A new study reveals that using computer advice for just three moves in a chess game can boost a player's victory odds from 51 percent to 84 percent. Researcher Daniel Keren simulated thousands of matches to demonstrate how selective cheating evades detection. The findings highlight vulnerabilities in online chess platforms' anti-cheating measures.
Daniel Keren from the University of Haifa in Israel conducted a simulation of 100,000 chess matches using the Stockfish engine, pitting engines rated at 1500 Elo—equivalent to an average player—against each other. In half the games, no external help was provided, while the other half allowed interventions from a superior 3190 Elo engine on select moves.
Normally, the player with the white pieces holds a slight edge, winning about 51 percent of the time due to the first-move advantage. However, Keren's results showed that seeking computer advice on a single move raises this to 66 percent. With three such interventions, the win probability surges to 84 percent on average.
"I thought that one cheat would increase the ratio to 55 per cent and another one to maybe 60 per cent," Keren remarked. "Cheating three times and you reach 84 per cent – to me, that was astounding."
Timing proves crucial: a single intervention around the 30th move can enhance win chances by 15 percentage points, outperforming five randomly timed cheats, which yield only a 7.5-point gain. The simulation's algorithm intervened only when the suggested move substantially improved outcomes, with thresholds tightening as the game progressed, offering a form of disguise against detection.
Keren emphasizes that his research aims to inform chess platforms about cheating risks, not encourage it. "The idea is to see what cheating can do," he said. "Know thy enemy, right?"
Expert Kim Schu from the University of Mainz agrees, noting, "A single engine ‘hint’ in the right position can be game-deciding, and because humans can sometimes find the same best move, that kind of selective cheating is unusually difficult to prove from move analysis alone."
To combat this, Schu advocates combining move analysis with behavioral monitoring, move timings, and account histories for robust anti-cheating systems, especially as online play grows.