Author: Pierre Bourget
Date: 10:36:36 02/14/05
Go up one level in this thread
On February 14, 2005 at 12:44:51, Kolss wrote: >Hello everyone, > >As I have always enjoyed reading other people's reports on tournaments, here is >my own brief account of CCT7 as the operator of Ikarus... > >First of all, I would like to thank everybody involved in the event, especially >Peter for organizing it, Andy (Zek) for his assistance in setting up accounts >etc., for IM Schroer for commenting throughout the whole tournament, for >everyone else who was involved and whom I don't know or may have forgotten. >Also, thanks a lot to all participants in helping to make this a great >experience, especially to our opponents for some quite elucidating games. > >All in all, I think CCT7 went extraordinarily smoothly. There were only a few >small incidents, but that is of course unavoidable... Everything was perfectly >organized, the rounds could always start almost on the minute. That is to some >degree due to the time control, which at 50'+3'' was very well chosen and can >only be recommended for future CCT, but of course also due to Peter's incredible >speed at updating results and posting them as well as the new pairings on his >web site. > >The next part will be a very subjective story - from the perspective of Ikarus >or rather one of its developers. If you are not interested, this is the point to >quit! :-) > >Before the tournament, we had several goals, of course (not necessarily in order >of descending importance...): >1. Test our current version - there is no better way of getting new ideas and >insights than from intensely watching a real game. >2. Play against some new and interesting programs - this is sort of the same as >#1. >3. "Meet" other programmers and exchange some thoughts. >4. Have fun. >5. Make Ikarus known a bit more. >6. ? > >#5 partially stems from the fact that Ikarus was only seeded at #21. We both >(i.e. Muntsin and myself = Munjong) thought that that was rather off, as we >would probably have seeded Ikarus in the top ten. But then again, people did not >seem to know much about Ikarus, and I had also failed to publish our hardware in >time (Centrino 2.0 GHz, 768 MB RAM, by the way). Besides, we did not care that >much about the seeding anyway - you have to win the games in the end, no matter >against whom. Of course, it also had the positive side that we were quite >certain we would overfulfill the expectations... > >As for the goal in terms of final ranking, it was hard to predict. We thought >there were at least seven or eight programs that would seriously compete for the >victory, and about 20 to go for the top ten... :-) For us, top ten or something >like 6 to 6.5 points was sort of a minimum goal, though ending up worse would >not have been a catastrophe - in such a dense field, a bit of bad luck is enough >to catapult you down significantly. > >I personally expected 7.5 points to be needed to at least jointly win the >tournament - I did not expect that one program would conduct such a rout >though... > >So here goes the account round by round - don't worry: analysis not included!! > >Day 1: > >Round 1: Ikarus - SEE (1-0) >Being seeded 21st, we played 43rd seed, SEE. Not knowing much about SEE, we >chose to play something solid. Ikarus came out of a Nimzo-Indian with a minute >disadvantage, but that is normal for this opening (at least our program seems to >have difficulty judging it properly). After some moves, it gradually accumulated >a nice advantage and finally converted a passer into victory. That was a good >start! > >Round 2: Pharaon - Ikarus (1-0) >As a reward, we were to play Pharaon on the black side. This would be a tough >task, as we considered Pharaon to be one of the serious contenders for the top >spot. Both programs being out of book rather quickly, a quite balanced position >emerged. > >In this position > >[D]3r2k1/pp1r1ppp/4b3/3N1n2/8/1B1R4/PP2KPPP/3R4 w - - 0 21 > >Ikarus missed the strong 21. g4! and ended up in a knight vs. bishop endgame >with a passive king, which turned out to be hopeless, as Ikarus realized way too >late. Pharaon played the ending very well and easily converted it into the >deserved victory. >So we were stuck with 1.0/2 and thinking that it could only become better... No, >that is a bit exaggerated - while we did not think that Ikarus is really much >weaker than Pharaon, we knew that it would be perfectly OK to lose to this >strong program. > >An exclamation mark was set by DanChess by beating Yace, and also Matacz, seeded >way in the bottom half, ended up with a perfect score after this round. > >Round 3: Ikarus - Petir (1-0) >We knew nothing about Petir, but were lucky to get a good opening. Petir came >out with a locked-away bishop, failed to set up counterplay on the queen side, >and Ikarus, effectively playing with a piece more, quickly won the game. It was >nice to see a #13 announced without tablebases... > >Movei beat Crafty, for us perhaps the biggest surprise of the round besides >DanChess' draw against Pharaon. > >Round 4: CEng - Ikarus (1-0) >Again a completely unknown to us - please pardon our ignorance... CEng played >without opening book, something we did not know before. In a Sicilian, Ikarus >came out with a slight disadvantage, managed to keep the balance and gradually >equalized. Things started turning when CEng allowed Ikarus to enter a slightly >better bishops' ending (same colors). Ikarus increased its advantage, when >suddenly the following position came up: > >[D]8/8/6B1/4p2P/1p1pp1k1/1P6/2PK1Pb1/8 b - - 0 55 > >Here I thought that it had blown its advantage, but after a bit of shuffling, it >came up with a nice sequence which led to the following: > >[D]8/8/7k/7b/1p2B3/1P2p3/2P5/5K2 b - - 0 69 > >I suppose that this position is won for black and therefore the above one should >be as well, but deeper analysis would take too long, and I leave it to others... >The end was a tablebase mate with KBP-KB. >With 3.0/4, we were back in the game! > >Round 5: Chepla - Ikarus (1/2-1/2) >In a sharp Sicilian, Ikarus emerged with what it considered a substantial >advantage. However, Chepla did a good job at keeping the position complicated, >not trading queens and thus sustaining some annoying counterplay against the >open black king. In the end, Ikarus traded off into an ending KRPP-KRN where it >thought it was still better, but had to aquiesce to the draw after all. I am not >sure if black had a win anywhere at all - that will also have to remain open to >more detailed analysis. Well defended by Chepla, although fighting for the draw >is not really what you hope for as white... > >So we finished the first day with 3.5/5, which was OK, had some lousy moments >with the loss and the gross misevaluation in the last game - but hey, that is >where you learn most from! > >Zappa had beaten SOS to maintain a perfect 5/5 score - the tournament victory >seemed to be booked, especially considering the hardware Zappa was running on. >The only way to catch it would be to score 4/4 on the second day while beating >Zappa itself and hoping for it to give up another half point somewhere else... >That of course seemed ridiculously out of reach. > >Day 2: > >Round 6: Ikarus - ChessThinker (1/2-1/2) >ChessThinker was still undefeated, so we knew it would be tough. Besides it was >playing on superior hardware as well. Ikarus was surprised in the opening by a >pawn sacrifice, which left it with a poor bishop. However, ChessThinker probably >did not treat the position optimally, as suddenly things turned, and it was >black that was stuck with the bad bishop! I found that very remarkable, but of >course did not mind. Ikarus played the next part very well and increased its >advantage to 2.5 pawns. But when ChessThinker was almost on the ropes, it fired >up some counterplay by sacrificing a second pawn, thus activating all its >pieces. Ikarus completely underestimated this counterplay. In the end, it did >not have the guts to continue in this position: > >[D]4R3/kp2N3/p5r1/P7/7P/5rP1/P7/5bK1 w - - 0 65 > >With 65. Nxg6 (65. ... Bb5 66. Kg2 Bxe8 (Ra3!?) 67. Kxf3 Bxg6 68. g4 +-), it >might have played for the full point, but instead went for the perpetual with >65. Nc8+ Kb8 66. Ne7+ Ka7. Nice fight by ChessThinker after a maltreated opening >- another lost half point for Ikarus, although before the game, we thought that >a draw would be OK. I guess that is one of the big differences to the top >programs - they squeeze out full points out of these kinds of positions. > >Zappa beat Pharaon in a dramatic game and must have started to feel lonely up >there with 6/6 and 1.5 points to the next neighbor... > >Round 7: Amateur - Ikarus (0-1) >This was one of the games where I can only say: "Beyond the fourth ply..."! It >was simply way too complicated for me, but somehow Ikarus seemed to gather >centipawn by centipawn (which I actually missed completely, because I had a >couple of phone calls), finally trading into a good ending with the more active >pieces. In contrast to the game against Pharaon, this time Ikarus was on the >better side of the topical good bishop vs. not so great knight ending (although >there also was a pair of rooks left) and convincingly converted the advantage >into a full point. I particularly liked the move 63. ... Bh7! in the following >position, although it is not difficult for engines: > >[D]4R1b1/6P1/8/2p5/p7/3k1P2/2r5/6K1 b - - 0 63 > >The point is 64. Rh8?? a3! 65. Rxh7?? a2 66. g8Q a1Q# checkmate! >Things seemed to go pretty well now with 5.0/7. > >Zappa was closing in on victory by beating Movei for an incredible 7/7 and now a >two-point lead, which meant that it only needed one single draw from the >remaining two rounds to be sole winner. > >Round 8: Ikarus - Yace (1-0) >Yes, I confess: I had been hoping that we would play Yace in this tournament. >Yace is a really strong program, although Dieter would never admit that. >Besides, Dieter is always an enrichment in any tournament - it is nice to chat >with him, and we had participated at some of the Paderborn tournaments without >ever getting to play each other, so it was about time... The game commenced >quietly, when Dieter surprised me with the comment: "This could go quickly - are >you going to play Bxh6 soon?" I replied: "No, Ikarus does not play such >moves..." Sure enough, two moves later it played 16. Bxh6, although it was not >really tactically inspired, but seemed rather positional. I seem to have the >incredible talent of missing out the interesting parts - this time, my body was >called to sustainment measures (aka dinner). When I came back, Ikarus showed +9 >- it had converted more active piece placement in a way that is of course beyond >my modest capabilities. >Now things were definitely going well - 6.0/8, which meant we would get to play >"the monster", as Zappa had been tenderly called. > >Zappa meanwhile had wrapped up his victory by drawing ChessThinker. 7.5/8 was of >course an unbelievable demonstration by Zappa. On the other hand, there was to >be noticed the clear downward trend - obviously Zappa had some serious problems >with motivating itself to keep concentrating... We hoped to exploit this fatigue >and at the same time make use of our own forward momentum. No, just kidding - of >course we would have been quite contented with a draw... > >Round 9: Ikarus - Zappa (1-0) >Fortuna had an eye upon us, and because Zappa was seeded higher than Ikarus, it >(Zappa) had color change priority and therefore got black... Ikarus came out of >a rather short Scotch opening with maybe a minute advantage due to the better >pawn structure. Zappa had some counterplay against the king, but it turned out >to be easily driven off. Ikarus was daring enough to start pushing its own >kingside pawns (with opposite-side castles!), something which I naturally missed >again due to some other interruption I don't remember any more... It gained a >pawn, only to give it back in exchange for a passed pawn on the seventh rank. >After the score wobbled around a fair bit, this pawn proved decisive, as in this >position > >[D]4r3/kpp1P3/2p3p1/p7/P3QPPp/2P4P/3qRK2/8 b - - 0 45 > >Zappa had to give the rook for the passer, as 45. ... Qxc3 would fail to 46. >Qe3+ Qxe3 47. Rxe3+-, followed by the push of the f-pawn. The last hurdles >(avoiding perpetual check) were mastered, and when the queens went off the >board, the game was over. > >I feel we deserve some sort of charity price for this, as I do not think that >winning such an event without a single loss can be healthy! :-) Anyway, I guess >that Anthony was able to get over this easily - still an impressive tournament >victory for him with Zappa, so again congratulations to this marvelous >performance! > >I would also like to point out the IMO most remarkable showings at CCT7: It is >most noteworthy that ChessThinker was the only program that went undefeated >through all nine rounds. The (for me) most surprising (positive) result was by >DanChess - I had not even known it really before, but obviously Dan must have >improved it quite a lot recently; it should be watched in future events! Also >Gosu, CEng, and Chompster (the latter written in Java, if I gathered correctly!) >did very well. And of course everyone deserves applause for the mere fact of >producing a program that is able to play and entering it in the tournament - you >are all winners, at the very least from a learning perspective! > >To wrap up the tournament for Ikarus, it goes without saying that we were more >than pleased with the result. If someone had predicted 7.0/9 and a clear second >place for us, we would have said: "Well, yes, theoretically not entirely >impossible, but..." Apart from that, we were also very happy that the whole >tournament worked out perfectly from a technical point of view: not a single >crash, not a single disconnect - when we compare that to the pathetic failure >last year (router died, no internet connection, hastily set-up backup quit, >etc.), that was like ascending from Hades and flying up to the sun... > >There is of course one downside to this result: realistically, it can only go >downhill from here on...! > >Best regards - Munjong. Congratulations ,is Ikarus will be available for download sometime in the future ? Pierre
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.