Report: 30th World Amateur Go Championship - Tournament Conditions
Robert Jasiek
Date: 27th - 30th May 2009.
Venue: Ecopa indoor stadion, Fukuroi, Shizuoka, Japan.
Playing area: The stadion's hall was huge and suitable for physical,
athletic sports. For the tournament, it was divided by an artificial
visual wall in two halves. One half was used for the tournament, the
other half was used for side events. The temperature was a tiny bit
too high but this is better than a too low temperature due to an
exaggerated air conditioning might have been.
Spectators and journalists: Some local spectators visited the event.
They were very friendly to the players. Some took photos and presented
them to the players, others collected signatures etc. The greatest
surprise was a man with a huge bag full of very well selected Go
books, which he gave as presents to every player he could talk with.
Pieter Mioch and James Davies interviewed some of the players.
Sponsors: The original sponsor seems to have dropped out but the Nihon
Kiin ensured free flight tickets, accommodation, and meals for the
participants. According to the tournament booklet, further support
came from NEC, Autodesk, Chubu Electric Power, Kowa Kogyo, Suzuyo &
Co, Tokyo Club, Ministry of Foreign Affairs, Japan Foundation,
Shizuoka Prefecture, Fukuroi City, Shizuoka Shimbun, Shizuoka
Broadcasting System.
Information: Preliminary information was sent to the potential
participants half a year before the tournament. This made it easily
possible to choose a suitable flight. The tournament rules were
announced reasonably well and key questions were answered. The rules
of play were named but not published; getting answers about them was
much more difficult. During the tournament, a very big and visually
clear table showed the players' names, countries, flags, previous
year's place number of their country, opponent numbers, current
numbers of wins for every round, and colours. After the last round,
the final place was denoted, too. Unfortunately, the column SOS was
not filled and columns for the other tiebreakers were not provided.
The table was updated regularly and timely. For the 2nd, 3rd, 4th,
5th, 7th, and 8th rounds, result lists were printed, copied, and made
available for everybody. It is possible that some for also the 6th
round were printed but in that case not in a sufficient amount. The
result lists were delayed a bit but still available reasonably fast,
which might well be explained by a careful check by the tournament
organizers. The lists let it be possible for the players to verify
possibly existing mistakes, if any. For some rounds, there were two
versions of the lists: one in English with +/- for wins / losses, the
other one in Japanese with white / black circles for that purpose. The
lists showed initial place number, name, country, opponent number and
win / loss sign per round, score (number of wins), SOS, SOSOS, current
place, number of pairings up, number of pairings down (how impressive;
this may well serve as a model!), number of games with Black, number
of recorded games. The last tiebreaker (direct comparison) was not
shown in the lists though. Like many result lists, the lists showed
opponent numbers as initial places (i.e. previous year's places of the
related countries); especially Europeans find this very much harder to
read than the European style of result lists that show opponent
numbers as current places. In some cases (like online pairing lists),
the initial place column was missing and it required considerable
effort to find out the next round's opponent or a player's previous
opponents.
Furniture: The tables were arranged in three rows. Each table had much
space around it. All the tables were surrounded by an unobstrusive
fence to keep off spectators, atlhough occasionally a few kibitzes
came inside. Since everybody respected the players, there was no need
for a strict entry control to the surrounded playing area. Tables and
chairs were as simple as one might expect in a stadion but were solid,
convenient, and high enough to allow much room for one's legs. They
were also wide enough to carry board, bowls, clock, name cards, flags,
and possibly a notebook. In short, furniture was simple but very
appropriate. Food was not allowed on the tables except for closed
plastic bottles of water. Recalling the number of beers thrown across
the boards during the evenings in the hotels, this might be called a
wise rule.
Playing material: Good boards were used. The stones were doubtful
though: While there is little wrong with simple glass stones (except
that one might wish the luxury of clam shell), the plastic bowls
contained stones of varying thicknesses. Such is inappropriate in any
tournament, especially in a world championship. A simple checking of
the stones could have increased the average quality of the played
moves because varying stones on a position distract a player's
thinking. The clocks were a delight: Very easy to use, pleasant
design, extremely clear figures for both basic time and byoyomi always
visible to both players, automatic counting and showing of numbers of
remaining Canadian byoyomi moves, and - best of all - not any
disturbing voice! Some players might prefer to have clocks without a
red light indicating the player to move, but the light is not really
distracting. Like every unknown clock or time system, it may require a
bit of practice to apply it well. Leszek Soldan was a victim of this
inexperience and lost a won game in round 5 by forfeiting on time when
thinking for several minutes about a decision on his last stone of a
byoyomi period to decide between a good and a better move. Maybe some
players would appreciate a short beep not only upon entering byoyomi
but also 10s before the end of a byoyomi period. The author is very
happy with the general silence of the clocks though: they do not
distract the players by unnecessary sounds. The used clocks belong to
the best on the world-wide market of electronic clocks. Other
international tournaments should strive for using clocks of equally
high functionality and quality.
Noise: The playing area was reasonably silent. The side events in the
other half of the hall were held with a low noise level and everybody
was respecting the players' need to concentrate. The most distracting
noise came from occasional mobile phones, which their owners had
forgotton to switch off. Of course, also the few film cameras
installed and deinstalled right next to a table for maybe a minute or
two per game were a bit distracting but the camera men knew their
stuff and tried to be as decent as anyhow possible. Even when dozens
of kibitzes gathered (maybe a bit too) tightly around a table during
the late endgame of an important game, everybody was being
extraordinarily silent. Clearly go players (participants or fans) knew
how to respect still active players. Therefore tournament organizers
did not see any need to interfere and everybody was allowed to
appreciate the exciting games. A restricted overall number of visitors
may have helped a lot here. Last but not least, also the scribes were
extraordinarily silent while doing their jobs. If anything, then the
cables to the notebooks on the affected tables may have been a bit
distracting.
Organizers: There were sufficiently many tournament organizers,
referees, and translators.
Tournament system: Colours were chosen by nigiri in each round. This
led to extreme imbalances: Ukraine got White in each round, Malaysia
got Black in all 8 games. As the Law of Great Numbers tells us, it
requires an infinite number of rounds to balance colours determined by
nigiri. To get a colour balance within one tournament and its only
finite number of rounds, a better algorithm than random choice would
be needed. Until about round 4, it was not clear to the players that
nigiri would be used in all rounds. Before the start of the
tournament, some announcement said something like that nigiri would be
used during the first round to determine the colours in also the later
rounds. The tournament system was an 8 rounds Swiss with even games,
6.5 komi, 90 minutes basic time plus 15 stones in 10 minutes byoyomi
and the final placement criteria 1) number of wins, 2) SOS, 3) SOSOS,
4) Direct Comparison (called 'direct encounter' in the tournament
rules). The thinking time is maybe the minimum that might be called
appropriate for a world tournament, although the author would prefer a
significantly greater basic time. 90 minutes of the WAGC are a big
difference compared to the shorter thinking times of the WMSG or the
KPMC: The frequency of blunders during a player's basic time is
reduced very greatly. Sufficient thinking time has a very great impact
on the quality of the games. Any tournament system may have
difficulties to handle the very broad range of player strengths, as
they occur during a WAGC. Swiss is a pretty good solution in principle
and about 8 rounds suffice for determining the number 1 player. The
exact order of all other players depends mostly on the tiebreaker
lottery though. One may say though that every player gets his deserved
number of wins with a high probability; already this is a remarkable
achievement. Considering the very low degree of accuracy of
tiebreakers, final result lists should not be overinterpreted. Why do
many still state a player's final place in a manner as if it were the
gist of ultimate wisdom? The real problem, however, is that the
tiebreakers do not determine the tournament winner with meaningful
accuracy if there is more than one player with the most wins. The
author suggests two possible great improvements in future to replace
the usage of tiebreakers for the final placements entirely: 1) Let the
tournament have a flexible number of 7, 8 or 9 rounds and end as soon
as exactly one player leads on numbers of wins. 2) After 8 rounds,
break a tie among the players leading on number of wins by means of
playoffs with short thinking time (like 30 minutes basic time plus 20s
boyoymi). Note that tiebreakers should still be used for the purpose
of making pairings as fair possible.
Tournament rules: With a few noteworthy exceptions, tournament rules
were announced well. Those rules conveyed only verbally include the
following: a) A player presses the clock with the hand that played the
stone. b) White decides the position of the clock. If White is late,
then Black decides the position. c) The clock may be stopped to remove
several captured stones. d) A pass is performed by simply pressing the
clock. Stopping the clock in byoyomi for the purpose of going to the
rest room was unspecified. The tiebreaker Direct Comparison was
undefined in two respects: 1) 3 or more tied players that have not all
played a game against each other. 2) iterative or non-iterative
application to 4 or more tied players. Since a tournament organizer
was too busy and James Davies was one of the official translators and
is known to be a rules expert, the author pointed out to him that
problem of insufficient definition. Later Davies contacted the
tournament organizers who said only that (1) in case of 3 players
could not break the tie. So effectively the third tiebreaker Direct
Comparison remained undefined for most non-trivial cases. This is sad
because it could mean that the final placements and even the
tournament winner would have to be decided by an arbitrary ad hoc
decision AFTER the last round. Such a possibility of an unfair (for
example, player A having beaten players B and C might feel treated
unfairly when a decision after the last round would declare that a tie
for 3 or more players is not dissolved) and unworthy (when tournament
organization politics overrides part of the players' shown Go skill)
decision could be avoided very easily before round 1, e.g., by
declaring iterative application of Direct Comparison or by declaring
that ties are not resolved by Direct Comparison if 3 or more players
are tied. There was a light (no T-shirts or shorts etc.), reasonable
(allowing a great flexibility), and equal (no discrimination of
gender) dressing code. Due to the increased temperatur during the
later rounds, some players wore T-shirts nevertheless.
Rules of play: It was announced that World Amateur Go Championship
Rules were used. Since the 1980 version is newer than the 1979
version, one may imply that the former was being meant. Before the
first round, the author asked Davies how to apply those rules whenever
they contain mistakes that contradict Go theory (like specifying that
a seki consists of [exactly] two groups). Davies, being a translator,
was careful enough to accompany his statements by the remark that they
would be only unofficial. On the other hand, the author had tried to
ask a tournament organizer who, before interrupting with some more
urgent task, said that he would fetch the rules expert and looked in
the direction of Davies. Therefore it appears that the then official
answer would have been essentially the same. Davies answer was: "When
problems have occurred in the past, referee decisions have always
applied common sense instead of the literal rules text. One should
expect the same also during this tournament." In author's opinion,
rules that are not used ought not to be announced as being used,
heavily flawed rules ought never to be used at all, and the fact that
verbal common sense was used in the function of substituting rules of
play ought to have been announced to all the players since they have a
right to know how rules are actually meant to be applied.
Pairing analysis: The remainder of the report studies the quality of
the pairings. The statistical evaluation methods are the same as used
for analysis of WMSG and KPMC.
Used pairing methods as detected by reverse engineering from the
actual pairings:
Abbreviations:
I = initial place / previous year's country standing
C = current place by 'number of wins' - SOS - SOSOS
W = number of wins
Round 1:
- random pairing
Round 2:
- cross pairing by I
- the player with the mean I in his score group was paired down
- the player with the smallest I in his score group was paired up
Rounds 3-8:
- fold pairing by C, else by I
- resolve "already played" conflict firstly by next same W opponent
down the list
- keep each player's pairing up and pairing down values each below 2,
if possible
- the player with the smallest C, else the smallest I in his score
group was paired down
- the player with the greatest C, else the greatest I in his score
group was paired up
The following result list has been edited manually to show the
opponent numbers after the last round. For compression of the width,
names and countries have been abbeviated where necessary. sss means
SOSOS, pu is the pair up / pair down balance of the players having
been paired up or down at least once.
C Player|Country dk R1 R2 R3 R4 R5 R6 R7 R8 + sos sss pu
1 Yuqing Hu CHN 8d 43+ 32+ 9+ 6+ 11+ 2+ 10+ 4+ 8 42 323 -2
2 S.-H. Yoo KOR 7d 16+ 22+ 12+ 4+ 3+ 1- 18+ 13+ 7 46 309 -2
3 Nai Chan HK 6d 65+ 30+ 10+ 5+ 2- 15+ 6+ 12+ 7 39 311 -1
4 K.-T. Yeh TAI 7d 35+ 20+ 14+ 2- 38+ 7+ 5+ 1- 6 45 303 0
5 Kanazawa JAP 6d 23+ 38+ 7+ 3- 17+ 8+ 4- 15+ 6 44 297
6 Jia Tan SIN 6d 8+ 48+ 13+ 1- 19+ 26+ 3- 37+ 6 42 287 -1
7 L. Heiser LUX 6d 34+ 41+ 5- 44+ 12+ 4- 21+ 11+ 6 38 299
8 T. Debarre FRA 5d 6- 27+ 59+ 31+ 9+ 5- 17+ 10+ 6 38 294 +1
9 O. Silt CZE 6d 47+ 44+ 1- 29+ 8- 32+ 16+ 14+ 6 38 287
10 Juyong Koh CAN 7d 33+ 45+ 3- 34+ 13+ 14+ 1- 8- 5 42 295 +1
11 Csaba Mero HUN 6d 28+ 17+ 15+ 25+ 1- 21- 20+ 7- 5 42 292 -1
12 Merl. Kuin NET 6d 31+ 42+ 2- 18+ 7- 55+ 25+ 3- 5 40 277 0
13 David He AUS 7d 37+ 36+ 6- 33+ 10- 31+ 19+ 2- 5 39 292 +1
14 C. Burzo ROM 6d 46+ 21+ 4- 30+ 16+ 10- 24+ 9- 5 39 291
15 Eric Lui USA 7d 18+ 39+ 11- 52+ 22+ 3- 26+ 5- 5 39 282 -1
16 Iatsenko UKR 5d 2- 49+ 35+ 23+ 14- 45+ 9- 39+ 5 37 276
17 Ragnarsson SWE 4d 64+ 11- 20+ 32+ 5- 23+ 8- 30+ 5 37 274
18 Tormanen FIN 5d 15- 51+ 43+ 12- 46+ 33+ 2- 32+ 5 34 284 +1
19 Paisal Th. THA 4d 24- 56+ 45+ 27+ 6- 29+ 13- 28+ 5 34 269
20 Boon Teng MAL 4d 57+ 4- 17- 49+ 34+ 27+ 11- 33+ 5 33 279
21 Longyang Li NZ 4d 63+ 14- 34- 48+ 28+ 11+ 7- 29+ 5 33 270 +1
22 L. Soldan POL 5d 66+ 2- 40+ 28+ 15- 24- 36+ 26+ 5 33 266
23 Jan Ramon BEL 4d 5- 57+ 47+ 16- 43+ 17- 40+ 25+ 5 32 275
24 R. Jasiek GER 5d 19+ 50+ 25- 26- 62+ 22+ 14- 35+ 5 32 258 -1
25 D. Surin RUS 6d 29+ 40+ 24+ 11- 26- 35+ 12- 23- 4 36 276
26 Viktor Lin AT 5d 27- 53+ 50+ 24+ 25+ 6- 15- 22- 4 35 268
27 Stankovic SER 5d 26+ 8- 48+ 19- 36+ 20- 37- 47+ 4 34 262 -1
28 S. Tsolmon MON 4d 11- 58+ 39+ 22- 21- 41+ 31+ 19- 4 34 257
29 E. Aharoni ISR 3d 25- 55+ 46+ 9- 39+ 19- 38+ 21- 4 34 253
30 Simeunovic BH 5d 56+ 3- 42+ 14- 35- 60+ 34+ 17- 4 34 252 -1
31 D. Chacon SPA 4d 12- 54+ 41+ 8- 44+ 13- 28- 51+ 4 33 265
32 J. Reyes COL 5d 58+ 1- 65+ 17- 50+ 9- 44+ 18- 4 33 262
33 Matt Cocke UK 5d 10- 37+ 36+ 13- 53+ 18- 61+ 20- 4 33 257 -1
34 S. Tabares ARG 3d 7- 64+ 21+ 10- 20- 47+ 30- 49+ 4 33 256
35 M. Zakanj CRO 3d 4- 61+ 16- 56+ 30+ 25- 43+ 24- 4 32 264
36 Anh Tran VIE 2d 49+ 13- 33- 41+ 27- 42+ 22- 48+ 4 32 252
37 F. Sulak TUR 2d 13- 33- 62+ 50- 54+ 46+ 27+ 6- 4 30 251 +1
38 Wan Lou MAC 5d 61+ 5- 44- 58+ 4- 53+ 29- 46+ 4 29 259 +1
39 Bustamante MEX 4d 52+ 15- 28- 59+ 29- 48+ 45+ 16- 4 29 256
40 J. Taboada VEN 2d 51+ 25- 22- 54+ 45- 50+ 23- 44+ 4 29 247
41 Marquard. DEN 3d 60+ 7- 31- 36- 59+ 28- 52+ 43+ 4 28 245
42 Michelena URU 3d 55+ 12- 30- 43- 51+ 36- 57+ 45+ 4 27 252
43 A. Davies SA 2d 1- 63+ 18- 42+ 23- 49+ 35- 41- 3 35 243
44 M. Poliak SLK 3d 54+ 9- 38+ 7- 31- 52+ 32- 40- 3 34 250
45 T. Suc SLV 3d 53+ 10- 19- 57+ 40+ 16- 39- 42- 3 32 249
46 J. Flood NOR 4d 14- 52+ 29- 47+ 18- 37- 62+ 38- 3 30 243
47 Yamamoto PER 4d 9- 60+ 23- 46- 56+ 34- 53+ 27- 3 30 236
48 Garbarini IT 1d 62+ 6- 27- 21- 63+ 39- 55+ 36- 3 30 235
49 C. Neto POR 1d 36- 16- 63+ 20- 58+ 43- 54+ 34- 3 28 239
50 A. Aguilar ECU 1d 59+ 24- 26- 37+ 32- 40- 51- 60+ 3 28 234
51 F. Tehhani MOR 1d 40- 18- 53- 60+ 42- 56+ 50+ 31- 3 28 220
52 Gallagher IRE 1k 39- 46- 55+ 15- 57+ 44- 41- 62+ 3 26 232 +1
53 Petrauskas LIT 1d 45- 26- 51+ 65+ 33- 38- 47- 61+ 3 24 234
54 D. Mueller CH 1k 44- 31- 61+ 40- 37- 58+ 49- 57+ 3 24 229
55 N. Roussos CYP 4k 42- 29- 52- 61+ 65+ 12- 48- 59+ 3 24 228 +1
56 M. Singh IND 1d 30- 19- 66+ 35- 47- 51- 65+ 58+ 3 22 226
57 M. Garero PHI 1k 20- 23- 64+ 45- 52- 63+ 42- 54- 2 27 212
58 M. Aguero COS 2k 32- 28- 60+ 38- 49- 54- 64+ 56- 2 25 211
59 Sakajiri BRA 2d 50- 62+ 8- 39- 41- 61- 66+ 55- 2 24 210 -1
60 Harutyuny. ARM 2k 41- 47- 58- 51- 64+ 30- 63+ 50- 2 23 211 +1
61 S. Korbut BRU 4k 38- 35- 54- 55- 66+ 59+ 33- 53- 2 23 209 +1
62 M. Pina CHI 7k 48- 59- 37- 66+ 24- 65+ 46- 52- 2 21 215 +1
63 E. Estrada GUA 7k 21- 43- 49- 64+ 48- 57- 60- 66+ 2 20 213 -1
64 T. Nedev BUL 3k 17- 34- 57- 63- 60- 66+ 58- 65+ 2 18 208
65 Baghirov AZE 3d 3- 66+ 32- 53- 55- 62- 56- 64- 1 24 200
66 Rabin Raya NEP 5k 22- 65- 56- 62- 61- 64- 59- 63- 0 19 185 +2
The sum of all pair up/down excesses is TP = 3. Considering the
considerable number of players in the tournament and their great
variety of ranks, the value is impressively small.
Table with values V: For rounds 1 and 2, the values are set to the
default 1 to express triviality of relative places. Places within a
score group are distingusihed by SOS - SOSOS.
C Player|Country dk 1 2 3 4 5 6 7 8 + sos sss
V Vd
1 Yuqing Hu CHN 8d 1 1 16/3 8/3 4/2 2/1 6/6 3/2 8 42 323
73,67 -54,67
2 S.-H. Yoo KOR 7d 1 1 16/3 8/8 4/4 2/2 17/16 11/9 7 46 309
48,22 -80,12
3 Nai Chan HK 6d 1 1 16/3 8/3 4/1 11/9 6/2 11/11 7 39 311
83,00 -45,34
4 K.-T. Yeh TAI 7d 1 1 16/3 8/1 24/24 11/10 6/4 1/1 6 45 303
78,10 -50,24
5 Kanazawa JAP 6d 1 1 16/3 8/3 17/13 11/8 6/3 11/8 6 44 297
66,46 -61,88
6 Jia Tan SIN 6d 1 1 16/16 8/3 17/17 11/11 6/5 19/19 6 42 287
41,07 -87,27
7 L. Heiser LUX 6d 1 1 16/3 25/17 17/6 11/3 17/17 11/4 6 38 299
87,05 -41,29
8 T. Debarre FRA 5d 1 1 34/33 25/21 17/7 11/5 17/12 11/2 6 38 294
87,11 -41,23
9 O. Silt CZE 6d 1 1 16/3 25/21 17/11 20/17 17/9 11/5 6 38 287
66,37 -61,97
10 Juyong Koh CAN 7d 1 1 16/3 25/24 17/13 11/7 1/1 11/10 5 42 295
51,93 -76,41
11 Csaba Mero HUN 6d 1 1 16/3 8/7 4/3 20/20 17/14 11/7 5 42 292
54,31 -74,03
12 Merl. Kuin NET 6d 1 1 16/3 25/24 17/13 20/20 17/6 3/3 5 40 277
60,54 -67,80
13 David He AUS 7d 1 1 16/1 25/17 17/5 20/16 17/13 3/1 5 39 292
111,54 -16,80
14 C. Burzo ROM 6d 1 1 16/3 25/13 17/7 11/6 17/15 11/6 5 39 291
69,44 -58,90
15 Eric Lui USA 7d 1 1 16/3 25/25 17/12 11/4 17/10 11/3 5 39 282
84,82 -43,52
16 Iatsenko UKR 5d 1 1 34/4 25/15 17/10 20/18 17/8 19/18 5 37 276
70,65 -57,69
17 Ragnarsson SWE 4d 1 1 34/4 25/10 17/3 20/13 17/5 19/15 5 37 274
107,00 -21,34
18 Tormanen FIN 5d 1 1 34/4 25/1 24/19 20/9 6/1 19/17 5 34 284
196,09 67,75
19 Paisal Th. THA 4d 1 1 34/4 25/8 17/2 20/14 17/4 19/14 5 34 269
138,18 9,84
20 Boon Teng MAL 4d 1 1 34/4 25/23 24/13 20/6 17/1 19/13 5 33 279
189,77 61,43
21 Longyang Li NZ 4d 1 1 34/4 25/3 24/2 11/1 17/3 19/11 5 33 270
238,32 109,98
22 L. Soldan POL 5d 1 1 34/4 25/8 17/7 20/10 19/15 19/8 5 33 266
90,01 -38,33
23 Jan Ramon BEL 4d 1 1 34/4 25/10 24/13 20/8 19/17 19/3 5 32 275
118,22 -10,12
24 R. Jasiek GER 5d 1 1 16/3 25/23 17/17 20/11 17/2 19/9 5 32 258
112,65 -15,69
25 D. Surin RUS 6d 1 1 16/1 8/1 17/13 20/19 17/11 19/16 4 36 276
113,17 -15,17
26 Viktor Lin AT 5d 1 1 34/2 25/5 17/4 11/2 17/7 19/11 4 35 268
156,07 27,73
27 Stankovic SER 5d 1 1 34/4 25/17 24/21 20/15 19/19 16/14 4 34 262
61,24 -67,10
28 S. Tsolmon MON 4d 1 1 34/4 25/17 24/23 20/15 19/6 19/5 4 34 257
97,17 -31,17
29 E. Aharoni ISR 3d 1 1 34/4 25/5 24/17 20/7 19/13 19/7 4 34 253
101,65 -26,69
30 Simeunovic BH 5d 1 1 34/4 25/13 24/19 11/11 19/5 19/4 4 34 252
110,11 -18,23
31 D. Chacon SPA 4d 1 1 34/4 25/5 24/10 20/5 19/14 16/16 4 33 265
99,00 -29,34
32 J. Reyes COL 5d 1 1 34/4 25/16 24/17 20/4 19/10 19/2 4 33 262
154,11 25,77
33 Matt Cocke UK 5d 1 1 34/32 25/10 24/22 20/12 15/14 19/5 4 33 257
66,54 -61,80
34 S. Tabares ARG 3d 1 1 34/4 25/1 24/11 20/19 19/16 16/15 4 33 256
159,57 31,23
35 M. Zakanj CRO 3d 1 1 34/4 25/15 24/4 20/3 19/12 19/10 4 32 264
128,45 0,11
36 Anh Tran VIE 2d 1 1 34/3 25/15 24/3 20/16 19/7 16/12 4 32 252
117,83 -10,51
37 F. Sulak TUR 2d 1 1 16/15 25/3 17/12 20/2 19/1 11/1 4 30 251
324,62 196,28
38 Wan Lou MAC 5d 1 1 34/4 25/15 17/1 20/12 19/8 16/9 4 29 259
158,01 29,67
39 Bustamante MEX 4d 1 1 34/4 25/22 24/8 20/11 19/4 19/1 4 29 256
241,92 113,58
40 J. Taboada VEN 2d 1 1 34/4 25/19 24/8 20/8 19/3 16/6 4 29 247
126,43 -1,91
41 Marquard. DEN 3d 1 1 34/4 25/8 17/7 20/6 15/7 16/4 4 28 245
117,14 -11,20
42 Michelena URU 3d 1 1 34/4 25/11 17/16 20/5 15/11 16/7 4 27 252
91,73 -36,61
43 A. Davies SA 2d 1 1 34/4 25/13 24/11 20/17 19/9 16/13 3 35 243
75,78 -52,56
44 M. Poliak SLK 3d 1 1 34/4 25/7 24/15 20/14 19/10 16/10 3 34 250
82,46 -45,88
45 T. Suc SLV 3d 1 1 34/4 25/21 24/15 20/1 19/18 16/11 3 32 249
177,29 48,95
46 J. Flood NOR 4d 1 1 34/4 25/14 24/6 20/18 15/15 16/8 3 30 243
82,31 -46,03
47 Yamamoto PER 4d 1 1 34/4 25/11 17/8 20/2 15/5 16/2 3 30 236
190,22 61,88
48 Garbarini IT 1d 1 1 34/4 25/19 17/15 20/10 15/12 16/5 3 30 235
82,78 -45,56
49 C. Neto POR 1d 1 1 16/2 25/1 17/13 20/2 15/9 16/3 3 28 239
244,87 116,53
50 A. Aguilar ECU 1d 1 1 34/34 25/18 24/6 20/13 15/13 12/12 3 28 234
53,86 -74,48
51 F. Tehhani MOR 1d 1 1 16/15 8/6 17/3 11/4 15/2 16/1 3 28 220
233,87 105,53
52 Gallagher IRE 1k 1 1 16/2 25/1 17/11 20/6 15/8 12/11 3 26 232
173,58 45,24
53 Petrauskas LIT 1d 1 1 16/2 25/2 24/4 20/9 15/10 12/10 3 24 234
137,43 9,09
54 D. Mueller CH 1k 1 1 16/2 25/3 17/5 11/6 15/6 12/8 3 24 229
114,83 -13,51
55 N. Roussos CYP 4k 1 1 16/2 8/5 17/1 20/1 15/3 12/9 3 24 228
281,07 152,73
56 M. Singh IND 1d 1 1 16/2 25/8 17/9 11/7 7/4 12/7 3 22 226
81,34 -47,00
57 M. Garero PHI 1k 1 1 16/2 25/3 17/5 11/8 15/4 12/5 2 27 212
128,03 -0,31
58 M. Aguero COS 2k 1 1 16/2 25/8 17/4 11/5 7/6 12/6 2 25 211
95,12 -33,22
59 Sakajiri BRA 2d 1 1 34/1 25/3 17/10 11/10 1/1 12/4 2 24 210
181,43 53,09
60 Harutyuny. ARM 2k 1 1 16/2 8/3 4/1 20/1 7/3 12/1 2 23 211
287,00 158,66
61 S. Korbut BRU 4k 1 1 16/2 8/4 4/2 11/1 19/2 12/3 2 23 209
206,50 78,16
62 M. Pina CHI 7k 1 1 16/1 8/2 24/1 11/2 15/1 12/2 2 21 215
370,00 241,66
63 E. Estrada GUA 7k 1 1 16/2 8/1 17/1 11/3 7/7 1/1 2 20 213
178,00 49,66
64 T. Nedev BUL 3k 1 1 16/2 8/6 4/4 2/2 7/2 3/1 2 18 208
88,86 -39,48
65 Baghirov AZE 3d 1 1 34/4 25/23 17/16 11/9 7/5 3/3 1 24 200
60,29 -68,05
66 Rabin Raya NEP 5k 1 1 16/2 8/6 4/3 2/1 7/1 3/2 0 19 185
114,33 -14,01
Vm = 128,34
TP = 3
TV = 70,06
Tournament pairing quality: TQ = 1 / ( 4 * 71,06 ) = 0,00352.
Comparison:
The TQ of the 3rd Korea Prime Minister Cup 2008 is 5,15 times worse.
The TQ of the World Mind Sports Games - Men Team Group A is 1,67 times
better.
The TQ of the German Women Go Championship 2008 is 13,04 times better.
The TV of the 3rd Korea Prime Minister Cup 2008 is 1,38 times worse.
The TV of the World Mind Sports Games - Men Team Group A is 3,92 times
better.
The TV of the German Women Go Championship 2008 is 7,08 times better.
The TP of the 3rd Korea Prime Minister Cup 2008 is 4,67 times worse.
The TP of the World Mind Sports Games - Men Team Group A is 2,67 times
worse.
The TP of the German Women Go Championship 2008 is 3 times better.
Comments on the tournament pairing quality of the WAGC2009:
Although the pairing quality was good enough so that one would hardly
notice immediately where something might be wrong, there is still
quite some scope for improvement. While this is so from a theoretical
point of view, realizing it in practice will not be easy - rather it
requires very careful study before a tournament. Usage of different
pairing programs with algorithms of global optimization might be worth
trying, too.
The overall pairing up and down balance was as good as one could
achieve in a big tournament with a great variety of playing strengths.
Nevertheless, one might consider a more frequent per player balance (a
player who was paired up should be paired down and vice versa) to see
whether this can improve the standard deviation of TV.
The first round used random pairing and the second round used cross
pairing. Whether this makes sense depends on one's assumptions. The a
priori aim of letting each player in a world championship have
initially equal chances to win the tournament suggests random pairing
in rounds 1 and 2. The assumption that players of the same country
have similar playing strengths and that therefore one might use the
previous year's order suggests cross pairing in rounds 1 and 2. Using
each of random and cross pairing once in the first two round is a
compromise on the basis of acknowledging validity of both
presuppositions. One cannot say clearly which assumption might be the
best; therefore either of the four pairing schemes for the first two
rounds and including random and / or cross pairing would be valid.
From a view of preferring smaller TV, probably cross pairing in both
round 1 and 2 would be better. One should accept a somewhat greater TV
though if thereby the mentioned a priori aim is emphasized.
Improvements of TV should be sought by different means. The used fold
pairing in rounds 3 to 8 is considered reasonably good. There appears
to be scope for improvement though. The limited number of rounds in a
tournament might mean that for every round starting with round 3 some
hybrid between cross and fold pairing might be better. It will require
research though to find out exactly which. So far one can notice only
some possible causes of players with extreme V values: a) A player
with an opponent in round 3 so that the relative place in a scoring
group is extreme and belongs to a small subgroup has a rather high
probability to get an extreme V. b) Folding by SOS - SOSOS is not
always as balancing as folding by Vd might be. c) Pairing up / down a
minimal / maximal place's player of a score group might not have the
best balancing effect on TV. d) Global pairing strategies should be
used, too.
Comments on the usage of tiebreakers:
For the purpose of ordering the final result places, usage of SOS
(and, of course, also SOS - SOSOS) is very doubtful. This can be seen
already from one aspect. Consider players #2 and #3 after the last
round:
Player SOS_by_round_1_opponent SOS
#2 5 46
#3 1 39
The randomly chosen round 1 opponents of the 7 point players gave them
SOS contributions from those opponents that differed by 4 SOS points!
It is very likely that player #3 would have win his round 1 game also
if he had been given an opponent with finally 5 points. Compare the
difference of SOS of the players #2 and #3: The difference is 7
points. Thus over half of that difference can be explained already by
the pairing luck in round 1. Player #2 was lucky to draw the
tournament-stronger opponent, so player #2 had much greater chances to
be ahead of player #3 in the final results list! Hence the author
recommends not to use SOS for the purpose of final results ordering.
That V orders the players differently than SOS encourages the idea of
sharing places rather than splitting hairs.
It is hard to understand why SOSOS as the second pairing tiebreaker
would be any better or worse than lottery as the second pairing
tiebreaker would have been. Using lottery would have been more honest.