
my take on this: folks are generally satisfied with the current
mid-atlantic tournament scene, especially in the va, dc, md, and pa area.
some complain that there are not enough tournaments at the fringes,e.g.
nyc and nc. the frequency of tournaments seems about right. most folks
like to have the option of attending or not a tournament scheduled once a
month. most prefer 1 day, 4 round tournaments on saturdays. 2 day, big
event tournaments are okay every once in a while, but should not be the
norm (i.e. okay for quarterly, but not montly). entry fees and time
control seem not to be a big deal. while most players prefer to play
opponents most similarly rated, they also expect to take/give a handicap
for stronger/weaker opponents. most do not object to slight
underhandicapping (e.g. even games or adjusted komi with opponents
different by 2 rating points). some object strongly and feel that
handicaps should be according to the difference in rating , and that
ratings should be taken seriously. for the most part, prizes are seen as
necessary to attract the stronger players, but most players who responded
are not motivated to attend tournaments because of prizes. players tend to
go to tournaments to test their ratings and to play challenging games
against other players not normally played. given a choice between
challenging games and winning prizes, most players would choose
challenging games. most players prefer aga rules, but there is not a
strong objection to ing rules either. most folks would like to encourage
children to attend go tournaments, to seed them with adults according to
their ability, to give them discounts on entry fee, and to award prizes at
the same level as adults. to give children cash or not remains
contraversial and there is no clear consensus.
From ???@??? Sat Oct 26 11:31:10 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id IAA29895; Sat, 26 Oct 1996 08:36:30 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id IAA02717; Sat, 26 Oct 1996 08:35:41 -0700
Message-Id: <199610261535.IAA02717@lists1.best.com>
From: bull@netcom.com (michael bull)
Subject: Re: Putting mail list contributions on our web site
Date: Sat, 26 Oct 1996 08:35:39 -0700 (PDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Length: 4793
Content-Transfer-Encoding: 7bit
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
Having access to 'human-readable'tournament records will be helpfull in the
discussion and understanding of the various pairing and rating systems.
A record that is in a concise table form can be published or posted and has
some PR value.
The records produced by the programs Accelerat and MacTD are long and difficult
to read. The AGA's rating database also has records that can be represented in
table form.
> > Is it possible to add a feature to the accelRat program
> > that can produce a postable tournament record similiar
> > to the form used by the Europeans?
>
> The answer, in general, is yes, but I'd need to see an
> example of what you mean.
>
An example:
Rocky Mountain Open, Denver CO, May 4-5th, Open Section
#1 #2 #3 #4 #5 w/l p sos sdos agtc
1.Jung Hoon Lee 7.8 w7+ w5+ b6+ w3+ w2- 4/1 1 13 10 10
2.Michael Zeng 7.4 w4+ w6- b8+ w7- b1+ 3/2 3 10 6 5
3.Robert Zeng 6.9 w8+ w4+ b5+ b1- b7- 3/2 4 9 3 3
4.Eric B Kim 6.1 b2- b3- b7- b6- w9+ 1/4 7 12 0
5.Hyun Min Kang 6.1 w6+ b1- w3- b8- bye 1/3 8 10 2
6.Joseph Kim 5.5 b5- b2- w1- w4+ w8+ 2/3 5 10 2 2
7.Namkey Park 6.5 b1- w8+ w4+ w2+ w3+ 4/1 2 12 8 7
8.Sangwook Lee 5.5 b3- b7- w2- w5+ b6- 1/4 6 13 1 1
9.Burt Simon 3.5 b4- 0/1 9 1 0
Only the open section is listed, but a 40 player tournament can fit on
a single page. The aga id numbers could be included. It may be possible
to show the ratings/points change from each match.
I started with this: (and translated it to the above form manually)
## TOURNEY Rocky Mountain Open, Denver CO, May 4-5th, Open Section
## RULES AGA
## 1
## RULES AGA
USA5806 TMP20 W 0 5
USA1495 USA6441 W 0 5
USA3794 USA7658 W 0 5
USA3971 TMP6 W 0 5
## 2
## RULES AGA
USA5806 USA3971 W 0 5
USA1495 TMP6 B 0 5
USA3794 USA6441 W 0 5
TMP20 USA7658 W 0 5
## 3
## RULES AGA
TMP6 USA5806 B 0 5
USA7658 USA1495 B 0 5
USA3971 USA3794 B 0 5
TMP20 USA6441 W 0 5
## 4
## RULES AGA
USA5806 USA3794 W 0 5
USA1495 TMP20 B 0 5
TMP6 USA6441 W 0 5
USA7658 USA3971 W 0 5
## 5
## RULES AGA
USA5806 USA1495 B 0 5
TMP20 USA3794 W 0 5
TMP6 USA7658 W 0 5
USA6441 USA701 W 0 5
TMP6
NAME="Kim, Joseph"
RATING=5.50000
SIGMA=1.20000
USA6441
NAME="Kim, Eric B."
RATING=6.10000
SIGMA=0.60000
USA3971
NAME="Kang, Hyun Mun"
RATING=6.10000
SIGMA=0.60000
USA5806
NAME="Lee, Jung Hoon"
RATING=7.80000
SIGMA=0.60000
USA1495
NAME="Zeng, Dr. Michael"
RATING=7.40000
SIGMA=0.60000
USA3794
NAME="Zeng, Robert"
RATING=6.90000
SIGMA=0.60000
TMP20
NAME="Park, Namkey"
RATING=6.50000
SIGMA=1.20000
USA7658
NAME="Lee, Sangwook"
RATING=5.50000
SIGMA=1.20000
USA701
NAME="Simon, Burt"
RATING=3.50000
SIGMA=1.20000
## END
#Contact: Samuel E. Zimmerman 717 898 9363 szimmerman@cpcnet.com
----------
and another:
-----------------------------------------------------------------------
--------------------------------------------
Id Name Rank Stat Init Curr SODOS SOS
RD1 RD2 RD3 RD4 RD5
Band Band
-----------------------------------------------------------------------
--------------------------------------------
7346 Jiang, Danning 6D OK 10 14 48 61
+ 6626B + 6099B - 1890B + 4813B + 2998B
6099 Jeong, Jong In 6D OK 10 13 38 65
+ 5352B - 7346W + 2998W + 1890B - 4813W
1890 Kim, Woo Jin 6D OK 10 13 38 64
+ 3293B + 2998W + 7346W - 6099W - 5352B
4813 Yu, Lianzhou 6D OK 10 13 38 64
- 2998W + 3293W + 5352B - 7346W + 6099B
5352 Hung, Joey 6D OK 10 13 35 61
- 6099W + 6626B - 4813W + 3293B + 1890W
2998 Hong, Seung Hyun 6D OK 10 12 23 63
+ 4813B - 1890B - 6099B + 6626D - 7346W
3293 Q, David 6D OK 10 12 20 59
- 1890W - 4813B + 6626B - 5352W +bye D
6626 Lee, Won Sik 6D OK 10 10 0 61
- 7346W - 5352W - 3293W - 2998D -loss D
1st Place Jiang $1000
2nd Place Jeong 500
3rd Place Yu 250
4th Place Kim 125
5th Place Hung 75
6th Place Hong 50
From ???@??? Sat Oct 26 13:46:09 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id NAA28629; Sat, 26 Oct 1996 13:40:15 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id NAA07311; Sat, 26 Oct 1996 13:38:57 -0700
Message-Id: <199610262038.NAA07311@lists1.best.com>
From: Jared Roach
Subject: MacTD's Tournament Record Format
Date: Sat, 26 Oct 1996 13:38:54 -0700 (PDT)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
On Sat, 26 Oct 1996, michael bull wrote:
> The records produced by the programs Accelerat and MacTD are long and
difficult
> to read. The AGA's rating database also has records that can be represented in
> table form.
Choose "Save Section" from MacTD's "Section" menu on the Section
Micromanagement card to generate a table such as you describe. This will
be a comma delimited table which can be read by any spreadsheet program.
Or, just replace the commas with tabs using the replace command on a word
processor.
The long and difficult format you refer to is the "AGA Report"
which is used to submit game results to the AGA for ratings purposes.
MacTD also has a pretty good utility for perusing the AGA
database (and saving it as a comma delimited table).
---------------------------------------------------
"Better to play Go than to do nothing..."
Jared Roach, AGA 2D
Seattle, WA
roach@u.washington.edu
http://weber.u.washington.edu/~roach/
From ???@??? Sat Oct 26 14:37:02 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id OAA27481; Sat, 26 Oct 1996 14:30:42 -0700
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id OAA14006; Sat, 26 Oct 1996 14:30:04 -0700
Message-Id: <199610262130.OAA14006@lists1.best.com>
From: Steven Rushing <75444.720@CompuServe.COM>
Subject: Accelerat Reports
Date: 26 Oct 96 17:29:13 EDT
BestServHost: lists.best.com
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
On the Accellerat reports,
it does have a report feature which list all games each player has played.
Format is very readable. This is in addition to its AGA ratings input file.
Here is a sample:
The following data are shown for each game:
difficulty (probability of losing)
opposition (difficulty as a rating)
win or loss (+ or -)
color held (W or B)
opponent's name
opponent's updated rating
handicap
LABACH, KEN 2.0
0.85 3.4 - W vs COX, JIM -4.2 (6 stones)
0.26 1.1 + W vs LAING, MARK -3.0 (2 stones)
0.21 -1.1 + W vs COBB, WILLIAM S. -5.4 (4 stones)
LAING, MARK -3.0
0.74 -2.2 - B vs LABACH, KEN 2.0 (2 stones)
0.72 -2.3 - W vs CELMER, PAUL -4.1 (2 stones)
0.25 -3.9 + W vs GORNALL, DAVID A. -12.3 (8 stones)
0.10 -4.7 + W vs RILEY, RANDOLPH -14.4 (9 stones)
CELMER, PAUL -4.1
0.90 -2.5 - W vs COX, JIM -4.2 (2 stones)
0.76 -3.2 - W vs SHUTT, ROBERT L. -10.9 (8 stones)
0.28 -4.9 + B vs LAING, MARK -3.0 (2 stones)
0.22 -5.1 + W vs COBB, WILLIAM S. -5.4 (even game komi)
From ???@??? Sun Oct 27 09:57:39 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id IAA00505; Sun, 27 Oct 1996 08:51:39 -0800
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id IAA15861; Sun, 27 Oct 1996 08:51:11 -0800
Message-Id: <199610271651.IAA15861@lists1.best.com>
From: bull@netcom.com (michael bull)
Subject: Re: Accelerat Reports
Date: Sun, 27 Oct 1996 08:51:09 -0800 (PST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1198
Content-Transfer-Encoding: 7bit
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>
> On the Accellerat reports,
>
> it does have a report feature which list all games each player has played.
> Format is very readable. This is in addition to its AGA ratings input file.
>
But it does not show the pairings by rounds, and uses too much space.
> Here is a sample:
>
> The following data are shown for each game:
> difficulty (probability of losing)
> opposition (difficulty as a rating)
> win or loss (+ or -)
> color held (W or B)
> opponent's name
> opponent's updated rating
> handicap
>
> LABACH, KEN 2.0
> 0.85 3.4 - W vs COX, JIM -4.2 (6 stones)
> 0.26 1.1 + W vs LAING, MARK -3.0 (2 stones)
> 0.21 -1.1 + W vs COBB, WILLIAM S. -5.4 (4 stones)
>
> LAING, MARK -3.0
> 0.74 -2.2 - B vs LABACH, KEN 2.0 (2 stones)
> 0.72 -2.3 - W vs CELMER, PAUL -4.1 (2 stones)
> 0.25 -3.9 + W vs GORNALL, DAVID A. -12.3 (8 stones)
> 0.10 -4.7 + W vs RILEY, RANDOLPH -14.4 (9 stones)
>
> CELMER, PAUL -4.1
> 0.90 -2.5 - W vs COX, JIM -4.2 (2 stones)
> 0.76 -3.2 - W vs SHUTT, ROBERT L. -10.9 (8 stones)
> 0.28 -4.9 + B vs LAING, MARK -3.0 (2 stones)
> 0.22 -5.1 + W vs COBB, WILLIAM S. -5.4 (even game komi)
From ???@??? Sun Oct 27 21:04:29 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id TAA19061; Sun, 27 Oct 1996 19:56:45 -0800
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id TAA07860; Sun, 27 Oct 1996 19:54:51 -0800
Message-Id: <199610280354.TAA07860@lists1.best.com>
From: Phil Straus
Subject: Re: Accelrat Reports
Date: Mon, 28 Oct 1996 10:54:39 -0500
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>>
>>(Steven Rushing) On the Accellerat reports,
>>
>> it does have a report feature which list all games each player has played.
>> Format is very readable. This is in addition to its AGA ratings input file.
>>
>(Mike Bull) But it does not show the pairings by rounds, and uses too much
>space.
Yes, but the Accelrat reports shows who played tough games and who played
easy games. To me, this seems much more useful than the order in which
games were played. I can live with the extra space required, since the
information is so clear.
Taking the accelrat reports a step further, and printing the "opposition
rating" (the rating of the average opponent -- not be the exact definition,
sorry), gives a precise method of comparing the toughness of field for two
players who have the same win-loss record. This is an excellent tie-break
method.
Of course, most players are not used to this kind of report. Explaining the
meaning of such reports is necessary, and it will take some time for go
players to accept a more precise report and a different kind of tie-break
method. Player and TD acceptance is, of course, important and may delay any
change even if the change is theoretically for the better.
>> Here is a sample:
>>
>> The following data are shown for each game:
>> difficulty (probability of losing)
>> opposition (difficulty as a rating)
>> win or loss (+ or -)
>> color held (W or B)
>> opponent's name
>> opponent's updated rating
>> handicap
>>
>> LABACH, KEN 2.0
>> 0.85 3.4 - W vs COX, JIM -4.2 (6 stones)
>> 0.26 1.1 + W vs LAING, MARK -3.0 (2 stones)
>> 0.21 -1.1 + W vs COBB, WILLIAM S. -5.4 (4 stones)
>>
>> LAING, MARK -3.0
>> 0.74 -2.2 - B vs LABACH, KEN 2.0 (2 stones)
>> 0.72 -2.3 - W vs CELMER, PAUL -4.1 (2 stones)
>> 0.25 -3.9 + W vs GORNALL, DAVID A. -12.3 (8 stones)
>> 0.10 -4.7 + W vs RILEY, RANDOLPH -14.4 (9 stones)
>>
>> CELMER, PAUL -4.1
>> 0.90 -2.5 - W vs COX, JIM -4.2 (2 stones)
>> 0.76 -3.2 - W vs SHUTT, ROBERT L. -10.9 (8 stones)
>> 0.28 -4.9 + B vs LAING, MARK -3.0 (2 stones)
>> 0.22 -5.1 + W vs COBB, WILLIAM S. -5.4 (even game komi)
From ???@??? Mon Oct 28 10:09:42 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id HAA27685; Mon, 28 Oct 1996 07:05:02 -0800
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id HAA28593; Mon, 28 Oct 1996 07:03:47 -0800
Message-Id: <199610281503.HAA28593@lists1.best.com>
From: bull@netcom.com (michael bull)
Subject: Re: Accelrat Reports
Date: Mon, 28 Oct 1996 07:03:41 -0800 (PST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Length: 2426
Content-Transfer-Encoding: 7bit
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
>
> >>
> >>(Steven Rushing) On the Accellerat reports,
> >>
> >> it does have a report feature which list all games each player has played.
> >> Format is very readable. This is in addition to its AGA ratings input file.
> >>
> >(Mike Bull) But it does not show the pairings by rounds, and uses too much
> >space.
>
> Yes, but the Accelrat reports shows who played tough games and who played
> easy games. To me, this seems much more useful than the order in which
> games were played. I can live with the extra space required, since the
> information is so clear.
The table will also contain this information. I am refering to the form
of the data only, not the content nor the merits of the particular system.
> Taking the accelrat reports a step further, and printing the "opposition
> rating" (the rating of the average opponent -- not be the exact definition,
> sorry), gives a precise method of comparing the toughness of field for two
> players who have the same win-loss record. This is an excellent tie-break
> method.
>
> Of course, most players are not used to this kind of report. Explaining the
> meaning of such reports is necessary, and it will take some time for go
> players to accept a more precise report and a different kind of tie-break
> method. Player and TD acceptance is, of course, important and may delay any
> change even if the change is theoretically for the better.
>
> >> Here is a sample:
> >>
> >> The following data are shown for each game:
> >> difficulty (probability of losing)
> >> opposition (difficulty as a rating)
> >> win or loss (+ or -)
> >> color held (W or B)
> >> opponent's name
> >> opponent's updated rating
> >> handicap
> >>
> >> LABACH, KEN 2.0
> >> 0.85 3.4 - W vs COX, JIM -4.2 (6 stones)
> >> 0.26 1.1 + W vs LAING, MARK -3.0 (2 stones)
> >> 0.21 -1.1 + W vs COBB, WILLIAM S. -5.4 (4 stones)
> >>
> >> LAING, MARK -3.0
> >> 0.74 -2.2 - B vs LABACH, KEN 2.0 (2 stones)
> >> 0.72 -2.3 - W vs CELMER, PAUL -4.1 (2 stones)
> >> 0.25 -3.9 + W vs GORNALL, DAVID A. -12.3 (8 stones)
> >> 0.10 -4.7 + W vs RILEY, RANDOLPH -14.4 (9 stones)
> >>
> >> CELMER, PAUL -4.1
> >> 0.90 -2.5 - W vs COX, JIM -4.2 (2 stones)
> >> 0.76 -3.2 - W vs SHUTT, ROBERT L. -10.9 (8 stones)
> >> 0.28 -4.9 + B vs LAING, MARK -3.0 (2 stones)
> >> 0.22 -5.1 + W vs COBB, WILLIAM S. -5.4 (even game komi)
>
>
>
From ???@??? Mon Oct 28 10:09:46 1996
Received: from lists1.best.com (lists1.best.com [206.86.8.15]) by shellx.best.com (8.6.12/8.6.5) with ESMTP id IAA04825; Mon, 28 Oct 1996 08:09:47 -0800
Received: (daemon@localhost) by lists1.best.com (8.6.12/8.6.5) id IAA07072; Mon, 28 Oct 1996 08:07:19 -0800
Message-Id: <199610281607.IAA07072@lists1.best.com>
From: bull@netcom.com (michael bull)
Subject: Re: Accelrat Reports
Date: Mon, 28 Oct 1996 08:07:13 -0800 (PST)
BestServHost: lists.best.com
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Length: 4901
Content-Transfer-Encoding: 7bit
Sender: td-errors@lists.best.com
Errors-To: td-errors@lists.best.com
Reply-To: td@lists.best.com
To: td@lists.best.com
A tournament record in table form does not preclude or exclude the data in otherforms. It is possible to put the data for 'opposition', difficulty' 'rating'
and 'handicap' in the table. SOS and SODOS are also indicators of 'opposition'
and 'difficulty'.
A 'traditional' table derived from the Accelrat report that follows.
Northern California Open Goe Tournament June 8-9, 1996
win or loss (+ or -)
color held (W or B)
ID Name 1 2 3 4 5 W/G SOS SODS HtH Place
1. Shin, Hong Soo 7.0 w2- b6+ w4+ b3+ b5+ 4/5 14 10 0 1
2. Hung, Joey 6.9 b1+ w7+ b3- w5+ b8+ 4/5 11 8 0 2
3. Liu, Chen-Chang 6.6 b8+ b5- w2+ w1- b7+ 3/5 12 6 0 3
4. Phipps, Ned 6.6 w5 b8+ b1- w7+ w6+ 3/5 10 4 0 4
5. Mueller, Martin 6.5 b4+ b3+ b6- b2- w1- 2/5 16 6 0 5
6. Chien, James 6.3 b7- w1- w5+ w8+ b4- 2/5 11 3 0 6
7. Kim, Paul I. 6.1 w6+ b2- b8- b4- w3- 1/5 13 2 0 7
8. Roberts, Gary 6.1 b3- w4- w7+ b6- w2- 1/5 13 1 0 8
Northern California Open Goe Tournament June 8-9, 1996
The following data are shown for each game:
difficulty (probability of losing)
opposition (difficulty as a rating)
win or loss (+ or -)
color held (W or B)
opponent's name
opponent's updated rating
handicap
Shin, Hongsoo 7.0
0.44 6.8 + B vs Liu, Chen-Chang 6.6 (even game komi)
0.44 6.8 - W vs Hung, Joey 6.9 (even game komi)
0.38 6.6 + B vs Mueller, Martin 6.5 (even game komi)
0.34 6.5 + B vs Chien, James 6.3 (even game komi)
0.32 6.5 + W vs Phipps, Ned 6.6 (even game komi)
Hung, Joey 6.9
0.56 7.1 + B vs Shin, Hongsoo 7.0 (even game komi)
0.45 6.8 - B vs Liu, Chen-Chang 6.6 (even game komi)
0.31 6.4 + W vs Mueller, Martin 6.5 (even game komi)
0.28 6.3 + B vs Roberts, Gary 6.1 (even game komi)
0.21 6.0 + W vs Kim, Paul I. 6.1 (even game komi)
Liu, Chen-Chang 6.6
0.56 6.8 - W vs Shin, Hongsoo 7.0 (even game komi)
0.55 6.8 + W vs Hung, Joey 6.9 (even game komi)
0.49 6.6 - B vs Mueller, Martin 6.5 (even game komi)
0.37 6.3 + B vs Kim, Paul I. 6.1 (even game komi)
0.28 6.0 + W vs Roberts, Gary 6.1 (even game komi)
Phipps, Ned 6.6
0.68 7.1 - B vs Shin, Hongsoo 7.0 (even game komi)
0.42 6.3 - W vs Mueller, Martin 6.5 (even game komi)
0.39 6.2 + B vs Roberts, Gary 6.1 (even game komi)
0.37 6.2 + W vs Chien, James 6.3 (even game komi)
0.30 6.0 + W vs Kim, Paul I. 6.1 (even game komi)
Mueller, Martin 6.5
0.69 7.0 - B vs Hung, Joey 6.9 (even game komi)
0.62 6.8 - W vs Shin, Hongsoo 7.0 (even game komi)
0.58 6.7 + B vs Phipps, Ned 6.6 (even game komi)
0.51 6.5 + W vs Liu, Chen-Chang 6.6 (even game komi)
0.50 6.5 - B vs Chien, James 6.3 (even game komi)
Chien, James 6.3
0.66 6.8 - W vs Shin, Hongsoo 7.0 (even game komi)
0.63 6.7 - B vs Phipps, Ned 6.6 (even game komi)
0.50 6.3 + W vs Mueller, Martin 6.5 (even game komi)
0.47 6.2 - B vs Kim, Paul I. 6.1 (even game komi)
0.37 5.9 + W vs Roberts, Gary 6.1 (even game komi)
Kim, Paul I. 6.1
0.79 7.0 - B vs Hung, Joey 6.9 (even game komi)
0.70 6.7 - B vs Phipps, Ned 6.6 (even game komi)
0.63 6.5 - W vs Liu, Chen-Chang 6.6 (even game komi)
0.54 6.2 - B vs Roberts, Gary 6.1 (even game komi)
0.53 6.2 + W vs Chien, James 6.3 (even game komi)
Roberts, Gary 6.1
0.72 6.7 - B vs Liu, Chen-Chang 6.6 (even game komi)
0.72 6.7 - W vs Hung, Joey 6.9 (even game komi)
0.63 6.4 - B vs Chien, James 6.3 (even game komi)
0.61 6.4 - W vs Phipps, Ned 6.6 (even game komi)
0.46 5.9 + W vs Kim, Paul I. 6.1 (even game komi)
1. Shin, Hongsoo vs. Hung, Joey (even game komi)
2. Kim, Paul I. vs. Chien, James (even game komi)
3. Phipps, Ned vs. Mueller, Martin (even game komi)
4. Liu, Chen-Chang vs. Roberts, Gary (even game komi)
1. Mueller, Martin vs. Liu, Chen-Chang (even game komi)
2. Hung, Joey vs. Kim, Paul I. (even game komi)
3. Chien, James vs. Shin, Hongsoo (even game komi)
4. Roberts, Gary vs. Phipps, Ned (even game komi)
1. Liu, Chen-Chang vs. Hung, Joey (even game komi)
2. Chien, James vs. Mueller, Martin (even game komi)
3. Shin, Hongsoo vs. Phipps, Ned (even game komi)
4. Roberts, Gary vs. Kim, Paul I. (even game komi)
1. Hung, Joey vs. Mueller, Martin (even game komi)
2. Liu, Chen-Chang vs. Shin, Hongsoo (even game komi)
3. Phipps, Ned vs. Kim, Paul I. (even game komi)
4. Chien, James vs. Roberts, Gary (even game komi)
1. Mueller, Martin vs. Shin, Hongsoo (even game komi)
2. Roberts, Gary vs. Hung, Joey (even game komi)
3. Kim, Paul I. vs. Liu, Chen-Chang (even game komi)
4. Phipps, Ned vs. Chien, James (even game komi)
Go Tournament Directors Home Page