Forum

Hello everyone. Long time no posting here, I hope that everyone is doing well. I am with Helen in her final, under 10. It's for kids born in 2017 and 2016. Helen was born on December 22, 2017 so she is just about the youngest here. I haven't done chess in a while. Vasik goes to high school, is a special math program there, and what time he does have for chess, he works with Tomek Markowski. I am nowhere near the level he needs for a chess trainer. Simon is really into fencing, it's 20+ hours per week plus tournaments and camps, he's ranked #1 in Poland. He has stopped attending all chess lessons, group or individual. Helen on the other hand is just getting started. The junior stuff in Poland is really intense. At the final here the ratio of adults to kids in the dining room is at least 2:1. These kids have entourages, although the numbers are also boosted by grandparents. Saw one GM so far. This is all for under-10. Helen is mostly trained by Iweta. From the openings point of view they've been focusing on open positions, for example 1. e4 e5 2. Nf3 Nc6 3. Bc4 Bc5 4. c3 Nf6 5. d4, which gets pretty crazy. Helen will be playing this with white if given the chance, and I'll need to look at it myself before we prepare it. Against 3. .. Nf6 she'll be playing 4. Ng5. Iweta normally doesn't play like this with white but she decided, I think correctly, that kids should start with open positions. I've settled on the approach of showing her a couple of variations, no more than 10 half-moves, and then quizzing all subsequent themes from there. Tactical themes first, then positional ones if there aren't enough tactical themes in the position. The most important quizzes are about things to avoid. The philosophy is that the bane of preparation at this age is the kid forgetting the preparation and either playing the wrong idea in the wrong position, or deciding at the board that there is something even "better". What is difficult for a kid can be quite surprising. The round 1 opponent plays the French and the Kan. So, the first position I showed Helen was this, white to play. Should white play 6. e5: This was hard for her, it took her several minutes to work it out, with me having to shoot down her first several suggestions. In the French, we prepared 1. e4 e6 2. d4 d5 3. e5 c5 4. c3 Nc6 5. Nf3 Qb6 6. Bd3 cxd4 7. O-O, black to play: Here 7. .. Nge7 should be met with 8. cxd4 but 7. .. Bd7 should not because there are no checks on b5. I first showed her the position after 7. .. Nge7 and asked if white can play 8. cxd4, then I showed her the position after 7. .. Bd7 and asked if white can still play 8. cxd4. This seems much harder to me than the previous position, it's three plies further in compared to an immediate check, but somehow she was able to solve both in less than a minute. I've seen from preparing Vasik and Simon previously that a lot of these trainers have a really good sense of what the kids are capable of.

Yesterday Vasik and I played a blitz marathon at home. Time control was 3+2. He beat me in a longer match for the first time, the overall score was 17-11. The result wasn't entirely a surprise, last time it was just 10-8 in my favor and that was a few months ago and I trailed for most of the match. Below is a mini-report, for posterity. We both played Accelerated Dragons with both colors. I prepared several side-lines for the match as at this point Vasik knows the main lines better than I do. My strategy was to try to play endgames and quiet positions, as we've discussed here several times. I'm not actually sure that was a great idea, several times I got outplayed in endgames. My only defense as black against the Maroczy was 7. .. Ng4, then after 8. Qxg4 Nxd4 9. Qd1 I prepared and played both 9. .. e5 and 9. .. Ne6. I first played 9. .. e5 a few times, had a small minus there overall, with one cool win: Move 15, black to play and win: If this puzzle was a little harder I'd call it a "DP kind of position". Then I switched to 9. .. Ne6, which went a little better. Several times we had the following position, where the score was around even, move 12, white to move: Eventually Vasik switched to 5. Nc3 and then I had big problems. We had the following position several times, move 19, black to move: One of our rules was no computer analysis between games. I kept playing 19. .. Re8 here, which is a mistake, the only move for equality is 19. .. f5. I was playing 19. .. Re8 because one of white's threats is 20. h6 Bh8 21. h7+, and 19. .. Re8 defends against that by preparing 20. h6 Bf8 21. h7+ Kh8. What I didn't see is that the much more natural 19. .. f5 also provides squares for the bishop, for example 19. .. f5 20. h6 Be5. If we keep the no-engines-between-rounds rule, I'll need to prepare multiple defenses against 5. Nc3 next time to avoid getting stuck like this. When I was white we played this endgame several times, move 15, black to move: 7. .. Nxd4 is Vasik's main defense against the Maroczy nowadays, we call it the Markowski although it has no official name. I picked this endgame, white is supposed to be a little better and I liked that it's an endgame. However, I scored I think a small minus from this position. Not sure if that was bad luck or if the go-for-endgames heuristic wasn't as clever as I thought. Here is one interesting K+P endgame we got out of this, move 35, white to play: I played 35. Kd2. Who has the better chances in that queening race? We also got the main line a few times, I prepared the following, move 15, black to move: My idea was Nb5 and after the eventual Bxb5 cxb5, white will play Bc4, Qe3, and f3-f4-f5 with pressure against f7. We had the following position several times, move 22, black to move: This is another position where the no-computers-between-rounds rule had an effect. We're both on our own since 16. .. Rfc8. Here the computer wants to play 22. .. Na4 with a score of +0.3. Vasik always played 22. .. Nd7 here, then white can get +2.0 with 23. e5 Rxc4 24. exd6. Neither of us realized that so I kept playing 23. Qe2, then after 23. .. Nb6 black is already a tiny bit better. I still slightly preferred white's position after 23. Qe2 and didn't see anything better, so I kept repeating it.

The Polish Family Team Championship was held yesterday in Grodzisk. It was quite strong, there were 6 GMs, plus a bunch of tough teams which were evenly matched on boards 1 & 2. Iweta played with Vasik, they had 5.5/9 and took 5th. They played more top-heavy teams, Iweta scored 4/9 while Vasik scored 7.5/9. I played with Simon, we had 5/9 and took 18th. We played more evenly matched teams, I had 6.5/9 and Simon had 3/9. Simon actually played quite well, he beat a CM and had a bunch of better positions and near-draws with some (for him) very good opponents with ratings over 2000 FIDE. We took the prize for the top team from our club. As always there were a huge number of 1:1 matches, which is quite likely whenever an evenly matched team plays a top-heavy team. A total of 18 teams finished with 5.5/9 or 5/9. I had two games against GMs, a loss with Michal Krasenkow and a win against Monika Socko (who is a men's GM). Round 3 vs Michal Krasenkow: It went on for a while but I can't reconstruct it. We traded queens and he was two pawns up in a rook + opposite-colored-bishops endgame. I don't think it was ever particularly close to draw, but it also wasn't easy for a long time. Round 7 vs Socko: This one did not go on for a while. The game with Krasenkow was pretty decent, except for one huge meltdown on my part where my thinking was just a mess. Move 18, white to move: Obviously white has a very nice position. Computer gives +1.0. But how to proceed? I thought for several minutes. The first move I looked at is the obvious 18. f5, the problem is that black has 18. .. Nxf5 19. Bxf5 Rxe2. White isn't really happy with this, and I didn't see any convincing follow-up which would justify it. I thought white needed to play 20. Bf4, then black has several options. What else to do, though? If white doesn't play 18. f5, then black will play 18. .. f5 himself. 18. g4 is met with 18. .. f5. 18. Nd4 Nxd4 19. Qxd4 is also met with 19. .. f5. 18. Raf1 is another move which invites 18. .. f5. It's also not easy for white to challenge the e-file, 18. Re1 is kind of awkward. The answer is an obvious move which I didn't even consider: 18. Rd1, putting a rook on the same file as black's queen, controlling d4, and hitting d5. The e-file isn't that important, white's bishops defend all of the important squares. Instead I cycled back and forth between the other moves, then decided I can't spend this much time here and just lashed out with 18. f5, without seeing anything concrete. It's a big mistake, dropping the score to under -1. Move 22, white to move: This is a good example of mistakes coming in pairs. White should just "admit his mistake" and play 22. Bc2. I went for 22. Be4 Rd8 and now I realized that 23. Rd3 loses to 23. .. Nxd3 24. Bxd6 Nf2+, which is pretty obvious. So, I had to play 23. Bxe5 Qxe5 and the scores are -2.5. The game with Socko was a huge mess. Move 13, white to move: I thought for a while and couldn't resist 13. b5 Nxe5 14. b6+ Qd7 15. Bb5 Nxf3+ etc, which is actually fine, the scores are +0.0. Other moves such as 13. Bc5 are also around +0.0. The craziness is about to begin. Move 21, white to play and win: Black has just played 20. .. Qxh2, which turns out to be a huge mistake. White now has a win. I won't spoil it. I missed the win myself and played 21. O-O-O, then the score goes back to +0.0. But not for long. Move 22, white to play and win: White has another win, which I again missed. Instead I played 22. Bxd5 exd5 23. Rxd5, dropping the score to +1. But again not for long. Move 26, white to play and win: Now white has another win, which I finally found: 26. Rd6 and black can't meet all of the threats, the biggest being 27. Rxg7+ Kxg7 28. Rxe6 with further threats of 29. Qxf6+ and 29. Re7+. This motif I saw at the board. Move 27, white to play and mate: And now it's easy: 27. Rd7+ and black is mated.

Greetings. I know it's a long shot but if anybody is interested in visiting Warsaw, there are some good tournaments coming up. July 8-16: Najdorf Open in Warsaw, classical chess (https://chessarbiter.com/turnieje/2025/ti_2628/results.html) July 18-19: Najdorf Festival in Grodzisk, rapid, run by our club (https://chessarbiter.com/turnieje/2025/ti_2665/) Any of you are very welcome to stay with us during any or all of this, with or without kids, etc. Helen will be out of the house (camp, then grandma) so we even have an extra bedroom. In the meantime Vasik is in the middle of a chess marathon since the school year ended -- 2nd Junior league, now Kielce Open, then the above tournaments. Here is a cool game from Kielce, round 4: Move 18, white to move: Should white play 18. Bxa7? There is quite a bit of calculation involved. The opponent is my round 3 opponent from the 3rd league, that was the draw in the Caro-Kann with 3. e5 Bf5 4. h4 which I posted here where I messed up with cxd5 leading to Qc8+ Bd8 and then .. Qc6.

Greetings everyone. The Grodzisk club just got back from the junior second league. They took third place and advanced to the first league, despite being a very young team. Vasik and Ania got gold medals on boards 1 & 5 despite being 3 and 2 years younger than their age limits. Vasik got his live FIDE rating up to 2140, he's up 140 points this month. There is always a danger of an overshoot when playing this much but I do think he's close to me strength-wise now (the ratings have fallen, mine probably still has a bit more to drop). When we disagree about something, he seems to be right about half the time, etc. Here are three cool games. Round 6: Critical position was on move 20, black to move: Should black play 20. .. Nxb2? The computer doesn't like it but Iweta and I did. Round 8 was funny, until it almost wasn't: Move 5, white to move: White inexplicably blundered with 5. Nxd4 cxd4 6. Ne2 running into 6. .. Qb5+. There were apparently around 20 people looking in amazement. What's even more crazy is that white later got a chance. Move 32, black to move: The game hasn't been as easy as you might guess, Vasik ended up using almost all of his time. He's already needed to play a couple of accurate moves. Here he finally blundered with 32. .. gxf5, when white suddenly had a chance: 33. g4! with near equality. In round 9 Grodzisk played the top seeded team, which had already clinched first place, and somehow crushed them 5.5-0.5. Vasik's game: The game got off to a rough start, move 16, black to move: Black has 16. .. Nxe5 since it defends d7. Not only did black miss that, he blundered with 16. .. Rd8, missing 17. Nxb7. It's not over, though: 17. .. N8a7 18. Nxd8 Rxd8 and white has stuff hanging. Vasik continued with 19. Bxc6 Nxc6 20. Bc3: The score is only +1.5, the bishops are really strong and the extra pawn is nearly useless. The computer wanted to play 19. Bd3 instead of 19. Bxc6 with a score of +2.5. Black did a good job creating chances after this, activating all of his pieces. Move 38, white to move: Black's last move was 37. .. Nf4, which was a "mistake" according to the computer. White now had only one move which keeps any advantage, but it wins. Vasik used 17 of his remaining 34 minutes here and calculated everything correctly, and played: 38. h4! Black has five dangerous replied (38. .. Nxg2, 38. .. Bxg2, 38. .. Nd3, 38. .. Rxf2 and 38. .. Bxh4). None work. I first tried to calculate this myself and eventually also got it right, but it's not easy. You need to see that white's king gets h2 as an escape, and also that his rook on f8 is defended by the bishop on c5 (which I overlooked on my first pass).

We won in rounds 7 & 8 by scores of 4:1 and 5:0, finished in second place, and promoted to the second league. Karol and I were the draws in round 7. The three contenders for promotion all won their matches by big margins despite opponents coming with more-or-less their normal teams. My games turned out to not be that important for advancing, but they were very interesting. Round 7 featured another crazy double-hallucination, my third of the tournament and easily the worst of the three. Round 7 game: Round 8 game: In round 7 I got my preparation, 6. c4 against the Kalashnikov: I prepared 6. c4 specifically for this match, previously I had always played 6. N1c3. He knew it well. 6. c4 is actually quite interesting. Back in our day everyone played 6. .. Be7 here with the idea of a quick .. Bg5, but white has a cool move which completely foils that: 7. N5c3! Now 7. .. Bg5 loses to 8. Bxg5 Qxg5 9. Nb5! (better than 9. Qxd6 with some compensation), and 7. .. a6 is met with 8. Nd5 followed by 9. N1c3. No better is 7. .. Be6 8. Nd5 with further ideas of N1c3, Be3 and Bb6, since 8. .. Bxd5 9. cxd5 allows 10. Bb5+ due to the lack of .. a7-a6. Black can play 7. .. Nf6 but then .. Be7-g5 is off the table. It's funny that nobody figured this out 30 years ago, I guess N5c3 unprovoked instead of N1c3 is just too counterintuitive. Nowadays people instead play this with .. g7-g6. My opponent had done that several times, using multiple move orders. It's what I focused on during my preparation, and it's what happened. There are a ton of transpositions so the important thing was to understand the plans. Position on move 10, black to move: He played 10. .. Be6 here, which technically took me out of my preparation, but that's not that important, the bishop belongs on e6, the engine scores don't budge, and the plans are the same. Move 17, black to move: This is a normal kind of position. The engine always wants to leave the knight on d4, dance around it for a long time, leave the c-file uncontested (even if black doubles on it), and slowly play for f2-f4, with scores around +0.6 throughout. Black could just castle here, 18. Bxh6 is a double-edged sword. Instead he played the sharper 17. .. Qh4. Move 19, white to move: Here I started to go slightly wrong. I saw the best 19. f5 and thought that white would be better, but slightly overestimated my position after 19. fxe5. He actually mentioned after the game that he would have preferred 19. f5 here. Move 21, white to move: I played 21. Rac1, which is also slightly inaccurate. In general white does not contest the c-file in these positions. I knew that from looking at a ton of such positions with the engine yesterday, but here I decided that he had neglected the c-file enough that a change of plan was warranted. He understood the position well, and correctly exploited my move with 21. .. Rxc1, giving up the c-file and betting everything on .. f7-f5. This is of course the benefit of playing the same opening over and over, you know the ideas really well. Move 23, white to move: Best was 23. h3. I considered that, it was my top move at first, but in the end I concluded that I wanted my pawns on dark squares and king on a light square because my unmatched bishop was a light-squared bishop, so I played 23. g3 Kh8 24. Kg2. The logic makes sense but it walks head-first into 24. .. f5, which he played: Here I went into the tank and used most of the rest of my time. I saw the best 25. Qd2 fxe4 26. Bxe4 Qg4 27. Qd3 but decided that black's initiative was too strong. There are ideas of 27. .. Ne2, 27. .. Nf3, 27. .. h5, even 27. .. Rf3. The move I most feared was 27. .. Nf3, which among other things ties white's rook to defending against .. Ne1+. I also didn't like losing the threat against h6, the variation feels "inefficient" for white. All of this can be defended though, the computer even says that white is a tiny bit better, and prefers 25. .. f4 with a small edge for black. Instead I played 25. Bxd4 fxe4 26. Bxe4, sacrificing my queen. The resulting position seemed like it shouldn't be that hard to draw, although of course black has whatever chances there are. In theory this was probably the correct conclusion, although it's also easy to slip up, as I did. If I had understood that 25. Qd2 was only a little bit for black I would of course have played it. Move 30, white to move: This is the initial endgame. White has a grip on the light squares, and his d-pawn and bishop pair are strong. White also doesn't mind bishop trades since then black won't have a good piece to blockade the d-pawn with. We did trade those bishops. The double-blunder is coming. Move 37, white to move: My idea was to carve out a defended square for the bishop on f5, and stand pat with bishop on f5, rook on d3, and king moving around. The plan is fine, but I sleepwalked here with the immediate 37. Bf5??, allowing 37. .. e4+ with scores of -5. Even the resulting position doesn't look that trivial but I am sure the engine is right. Somehow he missed it too and a few moves later the draw was agreed. None of the spectators mentioned it either. In round 8, I wasn't ready with the Gruenfeld yet, that's a project for the second league. I also was unable to prepare because the pairing was unexpected. We got another King's Indian. Position on move 11, white to move: I looked at this position with Vasik during his final, and posted about it here. The main line is 11. Ng5, then 11. .. Nxg3 is unsound as we discussed. Instead he played one of the other moves, 11. Nh4. I couldn't remember if I should play 11. .. Nf6, 11. .. Nf4 or 11. .. Qe8 12. Be2 Nf6, went with the latter. The engine prefers the immediate 11. .. Nf6, after my 11. .. Qe8 the score climbs a little bit higher, to +0.7, but that's as high as it would get. From the point of view of engine scores this was easily my best King's Indian since I started playing again. Move 13, black to move: I had the thematic 13. .. f4 14. gxf4 exf4 15. Bd4 f3!! 16. Nxf3 Nb4 17. Qd2 Bf5 with a slight edge for black. I looked at this of course but didn't see 15. .. f3, which pulls the knight off f5. Move 15, black to move: I had a long think here, trying to decide between 15. .. Nd7 and 15. .. f4. It was basically a coin flip but I chose correctly (15. .. Nd7). Move 18, black to move: Another critical position. I played the obvious sacrifice, 18. .. Nc5, fairly quickly, and liked my position, but surprisingly, it's inaccurate. The computer scores 18. .. Nc5 as +0.0, which isn't shocking. What's surprising is how strong the simple 18. .. Bxd4 is, then 19. Rxd4 Nc5 20. Qd1 Qe7!, hitting the knight on h4, then if 21. Nf3 Bf5 and black has huge pressure. I underestimated this. A key motif in this position is the battle for the f5 square, which is of course the point of 11. Nh4. Move 20, black to move: Here I had another long think. I smelled blood but white's position is surprisingly resilient. I ended up settling for the slightly better endgame after 20. .. a4 21. Nxa4 Qxa4 22. Qxa4 Rxa4. This is actually black's best. Including 20. .. Bxg4 21. hxg4 a4 21. Nxa4 Qxa4 etc is equally good, white cleans up his pawns but loses Rg1. Move 29, white to move: This is the basic endgame, which I squeezed out. There were a lot of inaccuracies but no big blunders, and also no really critical decisions. We both played most of it with 2-3 minutes on the clock + 30 second increment.

Random kid at the park vs Helen:

We expect a King's Indian with 5. h3 d6 6. Be3 today, then 6. .. e5 7. d5 Na6: This time I helped with the preparation extensively, it's very hard to analyze these positions with just the computer because so many natural lines are +1. I am still not sure what to think about all of that but the King's Indian has been played by top GMs for several generations - Kasparov, Fischer, Tal, and on and on - so I think it's perfectly ok to play it, the practical chances are obviously there. It's just a question of how to prepare. In a lot of cases it's necessary to distinguish between high computer scores which are based on concrete defensive resources which the opponent can prepare, and high computer scores which are based on positional evaluations which won't help the opponent in a practical game. There are two main lines. The first is 8. Nf3 Nh5 9. g3 f5 10. exf5 gxf5 11. Ng5: Here is a typical practical question: can black play the thematic 11. .. Nxg3? The computer of course doesn't like it: 12. fxg3 f4 13. Nxh7. Now 13. .. fxe3 14. Nxf8 Qxf8 looks dangerous but the computer shows multiple defenses, for example 15. Rd1 e4 16. Bg2 Nd3+ 17. Rxd3 and white emerges with a better position. This is too concrete, in this kind of variation I think you have to defer to the computer. More interesting is 13. .. Rf5 14. Bd2 fxg3 15. h4! Kxh7 16. Ne4 with another score of +1. Position: Is this an acceptable position to include in the repertoire? This seems like a borderline case. The other main line is 8. g4 Nc5 9. f3 a5: The human statistics are dominated by old games. The two main moves are 10. Qd2 and 10. Nge2. In both cases, black waits with .. h7-h5 until either white has played h3-h4 (and is threatening h4-h5), or until white has played both Nge2 and O-O-O (and is threatening Ng3). For example, after 10. Nge2, black does not need to play 10. .. h5 yet and instead plays 10. .. c6, because he can still respond to 11. Ng3 with 11. .. h5 12. gxh5 Nxh5 13. Nxh5 Qh4+. It's only after white moves the king from e1 that Ne2-g3 becomes a threat. So, the computer wants to play the immediate 10. h4, to force black to play 10. .. h5 before he has played .. c7-c6, then 10. .. h5 11. g5 Nfd7: The computer's idea is that it is now difficult for black to play .. c7-c6 because after .. cxd5 white can play Nxd5 instead of cxd5, since there is no knight on f6 any more. That's certainly valid, but white has paid a high price. Black has other ways to play on the queenside, including .. c7-c6, offering d6 if white dares to take it, otherwise .. Nb6 and .. cxd5. Also .. a7-a5-a4-a3. In the meantime white is stuck on the kingside, and if he puts his king there after all then black will play .. f7-f5. To my eyes this all looks perfectly fine for black, even good. The human players who almost all played 10. Qd2 and 10. Nge2 must have agreed, yet the scores are +1 and black's play leading up to this position is for the engine not correct. In this kind of situation I think you have go with the human evaluation (or not play the KID at all).

I did a lot of thinking about this yesterday, and talking to Vasik. We had kind of put the King's Indian and Accelerated Dragon into the same category, as old school openings which computers dislike but where arguably black had good practical chances. Actually, they are very different from each other. The Accelerated Dragon is perfectly playable even today, while the King's Indian really isn't, at least not in the old-school way we've been approaching it. First, engine scores. Just running the engine on the basic King's Indian position is already +0.6. The Accelerated Dragon is +0.4. To keep the scores that low black has to play all kinds of weird stuff in the King's Indian, while in the Accelerated Dragon those scores are reached through natural moves. Second, my results have been quite different. Since I started playing again in 2023, I've had 10 blacks, which break down as follows: 1x 1. g3 & 3. b3: equalized early 1x London: equalized early 4x Accelerated Dragon: 3x equalized by move 10, 1x equalized on move 17 and won on move 23 4x King's Indian: was never equal in any of them until move 27; 2x scores near +2, 2x scores near +1 Third, the number of white options. In the Accelerated Dragon, white's only option to get any kind of advantage is the Maroczy, and then black is the one with the next few options. Around 50% of players play 5. Nc3 (without c2-c4) when the score is already below +0.2. The computer would rather have white against the Najdorf (the "best" Sicilian) than 5. Nc3 against the Accelerated Dragon. Other deviations such as 1. e4 c5 2. Nf3 g6 3. c3 and 3. d4 cxd4 4. Qxd4 also score below +0.2. In the King's Indian, everything is +0.7 or so, including all kinds of systems I've never seen before. In the third league I twice faced pet KID variations I'd never seen before, both were +0.8 and quickly got to +1.5 after my "mistakes". Fourth, we have some new ideas in the Accelerated Dragon. In studying for Vasik's fifth round opponent, we realized that his Accelerated Dragon move order is really cool and forces white into an early hard decision. (I'll skip the details for now, feel free to ask.) Conclusions: 1. I'm keeping the Accelerated Dragon. No Berlin. 2. King's Indian, I am not sure yet, but something needs to change. There are three options: a. Nimzo/Queens-Indian. b. Gruenfeld (playing 2. .. g6 allows a bit more reuse of knowledge against side variations). c. "Funky" KID variations, preferred by engines to keep the scores low (of course). When top players play the KID these days, usually in blitz and rapid, they also play these. For example, against the system Vasik faced in round 6, the engine wants to play 1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 4. e4 d6 5. h3 Nbd7 6. Be3 e5: The point of delaying .. O-O is to meet 7. d5 with 7. .. h5. This stops g2-g4 of course, and after an eventual f2-f3 black will play .. h5-h4. The engine scores after 7. .. h5 are only +0.45. These types of "funky" computer variations exist throughout the KID.

Vasik won nicely in round 4, is now at 2/4: Both players have nearly identical repertoires and play those Maroczy positions from both colors. Vasik has played this kind of BB vs BN endgame with a white pawn on d5 several times with black, after Nd5 Qxd2 Bxd2 Bxd5 Pxd5 etc. He looked at that for several hours before the game and was making notes for himself for what to also do with black. 11. .. Qa5 was new, the engine wants to respond with Bd2, either on move 12 or on move 13. White didn't get much from the opening but the position slowly turned in his favor. Black then defended superbly before getting low on time and blowing it at exactly the moment where it finally seemed clear that the game would end in a draw. Tomorrow we expect a King's Indian, did a lot of prep and will do a bit more when Vasik wakes up. Will open a new topic.

Round 5 was epic, but this time not in a good way. Patryk couldn't make it and was replaced by Karol Maciaszek. The opponents, Skoczek Piescirogi, were at full strength. Kacper and Karol lost and Magda drew, all medium-length games, leaving Vasik and me both needing to win to salvage a drawn match. We were the two last games in the entire tournament and ended up having only a 10-minute break between rounds 5 and 6. I was winning for a long time in a long endgame, and ended up winning cleanly. At the end my opponent played 15 moves in a hopeless position to avoid putting the point on the board, he resigned as soon as Vasik and his opponent finished. Vasik's endgame went between winning and drawn several times, in the end his opponent got the draw and we lost the match. In round 6 we came back with a vengeance and destroyed one of the competitors for 2nd place, 4.5-0.5. Karol was the draw. Two teams qualify for the second league and small points (ie. margins of victory) are the now-very-important tiebreaker. We have excellent small points, having won 3 of the 6 matches with scores of 5-0, 4.5-0.5 and 4.5-0.5. Vasik's round 5 opening was interesting. I don't have the full scoresheet right now but the opponent always plays 1. b3, so we prepared the following gambit: 1. b3 e5 2. Bb2 Nc6 3. e3 d5 4. Bb5 Nge7 !?: The opponent thought for 20 minutes. Of course the engine wants to take the pawn with 5. Bxe5 a6 6. Bxc6+ Nxc6, then 7. Bg3 (7. Bb2 allows 7. .. Qg5). The opponent surely realized this, but he also realized that the position was sharp and that Vasik had prepared it inside out, so he backed down with 5. Nf3. The engine then wants to play 5. .. e4 with an equal position, but we had prepared a second gambit, 5. .. a6: Again the opponent thought for 20 minutes. Again he surely realized that the engine wants to take the pawn with 6. Bxc6+ Nxc6 7. Nxe5, but again black has 7. .. Qg5 with complications which Vasik had prepared. So, again, the opponent backed down with 6. Be2 when black has a nice position with a score of -0.6. The gamble turned out great, Vasik had a nice lead on the clock and a better position. Players are often cautious in team tournaments. My games were crazy. Round 5 was a tactical slugfest where I landed the last blow. Round 6 was a positional KID where I liked my position but the engine stubbornly shows +1 for white for a long time. Round 5 game: Round 6 game: The round 5 opening worked well. He was surprised by my 7. g4 and had prepared for 7. Qf3, which both Vasik and I had played in the Polish databases. He had looked at 7. g4 previously, but not recently, and after I played it he looked for a way to get out of the sharp main lines, which he actually did in a clever way. Round 5, move 10, black to move: He played the perfectly playable 10. .. d6, which I hadn't prepared for. Black doesn't usually play that move in these structures, it's more of a Najdorf/Scheveningen move than a Taimanov move, but it's ok. The game continued with 11. O-O-O Bb7 12. Kb1 Be7, position: Here I had my first big think and went with 13. Ne2, which was fine. Against 13. .. e5 I planned 14. Ng3 g6 15. g5 Nd7 16. h4. He lashed out with 13. .. d5, which is both logical and crazy, and also fine, position: White can of course play 14. e5 Nd7 15. Nd4 and 16. f4 with a calm position which is a little bit better for white. Black benefits in this variation from not having played .. h7-h6, which makes it easier for him to later play .. f7-f6 without weakening his kingside light squares, but he has also lost a tempo with .. d7-d6 and .. d6-d5. I went for the sharper 14. g5, which was fine, then 14. .. Nd7 15. Nd4 Qc7: I spent a ton of time here looking at stuff like 16. f4 dxe4 17. f5 e5 18. Nb3, that's +0.4. In the end I played the best move, 16. exd5 Bxd5: So far so good. Now of course I looked at 17. Nf5, which was best, and correctly thought that that was a little bit better for white after 17. .. Bxa2+ or 17. .. Rc8. Instead I went for more with 17. Bg2, keeping the tension: White plans f3-f4 when black still seems to have some problems to solve. I thought that castling was clearly better for white, more on that below, and expected 17. .. Rc8 18. f4 Bxg2 19. Qxg2 O-O when white can think about 20. f5, which is actually best. The score is +0.7. But, I completely missed 17. .. Nb6, defending d5 and threatening .. Nc4, then 18. f4 Nc4 is -3.0. I just didn't even consider this move. My opponent actually mentioned it in the one-minute discussion we had after the game, but he couldn't come to a conclusion and decided not to risk it. Fortunately after 17. .. Nb6 white at least has the crazy 18. Nf5 Nc4 19. Qd4 exf5 20. Bf4! with a score of +0.0. Moves other than 20. Bf4 are already -3.0. Instead, black went for 17. .. O-O: I played 18. Nf5 quickly, thinking that 18. .. Bc5 was losing to 19. Qc3 so black has to settle for something like 18. .. exf5 or 18. .. Rfe8 with a nice position for white. My opponent did play 18. .. Bc5, then 19. Qc3: There is actually a crazy defense here, 19. .. exf5 20. Rxd5 Rac8 21. Rxd7 Qxd7 22. Bxc5 b4, diagram: 23. Qd4 loses to 23. .. Qc7! 24. Bxf8 Qxc2+, and 23. Qe3 is equal after 23. .. Rfe8 24. Qf2 Rxc5! 25. Qxc5 Qd2 with dual threats of 26. .. Qxg2 and 26. .. Re1+. Needless to say neither of us saw any of this. My opponent was already low on time and went down with 19. .. f6 etc leading to an obviously terrible endgame. The scores are over +3 and quickly climbed to over +5 and stayed there for the rest of the game. I am not sure what exactly he missed in the end. The sixth round game was interesting as well. Here was the position on move 11, black to move: I thought that this was a decent King's Indian, e4 is weak and black has ideas of .. b7-b5 and .. Nd7-f6-g4/e4 which combine well with each other because white's knight and light-squared bishop are overtasked with defending against both. Kacper analyzed with us and also thought that black looked fine. It's +1 though. It seems that King's Indians are always +1 no matter what. One critical position came later, move 22, black to move: I used a lot of time and played 22. .. Bb5, inviting 23. Bg4. Then I looked at stuff like 23. .. Rf1+ 24. Rxf1 Bxf1 25. Rc2. It's dangerous for black of course to allow the light-squared bishop to live on e6, but it also preserves the light-squared bishops. It's easier in that position for black to trade the dark-squared bishops than for white to trade the light-square bishops. It's also nice that the black bishop defends a6 (while white's bishop symmetrically defends h3). I felt that black should somehow be fine because of this. I also didn't want to allow 23. c6 Bc8, cutting off my rooks from each other. The decision was correct, 22. .. Bb5 was best, but the white score is still +1, in exactly the variation I had analyzed (23. Bg4 Rf1+ 24. Rxf1 Bxf1 25. Rc2). Instead of 25. Rc2 he played 25. Qd2, position: Obviously he wanted to prevent 25. .. Bg5 but black has many ways to trade those bishops, for example .. Bh4-g3-f4, and now white's rook is also running low on squares. I was sure that black was doing well here, and the score does drop, but only to +0.0. The game continued 25. .. Rf8 26. Kc2 dxc5 when I was really sure that black is much better, but even here white is fine. The critical move was 27. Bh6: I planned 27. .. c4 but even then white can still play 28. Be2 with equality. Only after his 27. Qa5 was black actually better, the score then is -1.5 (and his 28. Bc5 is losing, that's -10). It's crazy how much and for how long the engine likes white in these kinds of positions.

Round 3 is black vs Piotr Zareba. He plays a collection of slightly offbeat systems which are much less popular than their computer evaluations. One of these is 1. b3 e5 2. Bb2 Nc6 3. c4, against which we've prepared a pair of gambits which are similar to what we prepared against 3. e3 for round 5 of the third league. Vasik will continue with 3. .. Nge7 4. Nf3 d5: Now: A. 5. Nxe5 Nxe5 6. Bxe5 d4 followed by 7. .. Nc6. This is why the knight goes to e7 rather than f6 on move 3. B. 5. cxd5 Nxd5 6. Nxe5 Nxe5 7. Bxe5 Qe7 8. Bb2 Nb4 followed by .. Bf5 and .. O-O-O. Who knew that 1. b3 could be so sharp. Here is another cool position, 1. e4 c5 2. Nf3 g6 3. d4 cxd4 4. Nxd4 Nc6 5. Nc3 Bg7 6. Nb3 Nf6 7. Be3 O-O 8. Qd2 d6 9. f3 Be6 10. O-O-O: How should black continue? In general, white's plan is Kb1, h2-h4-h5, g2-g4, and Bh6. He usually wants to start with Kb1 because it's the most flexible of those moves. Bh6 allows .. Bxh6, .. Qb6, .. Rfc8 and .. Rxc3, g2-g4 weakens f3 and allows sacrifices on f3/g4/e4, and h2-h4 allows .. h7-h5. For example, the obvious 10. .. Ne5 11. Kb1 is pretty annoying. What move should black himself start with so that white needs to start with something other than 11. Kb1? The answer is 10. .. Qc7!, then 11. Kb1 runs into 11. .. d5!!, ie. 12. exd5 Nxd5 13. Nxd5 Bxd5 14. Qxd5 Rfd8. And here is a vicious combination which demonstrates the importance of Kb1. Position after 10. .. Qc7 11. h4 Rfc8 12. Nd4 Nxd4 13. Bxd4, black to move: Solution: 13. .. Bxa2! 14. Nxa2 Bh6.

Round 2 went very well: The preparation was again a problem. The guy had never played the Najdorf in any of his 50+ black public games. He obviously targeted Vasik's 6. Bg5. Vasik was aware of this, of course, but didn't have a backup line. Arguably, even then it's a mistake to play into the opponent's prep, and would be better to play something completely random. The opponent prepared it really well, until 11. O-O-O. Fortunately, Vasik played right through the guy's preparation, making only one small inaccuracy (8. Qe2) in that phase, and once the guy was on his own Vasik crushed him.

We won both matches yesterday. Round 3 was 4.5-0.5, I was the draw. Round 4 was epic. We played the third-seeded team, both teams were at full strength. Kacper won quickly against the Elephant Gambit on board 2. Patryk and Vasik made 30-move draws which never went far from equal on boards 3 and 4. Magda, who is our leading scorer the last two years, quickly got in trouble on the women's board. My opponent and I played most of our game expecting that it would break a 2-2 tie, which it did. It was a crazy game, I sacked a rook and then played most of the game down a full exchange, but actually it was pretty well played. Round 3: Round 4: Round 3, move 12, white to move: I played 12. cxd5 quickly, thinking that black has several options and I would figure out the complications of 12. .. cxd5 on the next move, but that was really dumb. 12. .. cxd5 is forced, of course, and 12. cxd5 is a big mistake. I then spent an hour on move 13, trying to sort through 13. Qc8+ and 13. Nb3, and looking at all kinds of crazy stuff like b2-b4 and Ne4, with or without 13. Qc8+. I also spent a ton of time looking at the endgame we got, trying to figure out if black's advantage was symbolic or real. Round 3, move 19, black to move: This is the basic endgame. We both thought that white should be able to draw it pretty easily, and in fact I got a draw offer a few moves later. The engine shows a weird and cool way to play for a win. Black's "problem" is that he can't trade both pairs of rooks, then it's an easier draw. He also can't allow white to dominate the c-file with both rooks. The solution is simple: trade one pair of rooks by playing 19. .. a5 20. Rac1 Rac8 21. a4 Nb4, now 22. Rxc8 Rxc8 23. Rxc8 Kxc8 allows 24. .. Nd3 so white should start with 22. Kf1, then 22. .. Rxc3 23. Rxc3 and now the best plan is the mind-boggling 23. .. Nfe7 with the idea of .. Nec6, .. Na2 and .. Ncb4. White then can't stay on the c-file and has to put the rook on d2. After that black has ideas of .. h5-h4-h3, and also .. Rc8 and .. Rc4 when a4 is hanging. The score is -1.0. Instead, black played the safe 19. .. g6, but then after 20. Rac1 Rhc8 21. Kf1 a5 22. a4 white is better off since 22. .. Nb4 can be met with 23. Rxc8 Rxc8 24. Rxc8 Kxc8 25. Ke2, controlling d3. Basically, 19. .. g6 gave white the tempo he needed to control d3. The engine still gives -0.5 after 19. .. g6 but it's a lot less dangerous. Round 4, move 14, black to move: I played the obvious (to me) 14. .. Nc6, already with the upcoming sacrifice in mind. It's not completely accurate though, better was 14. .. Bb7 with a small edge for black. Round 4, one move later, black to move: I had the fallback option of 15. .. Qb4+ and 16. .. Qd6 but went for 15. .. Nd4. The rest of the team was already done and watching, Vasik said he knew I would play this. Round 4, two moves after that, black to move: Another critical decision. I originally planned to play 17. .. Nc2+ but now I dug deeper into it and started to dislike it. After 17. .. Nc2+ 18. Kf2 Nxe3 19. Qxe3 Qxb2+ white can play 20. Kg3, then after 20. .. Rxa8 21. Rab1 Qc2 22. Rhc1 Qd3 white has to trade queens and activate black's rooks, although e4 is also weak. Instead I went for the more complicated 17. .. Rxa8. The computer says both are around +1.0. Round 4, one move after that: My 18. .. Qb7 was inaccurate, computer wants to play 18. .. Bb7. Round 4, move 22, black to move: This was my last big think of the game. I had seen 22. Qa5 and thought that I had several options, for example 22. .. Nf5 23. Qd8+ Bf8 24. Bc5 Qe2+ 25. Kg1 Nd6, but now I started to realize that white has a pretty big initiative, and the computer agrees. Instead, I switched to a move I hadn't even considered until this position happened, and it's the only move: 22. .. Bb7. White wins back the pawn and is up a full exchange after 23. Qd8+ Bf8 24. Bxd4 etc, but he has to give up his dark squared bishop and open the position for black's bishops. Round 4, move 26, black to move: This is the basic endgame, which went on for a while. White is up a full exchange but the bishops are really strong and there is no simplification for white in sight. The engine scores are around +1.0 to +1.5. My opponent was also quite nervous, playing several moves with under 10 seconds and constantly looking at the clock, while I always kept 2 minutes. For once I was the more practical player. Around this point Vasik bet Kacper that I would win, giving draw odds. We eventually got into a queen endgame. I was a little better, white's pawns were weak and I could pressure them with my king, although it's not much and the engine just gives +0.0. Even after white sacked or blundered one of his pawns, not sure which, the scores remained +0.0. There was one double blunder, move 60, black to play: 60. .. Qf6+ was winning. I saw it but didn't calculate it correctly, 61. Qxf6+ Kxf6 62. Kd5 Ke7 63. Kc6 f4 64. Kb7 f3 65. c6 f2 etc leads to another queen endgame which is winning, while after my 60. .. Qd7+ it's still a draw. Neither is obvious to me. After 60. .. Qd7+ he blundered a second time with 61. Ke5 allowing 61. .. Qc7+ 62. Kd5 Qd8+ with an obviously winning king and pawn endgame. Rounds 5&6 are in four weeks. In round 5 we have white against the second-seeded team. We'll be at full strength.

Hi everyone. I hope you all had a nice holidays. The Polish third league has started again, we are trying to promote to the second league after missing last year. Off to a good start, 2 match wins. First one 5-0 , second 3-2. Here are my games: The second game continued for another 30 moves, it was the last game and we were up 2.5-1.5 so my opponent had to avoid repetitions despite being nominally worse. In a normal game I'd try to win it with white. Game 1, move 17, black to move: I played 17. .. f5, which was good and correct. The point is that 18. gxf5 Nxf5 19. Bxd5 runs into 19. .. Nxd4. Three moves later, move 20, black to move: I correctly avoided 20. .. Nxd4, which is good for white. There are actually two refutations, I found one. Four moves after that, move 24, black to move: I played 24. .. Nxg2 with the obvious idea of 25. .. Bf5 and 26. .. Be4+. This bishop looks like a monster. But, the engine can apparently defend this, score is only -0.5. Better was a move I hadn't even considered, 24. .. Rf3. Game 2, move 12, white to move: I thought for 40 minutes and played 12. a3, which is a mistake. I wanted to play 12. h4 but didn't like 12. .. h5. Was calculating stuff like 13. gxh5 Rxh5 14. Nxd5 exd5 15. e6 0-0-0 16. exd7+ Rxd7 and didn't like it. However, there is a very simple answer to 12. .. h5: 13. gxh5 Rxh5 14. Bxg6 fxg6 15. Bg5 with +1.5. The rook is bad, black can't castle, the bishop can't be displaced with .. f7-f6, and d4 is defended. Sometimes it's so simple. But, 12. h4 is a mistake because of 12. .. Nxd4. My 12. a3 is also a mistake because of 12. .. Nxd4 with a score of 0.0. I wanted to play 13. Nxd4 Qxd4 14. Be3 Qxe5 15. Bc5, now 15. .. Qc7 or 15. .. Qf4 get hit with 16. Nxd5. But, black has 15. .. Qb8 with an advantage, then 16. Nxd5 is met with 16. .. Bxc5 and 16. Bxg6 hxg6 17. Qxd5 hangs h2. White should instead play 14. Nb5 Qb6 15. Be3 Qd8 with a score of +0.0. The engine just wants to play 12. Be3 to defend d4, then white is a bit better. Later we had a crazy triple-blunder sequence. Move 32, white to play: We're down to a few minutes each plus 30 second increment. I've been better the whole game but here I went blind and played 32. Ke3, walking into 32. .. Nxd4, which is -1.0 after 33. Bd1. My opponent not only missed that but blundered himself with 32. .. f6, allowing 33. Ng2 Bg5+ 34. f4. To top it all off, I missed that too, and played 33. Nf3 Bg5+ 34. Nxg5. The rest of the game was pretty decent actually.

There is another element of teaching openings via tactics, which is that sometimes you have a motif which works in some positions but doesn't work in others. I am thinking about the following organizational structure, will try it with Simon today: 1. Pick two similar opening positions, one where the motif in question works and one where it doesn't, and quiz them, without giving hints or answers. 2. Then quiz the downstream motifs. 3. Then go back and repeat question #1, and derive heuristics for when the upstream motif works and when it doesn't. Here is the example for today, London as black. Our system is 1. .. Nf6, 2. .. g6, 3. .. Bg7, 4. .. 0-0, 5. .. d6, 6. .. Nc6, then an immediate 7. .. e5 when it works and 7. .. Nfd7 followed by 8. .. e5 when the immediate 7. .. e5 doesn't work. Step 1: Quiz two similar positions. The question is whether black can play the immediate  7. .. e5 in the above positons, or should settle for 7. .. Nfd7 and then 8. ..e5. At this point I don't expect any competent answer, and only want to hear some initial theories about what might make 7. .. e5 more or less attractive. 2. Step 2: Quiz the downstream mofits: In the above positions, can white play 1. exd6? These positions need to be worked out correctly. I'll give hints until Simon gets it. The answer is that in the second one 1. exd6 Bxb2 2. Nbd2 is good for white while in the first one 1. exd6 loses to 1. .. Bxb2 2. Nbd2 Nc3. Steps 3 & 4: Now the first questions should be easier. Black can play 7. .. e5 when b2 is hanging and he can respond to 8. dxe5 with 8. .. Ne4. Tactical motifs are the core of so much of chess, and there is a finite number of tactical structures.

Simon is playing a local tournament with one round per day, it's part of the chess camp here. The games are not that important, nothing is at stake except ratings and categories. I'm preparing him for the games, and have been experimenting with a different way to structure the preparation. Instead of showing variations, we focus on tactics. I pick out the tactical ideas from his likely opening and ask him to solve them, one by one. He got one yesterday in his game, and then worked it out at the board, correctly. White to play: The position was reached via 1. e4 d5 2. exd5 Qxd5 3. Nc3 Qd8 4. Nf3 Nf6 5. Bc4 Bf5 6. Ne5 e6 7. Qf3 c6. The opening we prepared for today was the 3. .. f5 Spanish, from the black side. Here are some tactical themes, all black to move: The last one is really tough, obviously. Black should not allow 10. .. Kxd7 11. Qb5+, but what happens after 10. .. Nxd7 11. Qxe4+ Kd8 12. Qxb7? Of course Simon didn't find that, I probably wouldn't find it myself, but it's still maybe better to show it as a tactical quiz rather than as a variation to memorize. There are also tactics for the opponent, those are just as important. For example, white to play: I've never heard of a trainer using this method to prepare for "serious" games, but I wonder if this could be everything you do with the kid. Instead of telling him what to play, which is extremely prone to getting mixed up, you go over the tactics and then work out the repertoire backwards.

The Grodzisk team finished the first league. Unfortunately they dropped to the second league. It was terrible luck, four times they lost 3.5:2.5. Kacper was on site for the first five rounds, he and Jacek prepared the kids for those rounds, then Kacper had to return home so I prepared Vasik for rounds 6-9 by phone and email. Vasik had several interesting games. Here is round 7: Here is the position on move 8, white to move: We prepared for 8. e5 and 8. Kh1 and mostly focused on 8. a3, which forces black to spend two tempos to play .. b5-b4 (ie. .. Rb8 and .. a7-a5) and gives white an open a-file when black does play it. After 8. a3 we planned 8. .. Rb8, then we briefly talked about 9. Be3, for maybe a minute. I told Vasik that 9. .. d6 is then perfectly fine but he can always give up the c-pawn to get the dark squared bishop with 9. .. Ng4 10. Bxc5 d6 11. Bf2 Nxf2 12. Rxf2 since white's dark squares are so weak. We didn't really go further, I just threw out a few obvious dark-squared ideas such as .. Bh6 and .. e5, .. Qb6, .. a7-a5 and .. b5-b4, etc. I had the engine running as we talked to make sure that what I was saying wasn't complete nonsense. Instead, the opponent played the immediate 8. Be3, which is just a less logical version of 8. a3 Rb8 9. Be3. The question is if the same black idea works. It does, and Vasik used less than two minutes to play it. He figured, correctly, that at worst it can't be too bad. He also quickly settled on .. Bh6 and .. e7-e5 as the follow-up since he knew I had vetted it with the engine in the 8. a3 Rb8 9. Be3 position. It's nice to basically cover multiple variations with one shot. This was just an accident, I certainly didn't have it in mind, but coaching should probably somehow be optimized around this concept.

Here is a funny position, white to play: Who stands better and why? The variation is: 1. d4 Nf6 2. c4 g6 3. Nf3 Bg7 4. g3 O-O 5. Bg2 d5 6. O-O dxc4 7. Na3 c5 8. dxc5 c3 9. Nd4 Na6 10. c6 Nd5

Vasik finished the under-14 final a few days ago. It was pretty rough, 3.5/9. I was his trainer, it's probably time to find him a professional trainer. His roommate was preparing over three hours per game and had a GM trainer. We were doing 90 minutes. Here are Vasik's best and worst moves as puzzles, both interesting. Worst move: It was the last round after a tough week. We prepared 13. Qe2 but he got it mixed up and played 13. Qd2 instead, which is much worse than you might guess. How? Game: Best move: Black to play. Game: I talked to Bartosz Socko about everything after the tournament, he's a professional trainer. Showed him several games and the preparation I was sending Vasik, including the last round game. He laughed of course at how quickly it went downhill but said that it's normal. His students are in the 15-18 age group, that's his specialty, but even then they are often mixing things up.