Why you should not trust chess database statistics

Have a look at this position, which arises in Panov Attack in the Caro-Kann, or from some lines of Queen’s Gambit Declined.

1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 e6 6. Nf3 Be7 7. cxd5 Nxd5 8.
Bd3 Nc6 9. O-O O-O 10. Re1 and now Black can play Qa5, which is a very rare sideline

Say, the database will tell you that 4 games have been played with this move, and black scored 50% with Qa5. The move itself sure does not look quite right, but Viktor Korchnoi himself played it, and the score seems not too bad. The 50% however is completely deceptive, have a look at another position from Brodsky – Maiorov, one of the games played in this line.

  White to move

White (a Grandmaster) is completely winning, but he played the completely inexplicable 47. Rd2?? dropping the rook on c4, and resigned a couple of moves later. So that 50% should have really been 75% in White’s favour! 4 games is really too small of a pool to rely on the numbers, so you should really look at objective value of Qa5 instead of relying on rather meaningless percentage from a database. I had failed to do that, so Black’s other loss is mine – from my 2002 game against Stephen Glinert! The title of the post is of course not fully true – sometimes you should check how well each side is scoring in a line before playing it, but it’s more important to understand the meaning of moves while building up your opening repertoire, especially if you rely on sidelines like Qa5.

Bishop Sacrifice in Panov Attack (Caro-Kann)

One of the main points of playing blitz on the internet is to go over finished games, and draw certain conclusions from them. I learned a fair bit from the game I played online today – something that can be useful in tournament games I play later, because the pawn structure and combination ideas are typical for the opening that I play.
This position occurred in this game
White to move
Black just played 11… Nde7, instead of the more standard 11… Nce7. I continued with 12. Be3 and got a position with isolated Queen’s Pawn (IQP) that I like, but putting the bishop on e3 did feel a bit passive.
After the game, I looked up this position in Karpov and Podgaets’ book on the Panov attack since the move Black played took me a bit by surprise. Turns out White has a nice way to exploit the fact that Black reduced his control over g5 square, and play 12.Bg5!? The key point is that if 12…Bxg5, then White can strike with a typical sacrifice on h7 with 13. Bxh7+!?:
White does not win on a spot, but the book shows that his position is better. This is the kind of guidance I would expect from an opening book, so I recommend it for its thoroughness! I took a longer way around, but in the end did create pressure against Black king and won (who said that analysing blitz games is a waste of time?!)
Replay the game in the viewer:

Dilworth attack – positional pawn sacrifice

Aseev – Mikhalevski , 1999

Black to move

28 … d4! was the best move in this position. White was threatening to fully consolidate on the dark squares (maybe bring the knight over to c5), so black had to create counter play as soon as possible.

After a few moves this position arose:

Now Black king will threaten to invade on light squares, so the White knight won’t have time for Ne4-c5xa6. Black got reasonable counter play and drew the game.

Review the game in the viewer:

Two Knights defence – bishop sacrifice

Leonidov – Jiganchine, Moscow, 1997

Black to move

Yet another silly nostalgic post: When I was a junior, I was spending a lot of time analyzing (without a computer!) positions with initiative for a pawn. My coach wisely set me up with an opening repertoire that was based on home-grown gambits. There was no books explaining how to play in these positions, just me, and the wooden board. This game was one of those kind of games where you feel that your homework paid off, and you have a good sense for positions that you get. I sacrificed the bishop on f2, won a nice game, and kept a good memory of it. But now computer puts doubts into my head – was the sacrifice really winning? Or was it just enough for a draw? Either way, in the game White got afraid to take the bishop and got mated very quickly:

12… Bxf2+! 12. Ke2 Nh5 13. Ne4 Qd4 14. d3 Ng3+ 15. Nxg3 Bxg3 16. Qg1 Rf2+ 17. Kd1

Black to move
17… Bg4+ 18. hxg4 Qxg4+ 19. Ke1 Qe2#

Replay the game in the viewer:

Understanding Chess Classics - Attack by Jorge Pelikan in Taimanov Sicilian

White to move

Jorge Pelikan is a player credited for popularizing the opening system that later transformed into the Sveshnikov variation. But in this game he demonstrates his attacking skills on the White side of Open Sicilian. I really like attacks like this one, that come from superior development in open positions. White’s pieces are very active, and Black also has some weaknesses on the dark squares, in particularly on d6 (typical for Taimanov Sicilian). So in the position on the diagram White uncorked the following:

15. Nf5 !! Be5 (15... exf5 16. Bd6 Qa5 17. Rxb7 Qxa2 18. Bd5 Qa4 19. Rxd7 Kh8
20. Rxf7 Qg4 21. Rd7 Qxd1 22. Rxd1 a5 23. Bc7 Ra7 24. Bxc6 Raxc7 25. Rxc7 Rxc7
26. Rd8+ Bf8 27. Rxf8+ Kg7 28. Rxf5 Rxc6 29. Rxa5 Rxc3 30. Ra2 +-) 16. Nh6+
Kg7 17. Ng4 d5 18. c4 Bc3 19. cxd5 Nd4 20. Ne3 f5 21. Bd6!
another piece put en-prise
Qxd6 22. Rxb7+ Kh8 23. dxe6 Qxe6 24. Nd5 Qe5 25.
Nb6 Rab8 26. Nd7 Qf4 27. Rxb8 Rxb8 28. Nxb8 Qxb8 29. Qd3 Qb2 30. Qe3 Kg7 31.
Qg5+ Kf8 32. Qf6+ Ke8 33. Re1+ Bxe1 34. Bc6+ 1-0
Replay the game below in the viewer:

Trapped Rook

White to move

Here is a chance I missed in one of my blitz games. The position is fairly typical for Accelerated Dragon and Maroczy Bind. Surprisingly White could win by simply attacking the rook:

25. Bb5! Ra8 26. Qc6 Qe3+ 27. Kh1 Rd828. Qc7 Qb3 29. Rg1 Ra8 30. Bc6 and amazingly the rook is trapped!!.

Black could pin the bishop with by playing Qc2, but then White just picks up Black pawns one by one.

Becoming a chess master

What is involved in becoming a chess master, and why is that so hard for most people? What differentiates a master from an expert, and what distinguishes a master from a grandmaster? One experienced player told me that there is little difference between his understanding of the game, and that of Kevin Spraggett, but Spraggett, being an active professional player is a lot faster at making correct decisions, making speed at evaluating positions and calculating variations is the main factor. True or false, the gist of this statement is that there is nothing magical in being a grandmaster. You can "simply" be slightly better than weaker players at every component of the game, and that already can boost your rating by a couple of hundred points.

But of course, it's not just about speed. It's also about the depth of your knowledge and preparation. Let me give you my personal example, without hopefully not turning it too much into a rant. In Canada, a master is considered to be any player rated over 2200 CFC. I crossed that mark around 10 years ago, and remained slightly above it since then. To become a FIDE master, one has to be rated 2300. So why after 10 years of playing and studying – I am nowhere close to getting to 2300?
1) One obvious reason is, of course, that I have not studied for these years as much as I did 10 years ago. Simple metrics show that I started to play a lot less after I finished high school. So lack of practice and effort surely is to blame. You need to play regularly in tournaments to improve.
Things have kind of slowed down for me after year 2002…
2) A related reason for slowed down progress is age! Arguably, once you turn 20, your brain does not absorb information as quickly. Try to improve as much as you can while you are young!
Don't wait till you're too old!
3) All these personal reasons aside, there is one more big reason why I have not gotten to that 2300 mark yet, same reason as why there are several thousands IM's in the world, but probably less than a thousand GM's. The further you improve at anything, the more effort you have to put into it.
Are you ready to work hard?
Opening repertoire is one good example: to get to 2200, it was sufficient for me to have one defence to 1.e4, and 1.d4 for Black. For White, I knew one simple anti-sicilian system, so that I could get out of the opening and outplay 1800 rated players. Against 1…e5, I knew the difference between the Zaitsev and the Breyer in the Spanish opening, but my knowledge stopped there.
As I started to play players 2200 and over, I realized that I need to be able to surprise them with my choice of opening, or else their home preparation would put me into a worse position in every game. I also realized that I need to play more complicated systems than that simple anti-sicilian, that is never supposed to give White any advantage. So partially out necessity, partially out of boredom, I expanded my repertoire for both Black and White. The amount of opening information I need to be familiar with now is about 10 times larger than it was in 1999. I also need to know each variation a lot deeper. A book like MCO is no longer sufficient, it just scratches the surface.
The same applies to the middlegame as well – by virtue of expanding my repertoire, I also need to be familiar with a lot more middlegame structures than I used to. For example, I now need to play IQP not only for White, but also for Black! (yes, now I have to face that same anti-sicilian that I used to play myself as White).
4) You also need a mentor. Someone not necessarily to coach you weekly through rook endgames, but to observe your progress and make suggestions from the point of view of more experience. I was very fortunate to have this kind of mentoring from a couple of strong players between 1996 and 2002 and that helped me with my progress tremendously. Remarks about positional subtleties or flaws in your understanding you hear from another chess player stick to your mind a lot better than something printed in a book you flip through at local Chapters book store.

Of course, one could just be _good_ at chess, and become 2300, but unfortunately I don't know how to magically do that!

The King Walk – Chigorin – Caro, 1898

Black to move

Try to predict the outcome of this chess game! Amazingly, 10 moves later Black got mated with a sacrifice that many of you probably have seen while studying tactics. After the following moves 25. …Qg2+ 26. Kc4 b5+ 27. Kd3 Qf3+ 28. Kc2 Qf2+ 29. Kb3 Rc8 30. Rc2 Qxf4 31. Kb2 Na5 32. Ka1 Qc4 33. e6 Nc6 34. Qd1 h5 35. Rg1 Rh7 the players got to this position:

 White to move

36. Rg7!! ends the game on a spot. Notice how the white king now is safely hidden.

Replay the full game here. Below is the raw pgn that you can copy and paste into a Chessbase Light board window.

[Event "Vienna"]
[Site "Vienna"]
[Date "1898.??.??"]
[EventDate "?"]
[Round "?"]
[Result "1-0"]
[White "Mikhail Chigorin"]
[Black "Horatio Caro"]
[ECO "C29"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "71"]

1. e4 e5 2. Nc3 Nf6 3. f4 d5 4. d3 Bb4 5. fxe5 Nxe4 6. dxe4
Qh4+ 7. Ke2 Bxc3 8. bxc3 Bg4+ 9. Nf3 dxe4 10. Qd4 Bh5 11. Ke3
Bxf3 12. Bb5+ c6 13. gxf3 Qh6+ 14. Kxe4 Qg6+ 15. Ke3 cxb5
16. Ba3 Nc6 17. Qd5 Qxc2 18. Rac1 Qf5 19. Rhe1 Rd8 20. Qxb5 a6
21. Qb1 Qg5+ 22. f4 Qg2 23. Bd6 Qh3+ 24. Ke4 f5+ 25. Kd5 Qg2+
26. Kc4 b5+ 27. Kd3 Qf3+ 28. Kc2 Qf2+ 29. Kb3 Rc8 30. Rc2 Qxf4
31. Kb2 Na5 32. Ka1 Qc4 33. e6 Nc6 34. Qd1 h5 35. Rg1 Rh7
36. Rxg7 1-0

Chess Endgame Tablebases - online

As I was analysing my endgame against Jack Yoos, from the BC Closed Championship 2000, I came across this position below (Black to move).

 
Is it a draw? Or a win for Black? A regular engine thinks it’s –2.00, so that does not help. But I am down to 5 pieces, so this position has been already “fully” pre-analyzed, and as the online chess tablebase server shows – this is a draw! You just paste the FEN for the position into a form on the webpage, and it tells you exactly what’s happening here. A draw in this position means I have to keep looking for other ways for Black to win this endgame. Poking around earlier in the analysis, I see that I can reach an improved version of this position. Obviously – here Black is winning:

10 tips for Analysing your Chess Games

One of the major ways of studying chess in our Russian “chess school” in the nineties was analyzing your own games. That got me into believing that this is the best way to improve your game. I started doing it without computer, on a piece of paper, but obviously over time tried to use databases as much as possible for maintaining and updating a collection of my games. I also helped several students of mine to learn how to analyze their games in effective ways, so here are a few suggestions that you might find useful:

1. Maintain a database of all your games. I keep several databases: my games with standard time controls, rapid time controls, internet games
2. As soon as possible after the game has finished – put down thoughts you had during the game. That will help you later to remember and understand the reasons for your mistakes.
3. Let the computer engine run through the game in blunder check mode – that way you’ll know immediately about the major blunders you and your opponent made
4. Identify the critical moments of the game. How many times does evaluation of position change, and advantage shifts from one side to the other?
5. Analyse the opening, update your opening repertoire if necessary. Evaluate the position after the opening, to decide whether your openings need “repair”
6. Do not just analyse in terms of variations. Give verbal evaluations of critical positions. If white is better – say why. That helps you to better understand the true meaning of each position. That also makes you stop looking at the computer evaluation and think on your own for a few seconds.
7. When you’re done analyzing – summarize your the game. Why did the game end the way it did? Where was it decided – opening, endgame, tactical blunder?
8. Over time – look at the trends in your games. Do you lose more points in openings or in endgames? Is there anything you can study in particular to improve those trends?
9. Go back to your games, even years after they were played. I do that just to practice my analytical skills, and very often I find surprising how much new little details I can discover (e.g. the endgame I thought was drawn – is actually winning, etc).
10. As already mentioned – don’t fully rely on the computer engine. Try to find moves and ideas on your own, and only then let the engine give you hints. It is ok to guide the engine, but make sure you’re still the driver.

Chess Position Trainer – test your opening knowledge!

I have written a fair bit about how you’d want to be careful about choosing the openings that you play and opening preparation in general. But how do you actually remember the multitude of lines that constitute your repertoire? If you carefully analyzed certain openings, and your repertoire consists of your original ideas – memorization likely won't be an issue. But if you never played a line before, and barely can think of a couple of plans – as soon as a position on move 10 comes up on the board, you'll realize pretty soon that:

• The devil is in the details – understanding positional ideas is great, but you still need to play specific moves to implement those ideas
• Move order matters
• some positions look similar, but the move you remembered and thought should be played in this particular position is actually good in a different sub-variation, but is inappropriate in this particular case

To test and refine my understanding of an opening – I usually play blitz games on ICC, look at them later and see whether I deviated from my own repertoire, or whether there is actually a gap in my repertoire that this game revealed. While practice games are super-useful, that type of training has some limitations:

• you have no control over what opponents will throw at you (you really did not want to test your knowledge of 1…b6 !)
• in the rush of a blitz game - you may not focus too much on playing the move that you had prepared (I already spent 30 seconds on the opening, I’d better move something quickly)
• once the game has ended - you also need to have the stamina to go back and compare your moves against what you had planned in the preparation

So there is a need for a special tool to test your opening knowledge! I have always admired the developer who put up a ton of effort into creating a free tool that lets you manage your opening repertoire and then practice it in a training mode – Chess Position Trainer. It always felt like a tool I would want to write myself if I had the skills and time. As it often happens with free software - the author had to fill the gap that no other free or commercial application would cover sufficiently for him.

However I have not used it for a while as it seemed that all my data was already in pgn and chessbase format, and my opening repertoire existed outside of Chess Position Trainer ecosystem. What was even worse – my repertoire had too many gaps (it still does), so I spent the last year going through books, databases and my old games, deciding what I should play against such and such line. Today I imported the opening repertoire I had created with Chessbase Light into Chess position Trainer, and here we go – I can practice my openings again!

In the training mode – the UI looks like this:

Training mode

A few points I like about its training mode:

• games are thrown at you relatively randomly
• Your progress is tracked (I have not dug into the details)
• The board is large and I like its simple look – often underestimated but very important, but some $100 packages got that part wrong.
• You’re told immediately whether you made a correct move (“correct” means – the move you yourself defined to be in your repertoire)

Sometimes it keeps throwing the same line at you until you actually know enter all the correct moves:

Training Mode is a pretty fundamental part of Chess Position Trainer – you can tell that from looking at training options (that’s a lot of Tabs!):

I would agree with what Stefan (the developer) wrote in this blog post – "this time I have already at hand a chess tool to study my material."

Outside of the Opening Training mode – the UI looks like this. This is one of the most well designed user interfaces for a chess program I’ve seen so far. It sounds silly, but buttons are large and convenient to press. It feels that the author actually uses this program himself, so he tried to make things easy for himself – and therefore, for any other chess player using his product. At the core of the software - the focus is as much on positions as it is on the moves, which makes sense for opening preparation.

Attacking the king – rooks vs. bishops

I played a rare 15 minute game online today, and managed to get to an interesting position where bishops and rooks were tangled up in a curious way.

DDT3000 – Fundie, ICC, August 1, 2009

White to move

White is better because he controls the ‘e’ file, but here I was not sure what to do, but accidentally played the best move:
21.Ne4 This exchanges the main defender of Black’s king. Nxe4 22. Bxe4 Bg7 23. h5! And here Black started to collapse in time trouble f5 24. hxg6+

(Apparently also very good was 24. Bxf5! Bxf5 25. Re7
White’s attack is very strong: Qb6 26.
Bxh6! Rhg8 27. hxg6+ Bxg6 28. Bxg7 Rxg7 29. Rxg7+ Kxg7 30. Re7+ Kf6 31. Re6+
Kg7 32. Qg5 +-)

24... Kxg6 25. Bf3 Rf8 26. Re7 Bd7 27. R1e6+ Bf6

White to move
I thought this position was kind of curious – both bishops are attacking my rooks, but they are both pinned! Since none of my rooks can be captured – I can simply continue attack against the king.

28. Qe2! (28.Bh5+ was apparently mating immediately Kxh5 29. Rg7 Bxg7 30. Qe2#) 28... Qd8 29. Bh5#  1-0