The 95% Conjecture

Suppose you could try your best (or as best you can) 95% of the time. That sounds pretty good, right?

Let's assume the average chess game lasts about 40 moves. 95% of 40 is 38, so that means two moves each game you would not try your best. This observation led to a little thought experiment:

Suppose we have two players, virtual clones. One is rated 1800 and is able to try his/her best all the time. The other player is identical with that one exception: twice a game he/she suffers a "lapse" and can't put in the effort that the position and time on his clock would allow, essentially not even trying. What would be the rating of the second player?

I have had several players make estimates for me, and I agree with the consensus answers. These two lapses a game could easily make Player #2 1400-1500 instead of 1800, a loss of 300-400 rating points! That's a lot more rating points than a 1400 could get from memorizing a book on his favorite opening, for sure (even if that were possible).

This experiment shows that consistency in effort and concentration is worth quite a bit. It's no coincidence that if you go to a big tournament the ones who consistently take the most time are the better players.

What might cause such lapses? There could be many factors:

  • Forgetting to do all the things you know you should do, like make sure the opponent does not have a check, capture, or threat (forcing move) which might defeat your candidate move (See Real Chess, Time Management, and Care: Putting it All Together),
  • Fatigue or some sort of tiredness due to internal or external conditions, such as not getting enough liquids during the game or not having enough sleep,
  • Some sort of distraction, like worrying about what happened earlier in the game, or where you are going to eat dinner,
  • The inability to roll up the sleeves and take time to carefully analyze complex positions, whereas following general principles on more simple positions requires less effort, and/or
  • Quite a few other reasons are possible (such as the one in the example game below).
Whatever the cause, these lapses can be disastrous. Learning how to minimize or prevent them can make a big difference in your game. The first step is awareness.
     I have included a sample game. In this game White, the lower rated player, played much more slowly and carefully than his opponent for the first 40 moves (you can see the time remaining by clicking on a move). However, at the point when the endgame became very critical, he began speeding up, despite having plenty of time on the clock to at least try to work out the analysis. Around move 40 he had 11 minutes left (with a 45 second time increment) but at the end he was up to about 18 minutes. If you check with an engine, he missed wins on several consecutive moves near the end, finally allowing quite an undeserved draw. One reason he gave for speeding up is that he thought he was lost (despite having a winning position!). I always say there are only two gears you should use when playing:
     * Try your best (given the board and time conditions), or
     * Resign
So it turned out on the early, non-critical moves he played quite carefully but, once the game started riding on every move, that's when he speeded up! I call this syndrome "Getting Chess backwards!"

     One can only wonder if he would have taken several of those 18 minutes on some of his moves if he would have found at least one of those many wins...


  • 6 months ago


    This is a brilliant article with content that has mostly crossed my mind; but I had not idea the drop in elo would be so drastic!  I do not doubt it, though.  Thank you for laying your thoughts out so clearly for us.

  • 2 years ago


    Great stuff. So true yet difficult to accomplish.

  • 2 years ago


    Great article! After following your advice to examine any possible checks, captures, or threats by my opponent (as well as studying lots of tactics, per your advice as well), my rating on increased from 1200 to 1500. I still only use my mental effort on around 80 percent of the moves, so this article is great enouragement to keep up the effort throughout the whole game- Thanks Coach Heisman!

  • 2 years ago


    Very interesting, 57.Rf8+ would draw also by perpetual check. cause Re8 58. Rxe8+ Kxe8 59.Kc7 Kf7 (a2 bad move) 60.d7 a2 61.d8=Q a1=Q and after a lot of checks would draw

Back to Top

Post your reply: