We want to know what percentage of games from a position resulted in wins, losses or draws. While useful, these statistics can also be very misleading. Reasons for misleading statistics are 1) the number of games is too small to give reliable results. 2) Most games are decided by mistakes occurring later in the game and not as a direct result of the opening. 3) Unless the database is current, the statistics don’t take into account later improvements. In an extreme example, White may have won 100 games in a given line, but a refutation may have been found that busts the line. So the +100 -1 =0 stat means nothing. 4) Most DB’s don’t take into account the strength of the players. 5) Most stats don’t take into account the number of draws. What does a score of 50% mean? Out of 100 games were they all drawn, did White and Black each win 50 games or was the score +25 -25 =50? With most DB stats, you just don’t know.
This latter factor can make a difference in your decision. If you are the lower rated player who wants to gain some rating points or if you are playing for a title then you might want to consider playing the 100% drawing line. On the other hand, if you are the higher rated player or a draw won’t do, you might want to play the +25 -25 =50 line.
So, to successfully use opening statistics you should have a large selection of games by highly rated players, it should be kept up-to-date and before actually playing any line you need to examine the entire game to see if the opening was responsible for the result or if it was the result of poor play occurring much later in the game.
Also, it is wise to select an opening you are familiar and comfortable with so when you get out of the book you will have some idea of how to continue. You also have to pay attention to the number of draws that resulted from a particular line. If you want to avoid a draw, then don’t pick a line that lead to 60% draws. Most important of all, play over the entire game when evaluating the stats! You will want to know if the result was influenced by the opening or an endgame blunder on move 75!
Recent games annotated by GM’s is another source for obtaining good information. GM’s still have a better positional understanding than engines. Once reaching a position of interest you can have the engine spend time searching deeper into positions the GM thought were worth examining.
CC GM Robin Smith suggests that you use the following procedure: 1) on the first pass have the engine analyze at just a few seconds per move, say 5 seconds. 2) eliminate unpromising suggestions. 3) if you find a variation where the evaluation reverses, say from favorable for White to favorable for Black, delete the moves after the evaluation reverses.
Now you have an annotated game with lines of interest. Let the engine examine the suggested moves for as much time as you can. It’s also a good idea to use more than one engine when doing this because they will often not agree on the best move and some engines are more optimistic than others.
Once done, it’s time to take a look at what you have. What do you do when the analysis suggests a move that the GM didn’t? Who do you believe? My experience is that when it comes to evaluating a position the GM’s ‘gut feeling’ is usually more reliable than the engine. Still, the dilemma remains. Is the engine suggested opening novelty a good one or not?
We aren’t strong enough to make that decision, so what do we do? One of the best ways of resolving the dilemma is probably playing a series of engine matches using the novelty and seeing if it is likely to result in success.
All of this will take time and patience, but it’s the way high level CC is played these days. Even if you’re not playing ‘high level’ you can still use this technique to search out promising opening ideas in your games.