Upgrade to Chess.com Premium!

Openings Chessmetrics

Chessmetrics Statistics Chess Opennings

 

Openings Chessmetrics

by Jim Fox, Ph.D.

 

 



Most chess players have their favourite opening.  Some have several that they know well, and advice abounds on which one is best. This is really true when it comes to black openings.  What is the best defence against “e4”? How should you counter the English opening?

 

Many openings bear the names of self-styled masters who created them. These include Ruy Lopez, Van Geet, Bird, Zukertort, to name but a few.

 

Chess, and the study of the probabilities of chess (chessmetrics) make up the backbone of chess theory.   Chess theory, you may think, is all about numbers, but it is not.  

 

Because of the billions, perhaps trillions, of possibilities there is no way to quantify the theory.  Chess theorists depend on chessmetrics, but they are making subjective analyses.  There are too many possibilities to even come close to a “right” answer every time. But the modern databases of openings by Masters and the outcome of the game is a quantifiable number. But it has not always been so. 

 

Let us turn the clock back to the late 1980s and early 1990s. Garry Kasparov was the  long-running world champion of chess.  IBM was the long-running champion of building mainframe computers.  IBM decided to build a computer to take on Kasparov.  Little doubt in the marketing or chess world it was a publicity stunt.  It was hoped it would raise stock prices.

 

IBM’s first attempt, Deep Thought, did not fair well at all.   It lost every game to Kasparov and even lost to a computer tournament in 1989 to programs running on a desktop PC.  So IBM built a mainframe based on brute force computing power called Deep Blue, a play on IBM’s nickname, Big Blue.   IBM also hired Grandmaster, Joel Benjamin, to help it develop its opening repertoire or “book”.

 

The first version (Deep Thought II, or Deep Blue Prototype, as it was known later in a PR move) lost to WChess and Fritz, both running on personal computers.

 

In 1996, Deep Blue became the first computer to beat a reigning World Champion.  It won a battle, but ultimately lost the war and Kasparov won the match.

 

A year later, working harder and increasing the computer brute force strength to Deep Blue (unofficially called Deeper Blue) it beat Kasparov.  But Kasparov accused IBM of cheating by using human players to intervene in the game.  Traps the computer fell for earlier, it did not in the final game.  Kasparov simply asked for the log print of the moves.3   IBM denied his request, adding more controversy.  Later it made the logs public on the internet, but no one knows if they were the true logs or not.

 

A documentary film was made in 2003 titled Game Over: Kasparov and the Machine”.   The film implied strongly that “Deep Blue” and its development was a traditional publicity stunt.   But it also implied that IBM did it to raise its stock value by showing it could build the world’s best computer. 

 

That was the last time IBM was involved with chess.  It dismantled Deep Blue and never again tried. The documentary came to the conclusion IBM was in it for the publicity and to drive the stock price up.  Dismantling Deep Blue kept the truth of Kasparov’s accusations hidden forever.  

 

But what IBM did set into motion was the building of a database of openings.  Anyone who has a chess computer or a program for their home computer knows or finds out that it comes in two parts:  the computing engine and the database.  Commercial databases of master games are available in standard formats to add to your computer.

 

Today, Joel Benjamin would not be needed.  Databases abound.  Hence we have databases of openings, complete games and, of course, the conclusion of the game.  We can say if two Masters played and one used this opening, he/she may have a calculated probability of winning, losing or drawing at well-published rates.

 

What I would like to do, using the database available to Chess.com Diamond Members, is find the top 10 statistical openings for white and black.  Then using Alan Bozon’s 5 research about how to beat computers, try to extrapolate that knowledge into to how to beat someone who you play who might have a higher rating. 

 

But here I want to reiterate what IM David Preuss has said (I think it was on a TV broadcast) that a good player, who knows their opening and the theory behind it, will beat you with it,  even if you know it is coming.  Why?  Because if you learn more than just the moves, and understand what the moves are doing, you can adjust your play to accomplish the same goals, even if your opponent throws the kitchen sink at you, if you know the theory you can find different routes to victory.

 

Now Bozon researched from from a Master’s database.  I compared it to the database available on Chess.com.  There were only minor differences.  

 

He decided to use chess tournament scoring.  He took the wins and gave them a value of 1 and took the draws and gave them a value of 1/2.  This is standard scoring.  Some would argue when selecting an opening, that the draws should be considered less important.  But if tournaments give them one for a win and 1/2 for a draw -- I think that is what should be used.  So, here are the top 10 white openings:

 

 


Opening

Win

Lose

Draw

Pts

Queen’s Gambit

40%

24%

36%

58

Blackmar-Diemer Gambit*

49%

35%

16%

57

Ruy Lopez

40%

27%

33%

56.5

Bishop’s Opening

41%

29%

30%

56

Vienna Game

41%

30%

29%

55.5

Benko Opening

38%

27%

35%

55.5

Reti Opening

37%

26%

37%

55.5

Centre Game

44%

34%

22%

55

Scotch Game

40%

30%

30%

55

English Opening

38%

28%

34%

55


 

 

 

 

Firstly the Blackmar-Diemer Gambit is almost 10 full percentage points higher in the “win” column.  It also has a very low draw rate.  But it has the largest losing percentage in the top 10.  No doubt this is a strong and sharp opening, which requires precision.  Not something you read a few pages of in a book of chess openings and play the next day.  You  will most likely get killed.

 

Secondly, all the top 10 are very close when you look at the “points” column.  Back to what IM Preuss said.  You must know your openings very well and the theory behind them.  Just knowing 3 to 7 chess book moves will also get you killed.  So what is a player to do with such data? 

 

First, you need to know 2 or 3 openings very well.  If you find one you like (and thematic tournaments at Chess.com are a great way to discover them) then study it and its theory.   I have one of the top 10 openings as one of my favourites.  I bought a book written by a Grandmaster just on that opening so I could learn it well.  My rating took a beating for a while but is now beginning to climb as I see, from the book, how my two white and two black openings fair and how to adjust.

 

Here are the top 10 openings for black:

 

 


Opening

Win

Lose

Draw

Pts

Sicilian Defense

34%

37%

29%

48.5

Nimzo Indian

30%

33%

37%

48.5

Robatsch Defense

33%

38%

29%

47.5

Alekhine Defense

32%

38%

30%

47

Nimzowitsch Defense

34%

30%

25%

46.5

Rat Opening

30%

37%

33%

46.5

Benko Gambit

32%

40%

28%

46

Modern Defense

31%

39%

30%

46

Queen’s Indian Defense

25%

30%

42%

46

Pseudo King’s Indian

29%

38%

33%

45.5


 

 

 

 

 

 

Again, just by glancing at the table, some things should stick out.  First of all, at the master level, the extra tempo that white gets makes a big difference.  If you win playing black you are certainly beating the odds and drawing is no shame.  And again, from top to bottom of the Top 10 list, there are only minor differences.

 

Erik, the owner and CEO of Chess.com, on a TV broadcast, was asked by a viewer, “What is your favourite opening?”  Erik kind of quipped, “e4, is there any other opening?.”  Well, look at the table, e4 does not bode well.

 

The reason is the common use of the Sicilian Defense.  It has a chilling effect on the e4 opening.  But still, e4 is the most common opening for white.  

 

To me, this suggests that everyone should know the Sicilian Defense.  Everyone.  Black is at a rather large disadvantage down one tempo (a tempo is a 1/2 turn) that black needs to do what it can.  Many of the Top 10 black defenses were designed to stop the e4 opening.  And notice that e4 is not as popular as some might think in the Top 10 white openings.

 

So what does this add up to?  That is up to you.  Chess is logic, science, sport, competition, and art.  The number of possible moves is so large, it might as well be infinite.  The human element plays the largest role.  To see that more clearly, read IM Silman’s book “The Amateur’s Mind6.   No doubt some of the smallest things make the biggest differences.  Confidence is a large part.

 

Now some ways to perhaps throw a higher (less than Master-skilled) rated player off his/her game.  This too is taken from Bozon.7  I took some of his ideas and extrapolated parts of them to fit the needs and mindsets of humans.  He wrote this on how to beat a chess computer.  But humans are more irrational at time, can be easily confused.  If it works on a computer, most likely with a little adaptation, it should work in many cases by a lower-rated player vs. a higher-rated player.  I seriously doubt Masters will be thrown off their games, but for good amateurs, especially those who have built their rating by playing low-rated opponents much of the time, these ideas might be worth considering.

 

First, an experienced player will have seen many of the “major” openings.  This is a stronger player’s advantage.  The most well-known example was when Fischer played Spassky for the World Championship.  The Soviet Chess machine had been preparing Spassky for the now famous “e4” opening used almost exclusively by Fischer.  In a game Fischer especially needed, he opened with “c4”.  

 

So perhaps find a solid, but unusual opening and know it well.  It could give your higher-rated opponent some problems.  At the very least, he/she will probably take a lot more time in the beginning.  Also, it may give you the upper hand before your opponent can recover.

 

In the mid game, keep as much pressure up as possible.  If your opponent is using his/her time to figure out a defense, not much time can be spent on offense or counter-offense.  The opposite is true for you.  Even if you are playing defensively, be thinking what  opportunities playing offensively in your opening could bring you when your time comes.

 

And in the end game, the more pieces, the more complicated.  Do not trade pieces in the end game unless it really helps your position.  More pieces, more possibilities and more for your opponent to worry about. 

 

Lastly, be thinking during your opponent’s turn.  Think ahead. The faster you move with precision, the more you will shake your opponent’s confidence.  At Chess.com, one of the most annoying things you can do is set up conditional moves.  Your opponent just played and oops. it is their turn again.  No time to think.  If you can out-guess your opponent with conditional moves several times, they can become really annoyed which could affect their performance.

 

Jim Fox, Ph.D.


[1] Wikipedia   (chess_computer)   Oct. 24, 2009

[2] ICGA Tournaments   No Date Given.

[3] IBM   No Date Given.

[4] Weeks, Mark:  About.com Column “Game Over: Did IBM Cheat Kasparov?  “  ;  June 2005.

[5] Bozon, Adam:   No Date Given.

Comments


  • 5 years ago

    Bill_Buchanan_tie

    Thanks Jim. Very interesting. I will take youe advice and learn the Sicilian Defense. Keep them coming.

  • 5 years ago

    JollyPlayer

    A Professional proofreader helped with the proofreading.   Should read better now.

  • 5 years ago

    JollyPlayer

    I'll fix the grammar errors.  These ideas come to me late a night sometimes.

    As for a large database, this was done on a very large database.  I compared a sample of chess.com to see if they were close -- and they were.  This was done on hundreds of master level games.

    Bozon used a huge database.  I used his numbers only after checking to see if chess.com numbers were close.

    Jim

  • 5 years ago

    chessoholicalien

    Very interesting article (a few grammar errors though). I'd like to see the same method used on the much larger Mega Database 2009 (ChessBase).

  • 5 years ago

    callastyle2k

    Nice analaysis. I'm relearning the game, and this is great food-for-thought.

    Thanks Smile

Back to Top

Post your reply: