Kommentarer från rec.games.go

Diskussioner kring EM 2008 i Sverige
Locked
Mcgreag
Posts: 101
Joined: 12 May 2004, 08:02
Location: Falun

Kommentarer från rec.games.go

Post by Mcgreag »

Kanske inte alla som läser rgg så här är de kommentarer om egc 2008 som postats där. Än så länge är det bara Robert Jasiek som kommenterat.

Första delen är om top grupper, regler, tiebreakers etc. Andra delen mer om kongressen i övrigt.
Number of players

According to the walllist, there were 718 players. None of them are clearcut ghosts; only in case of a very few with just one or two lost games, especially if the last game is a default loss, one might wonder whether they might have been ghosts. So it is a reasonable guess to say that the EGC 2008 was the EGC with the greatest number of players. As usual - estimating very roughly - ca. 70% played in all 10 rounds.

The median rank was about 2k or 1k, what is also usual. Down to 20k the field was thick. 13 players had a weaker rank.

The top was pretty strong. This year the playing strengths of the top Asians were not distributed commonly but most of them were 7d. I.e., the number of Korean 5d and 6d was surprisingly low. Apparently the EGC (and the other tournaments) has become attractive mostly for those with chances to win the tournament. Considering the expenses for flight and accommodation, this is rather natural though. Although the Koreans seem to welcome the EGC as a sort of inofficial Korean amateur championship, there have also been top players from, as usual, Japan but also from China and Taiwan. Some top players had been preregistered but could not come because of, as it seems, visa trouble; among them a Chinese 3p.

There was a supergroup of 32 players and below it a top bar group of 56 players at the start of round 1 and 5 more until the end. This is huge.

Formation of the supergroup

One of the top players was cancelled his flight due to double booking. So he was stranded in Stockholm in the middle of the night. It took some time to figure out whether he could make it to the congress site in time. Luckily another player with a car was around to pick up him.

Unfortunately as usual, Korean holiday efficiency of running to the hotel first and letting only a group leader do the registrations led to unclarity about the presence and desire to enter the supergroup of a few top Koreans. They really should know better by now that each top player has to prove his presence by signing the tournament agreement form.

Otherwise the tournament supervisors were struggling with getting the latest rating list, finding ratings of Asians at all, and sorting out additional background information for unrated Asians. Alexander Dinerstein's help in advance was very useful here.

The local organizers insisted on letting Fredrik Blomback 4d S play in the supergroup despite his low rating and later he confirmed his strength by a 5:5 score. This meant that still 7 Asians playing all 10 rounds could be in the supergroup. Luckily this year this fit very well by chance also because Kim Jung Hyeop, who later won the lightning, did - according to our information - not want to play all rounds at the beginning of the congress and only later changed his mind (and got 8 wins to contact the final top field again).

In earlier years, the minimally sufficient rating was around 2415. This year it was ca. 2445.

The supergroup players can be identified by an initial MMS of 29. The top group players had 28.

Judging from the final results, only one player, Robert Rehm, with 4:6 score did not justify his privilege. The others got at least 5:5, what was not too difficult in view of the big top group. 8 European players were a bit unlucky not to have been included in the supergroup: In particular, Guido Tautorat 4d D scored an impressive 7:3. He is known to play much stronger than usual in games with very long thinking times because he is a slow thinker and can discipline himself well to avoid blunders then. Taking into account only this tournament, you might wonder why he does not call himself 5d or 6d - but his achievements in tournaments with shorter thinking times are that of a solid 4d indeed. - Among the 5 non-supergroup Europeans with 6:4 were three Dutch players (apparently their ranking system is still a bit strict; or should we blame the rather great number of Dutch participants?), Lothar Spiegel 4d A (apparently it was an above average tournament for him), and the author. Then there were two Finnish 3d with 7:3, and everybody does know that Finnish ranks are 1 grade too strict...

Strange incidents

Simplified Ing Rules were used on all boards. The only rules troubles the author has heard of or watched were a) players still unfamiliar with the Ing fill-in counting or b) players misinterpreting the 4 komi stones as 8 points komi instead of the actually used 7.5 points (with no different effect on the result though). However, it was reported that one of the 5d players lost by 0.5 points because of not having attended the author's lecture on Endgame Strategy under Different Rules and therefore making the standard strategic mistake of first filling the dame and then fighting and losing the last endgame ko.

A different 5d appeared to have made an attempt to move twice in a row by reinterpreting the timing of quick alternate moves and pressings of the clock. The appeals committee would not let him though.

More important were the cases of 7d's losing on time. The new Ing clocks, which are terrible in quite some respects (e.g., the language button can easily cause a malfunctioning display and they beep to confirm each pressing of the clock - a real horror in a room with 100 clocks!), could not serve as a excuse though because it was still easy enough to see seconds counting down 60, 59, ... and so not confuse that with the new placement of the 03 byoyomi periods. - In the round 3 game Hong Seul-Ki 7d KOR - Hong Seok Ui 7d KOR, a difficult connection fight was taking place, Seul-Ki was taking his time and continued to do so excessively when in byoyomi. So naturally the time ran out. A few minutes later, Seok Ui could not resist any longer to point out that fact. A referee had to come and arrived about 10 minutes later (the main building was 5 minutes walking afar). - In the round 10 game Alexander Dinerstein 7d RUS - Park Jong Wook 7d KOR, the clock was running properly except that it did not speak (isn't that even an advantage?:) ). Alexander claimed that his loss on time should not count because of that but the referee Andreas Ensch decided that it was the player's responsibility to check whether clocks would be behaving regularly and to avoid running out of time. A very reasonable decision. According to a 5d kibitz, Alexander was losing the game deciding semeai by 1 liberty anyway. The outcome of the game was very important for both players: Park won the European Open title (otherwise his SOS would have been too low when tied on MMS with Hong Seok Ui) and Alexander missed his chance to be the only European with 8 wins and got second European place by being 1 SOS behind this year's winner of the European Championship title Catalin Taranu 7d ROM.

"European" was still undefined but during this year's AGM the EGF Committee declared a preliminary, trial definition "citizen of a European country, citizen of an EGF country, or resident in Europe / an EGF country for at least 5 years". This let the tournament supervisor, your author, wonder whether that new rule should also apply to Noguchi Motoki 7d J, who has lived in France for 8 years according to him and some witnesses one and a half days before the end of the congress. Then it was difficult to get a proper decision by the present members of the EGF Committee and the other tournament supervisor in time and transmit it to the pairing director. So effectively it could only be told definitely just after round 10, and the decision was that for this year the rule change would be too late because the tournament had already been started days earlier. Previously the player had been listed as J in the walllist. However, apparently some congress organizers (unduly) considered it their responsibility to change it to F out of their own initiative. - The author had been watching the last round 10's games in the room for boards 4 to 18. Among them Svetlana Shikshina 7d RUS - Ilya Shikshin 7d RUS (her brother, they are the previous two European Champions) was the last game to finish. In an exciting semeai in the middle game, Ilya found a very astonishing clamp tesuji leading to a ko and giving him the lead. Later he made a strategic mistake by starting to run on dame with a stone that could have been sacrificed easily and efficiently. So it needed an overplay tenuki by Svetlana and kill by Ilya to settle the game in his favour. After round 9, he and Mero Csaba 6d H had good enough SOS to have chances for the title if Alexander would lose his game. Catalin gave a convincing show about cutting and capturing options in his last round game against Mero: whichever option Mero would choose, it would favour Catalin, and this state of affairs lasted from the middle game until the end. - Back to the nationality issue, in the second the last game was finished the author hurried to the main building (which, as said before, was 5 minutes away) to fulfil with Matti Siivola their tournament supervising duties of verifying the top players' order, i.e., in particular whether the program did calculate the SOS correctly. Such takes some minutes! The author asked the congress organizers to wait another 5 minutes with the prize giving but they wanted to start it immediately while SOS, etc. were still being checked. Nothing was wrong with the SOS but only during the beginning of the prize giving could the author notice the sudden mistake of Noguchi's nationality in the walllist. Immediately he informed one of the organizers besides the prize giving stage who did not want to interfere before the announcement of the main tournament top winners. So it happended what must but should not have happened - Noguchi was announced as the 6th strongest European and given the prize for that. It took some discussions after the ceremony to award also Ondrej Silt 6d CZ the same prize. Another Swedish organizer wanted to prevent that arguing that the EGF should pay. Regardless of whose fault the mistake might be considered to be, such would have been an unwise decision because several tournament participants noticed the mistake already during the prize giving and wouldn't have to be told explicitly about its occurrence. Issuing the prize twice has been the reasonable decision by other congress organizers and EGF committee members to avoid later criticism. Nevertheless one wonders why the prize giving needed to be held in such an extreme hurry just a few minutes after the last round. One does not even want to imagine what might happen if more serious mistakes did occur: like wrong SOS and announcement of the wrong person as a European Champion. A punctual time schedule (like, e.g., for round 1) is a nice thing but correctness about who are the winners is of much greater importance.

Future rules changes

Same procedure as every year: During the AGM some major system changes would be proposed and rejected. Actually the Finnish proposal was pretty good except for some details and the missing consultation of the strong players before. The Dutch proposal (KO of 8 players after round 7) was rejected with 8:9 votes; although it was the maybe best McMahon-KO proposal so far, it failed to address the great variety of MMS of seeded Europeans after round 7, who might have been 4 and 7 wins if many strong Asians have entered the main tournament. Avoiding tiebreakers at the end would be great but the price of very high imprecision at the qualification to the KO is simply too high.

As should have become obvious, a clear definition of European has been in more urgent need than ever. Also its application during the registration day is of very great importance. It does not suffice to ask a player whether he is European - rather he needs to understand the difference between citizenship and residence even if his English is weak.

At 23:59 and two minutes before closure of the AGM, the author initiated at least two minor improvements of the EGC system: the top bar and number of non-Europeans in the supergroup may be handled a bit more flexibly now. So too huge top bar groups and too doubtful lotteries for expelling 7d's from the supergroup will become much less likely. These improvements should have been made many years earlier but better late than never.

Tiebreaker discussion

The European Champion title depended heavily on SOS. Let us study this in greater detail for the top Europeans with 7 wins. Place 10 is actually place 9b shared with a Korean.

Opponents

6 Catalin Taranu 34+ 1- 9+ 15- 51+ 26+ 18+ 8+ 2- 11+
7 Alexander Din. 2- 9- 59+ 78+ 16+ 36+ 14+ 3+ 5+ 1-
8 Ilya Shikshin 14+ 4- 20+ 23+ 25- 27+ 13+ 6- 55+ 16+
10 Pal Balogh 16+ 3- 40+ 25- 93+ 35+ 29+ 11- 21+ 13+

Direct comparison opponents

6 Catalin Taranu 8+
7 Alexander Din.
8 Ilya Shikshin 6-
10 Pal Balogh

Due to the great percentage of non-Europeans at the top direct comparison could not have produced non-trivial information. The criticism of the strong Europeans to play too few games among each other is justified when there is a great percentage of non-Europeans at the top.

SOS normalized at 0 for the 7 point players

# Name SOS per round SOS SOS-1 SOS-2

6 Catalin Taranu -2 2 0 -1 -2 -2 -1 0 1 -1 = -6 -4 -2
7 Alexander Din. 1 0 -3 -3 -1 -2 -1 0 0 2 = -7 -4 -1
8 Ilya Shikshin -1 0 -1 -1 -1 -2 -1 0 -3 -1 = -11 -8 -6
10 Pal Balogh -1 0 -2 -1 -3 -2 -2 -1 -1 -1 = -14 -11 -9

Catalin and Alexander were just 1 SOS apart. If we ignore the other doubtful noise contained in SOS, then Catalin won the title because he got only one Korean 7d in the first two rounds while Alexander got two of them (and these were two of his lost games). This meant that Alexander (played in the - at that time - second room and) got low SOS opponents in rounds 3 and 4. SOS always tends to give low SOS opponents if one loses in the early rounds. Since SOS is used, one has to live with that "feature". However, the real problem here is the timing of when he got two of his Korean 7d opponents. In the first two rounds, with different pairing luck, he might as well have gotten two 4d's of the supergroup, who scored a normalized -2 SOS, and only then the 7d's. So it amounts to pairing luck in the first two rounds which of Catalin or Alexander was to become the European Champion under the retrospect assumption that both would make 7 wins.

Usage of the tiebreaker SOS-2 instead of SOS would have swapped their places. Some say that early pairing fate has to be accepted so that all opponents' tournament-strengths can be measured. The author says that disregarding the worst noise (as SOS-2 does) is better. Anyway, the fact that different tiebreakers lead to different winners means that sharing the title would be more appropriate. One approach would be to share it among all MMS 7 Europeans. Another approach might be to share it among all those that are at place 1 under at least one of the tiebreakers SOS, SOS-1, SOS-2, or DirectComparison:

Places under different tiebreakers

(0 means "not applicable".)

SOS SOS-1 SOS-2 DC

6 Catalin Taranu 1 1 2 0
7 Alexander Din. 2 1 1 0
8 Ilya Shikshin 3 3 3 0
10 Pal Balogh 4 4 4 0

So, if one does not want to share the title between all 4 players, one might at least have shared it between Catalin and Alexander.

For reference, here is a study of 2007:

Places under different tiebreakers

SOS SOS-1 SOS-2 DC

4 Ilya Shikshin 1 2 2 2
5 Alexander Din. 1 1 1 1

In 2007, the title would have been shared as well under the study of all players placing 1 under at least one "reasonable" tiebreaker. (SOS-SODOS was still being used; Ilya won that tiebreaker combination to get the title.)
Del 2.
Preface

This report on the congress does neither mention names of winners nor discuss the main tournament. You can find both elsewhere.

Venue

The congress was held in the town Leksand, which is situated in the middle of Sweden. Although some typical Swedish architecture and a nearby lake could not be overlooked, one still felt like just somewhere in Europe. It was not a remote place amidst pure nature but a town with a reasonable train connection. Reaching it was not the easiest thing though because some planes would arrive in the middle of the night.

Wheather

The first few days were sunny and rather hot. Afterwards it was cloudy with partly soft rain. Almost ideal conditions for thinking well during the games.

On August 1, 10:38 the maximum of a partial eclipse was hidden behind light to medium grey clouds. If one did not know about it, one might have overlooked that the light was darker than usual for the kind of clouds - it was a bit like dawn. From ca. 10:40 to 10:43 the eager observer (most were unsuspectingly continuing their games) could see the sun covered for about 30% when shining through tiny gaps of less thick clouds. No specific security glasses were needed; the clouds fulfilled that purpose very well. So it was an impressionistic grey in grey show of light and shadow. Not particularly spectacular and without change of environmental colours but still interesting. The total eclipse was taking place in the Polar Sea and in deep Russia.

Registration

Some organizers of the 2000 congress were helping. Therefore the registration was very efficient. After only a few minutes, a participant had gone through the procedure including the Tournament Agreement form. - Although the station was not too far, there was even a congress taxi service when a train would just have arrived.

Professionals

Presumably there was a record of numbers of professional teachers. One just would not notice. As usual, most professionals were doing rather little (like only giving a few simultaneous games on the main tournament's days) and only a few like Saijo Masataka could be seen regularly. Even he is not as active as in previous years though - age is paying its price. - For congress organizers, it appears to be difficult to get more out of the professionals. Weak English meets weak organization effort? Not that the organizers were weak overall but also in this year they did not generate an amount of professional activity that would have justified their great number. So again one gets the impression that professionals were enjoying having holiday. Of course they may but still the strong feeling remains that much more would be possible.

Part of the problem was an uningenious note "professional" on the name tag, which was as white as that of the 718 ordinary players. It was easier to identify an organizer because of his yellow tag. Why not use orange or red or yellow on violet instead of white? It is rather difficult to distinguish Asian players from Asian professionals unless one already happens to know their faces.

The schedule for professional activities was a desaster, too: E.g., every day lectures on 14:00 and 16:00. 14:00 was a time when one might still be playing or watching top players. 16:00 was the start of rapid. So every day Saijo would have his lecture at 16:00 and the author had to miss it because of playing rapid. What the hell is so difficult about being more flexible?! Nothing prohibits lectures at different times like 9:00, 15:00, or 22:00.

So it was very easy to get rooms with demonstration board and multi-media facilities. A few amateurs or semi-professionals used that to give their own lectures and show a movie: Catalin (Teaching in Europe), Gunnar Farnebäck (Why Recent Programs Have Become Stronger), the author (Endgame Strategy), Kalli Balduin (movie "The Go Master", 2006). Judging from attendance, applause, and demand for more, amateur lectures are at least as welcome as professional lectures. What makes a good lecture then? Careful preparation and solid knowledge of the pupils' demands and weaknesses! This also explains why Saijo offers the best of the professionals' lectures: He spends a lot of time to prepare each! Amateurs with the courage to hold lectures as well do very thorough preparations, too.

There was also a history symposium with amateur and professional speakers. But it was mostly disappointing. Speeches that could have been held in 3 minutes took an hour. Well known facts were warmed up as if they were new. There were a few exceptions but no really impressive research came up. E.g., a highlight was an elegant hypothesis about the origin of the go board but very little evidence was supportive.

Side tournaments

All the usual and some uncommon tournaments were held. Here the report is highly selective in line with the author's preferences and attendance.

The weekend tournament had 667 players in the walllist. Due to the AGM decision two days earlier, reasonably sized supergroup and top bar group could be formed.

The registration procedure for the rapid was efficient for avoiding ghosts but a real problem: Every day a player had to mark his name afresh and he had to do so in the main building. Some praised the congress as the best organized ever; if so, why were there no registration walllists in the other buildings? It was a great pity having to leave watching the top board games just to register in a buliding 5 minutes walking afar.

The pairings of the side tournaments were done reasonably efficiently. The criteria for seeding to a second stage were kept too secret though. E.g., during the 13x13, the author first had to ask whether the SOS-SOSOS shown in the walllist were indeed the seeding tiebreakers. By chance, it did not matter because exactly 16 players had 3 or 4 wins after the 4 rounds Swiss preliminaries. E.g., the 9x9 and lightning had groups. Neither would one know how many should be qualified nor which kind of selection would be made. As a participant, this always leaves a very bad feeling about whether the organizers might be fair, careless, or corrupt. In retrospect, everything was fine: The unique group winners were qualified while shared group winners (like 4 players with 2 of 5 wins each!) on points (fortunately no tiebreakers!) would have to play one further game in the finals. A nice solution but it would have been much better if that procedure had been announced before the start of the tournaments.

The so called "lightning" was 10 minutes with 1 move per 5 seconds byoyomi and had up to 9 Japanese stones handicap. Only experts on the Ing clocks like the tournament winner in the final game could take advantage of that. Already in the 1/32th final, a lot of kibitzes were watching.

The 13x13 handicap was one Japanese stone per 5 ranks and komi values of 8.5, 4.5, 0.5, -3.5, -7.5. The 9x9 handicap was one Japanese stone (on the 3-3 points though) per 8 ranks and komi values of 6.5, 4.5,..., -7.5. These are rather tough handicaps against the stronger players. That the dan players still won a good percentage has to be blamed on the kyu players though: Compared to years with more reasonable handicap systems, they played as much weaker as the better handicap allowed them... It is a shame to encourage weaker strategy like this. Instead stricter handicaps should be used to motivate the kyus to play stronger! E.g., even a kyu can read ahead the remaining 10 moves of the endgame, especially since in the 13x13 the time was 20 minutes and 3*5 seconds byoyomi. But almost all simply would not do it and then wonder after the game why they lost by 4.5 instead of winning by 0.5. Answer: The dan player sold a gote as sente while elsewhere a bigger sente was still available and the kyu player would simply be afraid and answer everything. - The number of participants was rather small: around 45. Of course. The tournament was held on a Wednesday. Only the very great number of congress participants prevented a tiny number participants like in the previous year. When in other years congress organizers dismiss the tournament as unimportant, they even have a reason because they create it by themselves when putting the tournament on the excursions day.

Overall conclusion

Given the relatively small number of organizers, they produced about as good a congress as they could. Still it could have been better in many details. The organization was lacking a remote, overall view with that possibilities for improvements could have been perceived. This is a problem of perspective in every year: The organizers are lost in pride about their great amount of work and so fail to see a lot of easy things with a potentially great effect. Besides the examples above, here are a few more: Publish the pairings 30 minutes before the round instead of 0 minutes. Publish the main tournament results until 22:00 instead of until some unpredictable time around midnight or the next morning. Remove outdated announcements. Do not set 1 second byoyomi when there is announced to be none. Explain how to verify the set byoyomi on a new Ing clock. Always (not only sometimes) provide a pen at a walllist for entering a tournament. All such is easy, fundamental, and done wrongly year after year. Leksand may have been one of the better organized congresses but it did not shine on the top. With comparatively little extra effort it might have done.

Locked