One time, I watched one of my students playing a much higher rated player online. After the game he accused the student of cheating and using a computer in the endgame phase of the game. I thought it was pretty funny at the time, because the use of a computer to play was the last thing to think about using. I would have thought he would have made the accusation about the play in the first seventeen moves, a “book” line in the Ruy Lopez which my student partly studied and partly figured out through trial and error in other games.
Only a couple of times, the student turned to me and asked what to do. I gave a general answer and said it was the student’s choice. The correct move was selected and they moved on to the endgame. The opponent played it abysmally, much lower than a player with a similar OTB player would have conducted it. Only once in that ending did my student ask for advice. I mentioned that rook activity was better than greedy pawn grabs. I asked which pawn seemed most important to win. After about a minute’s thought, the right move was selected and the rest of the game was a massacre. Then the accusation came. It was denied. I advised the dad watching the game to no longer engage in any chat with opponents, as no good comes from that when a disgruntled player makes excuses or accusations.
The endgame was played all out of proportion to the ratings involved. The higher rated player demonstrated that his endgame skills were nowhere near what they should be. My student, out of frustration in playing badly in some other endgames, especially rook endings, had decided to pay more attention to that phase. I put together a group of endgames by Rubinstein, Capablanca, and other models to look at to get a better feel of the style and methods used. It was quite obvious to me that the work the student did had stuck. We also looked a couple of classic Fischer, Karpov and Kasparov games in the Ruy Lopez. Their clarity and some extra explanation helped in learning move orders. In some positions, it is a matter of style and temperament whether to fully close the position or keep it more open. My student chose to close the game and maneuver in a Karpovian way. The opponent went for an endgame. I thought it was too soon, but let the student figure out why.
While I offered a couple of pieces of advice, it was clear to me that the opponent was using a computer for long stretches of the game, especially the opening, which came back rapid fire. The poor endgame play came after a bad decision, which showed that the program was not up to the task. When we looked at the game afterward, that move was the first choice of the program we used to analyze, but was a bad move in general in rook endgames. After my student’s move, it flipped the evaluation and went passive, trying to hang on rather than seek counterplay. Not the first time I have seen that happen. He followed that style of computer recommendation and perished, losing key pawns, watching the king move in to put the finishing touches in a simple endgame. When we went back to the key move again, it took a while for the “best move” to pop up in the list of choices. So, if we were “cheating,” it was not with a computer. Human experience and study were being used optimally. The kid made all of the choices, some of them second best in the middlegame but not that bad, and ground the opponent down in a way you rarely see kids do.