Skip to content
All posts
For improvers3 min read

Which openings punish the exit

In the King's Gambit Accepted the move that leaves theory is a mistake in 30% of cases at 1400–1799 Elo, in the Polish Opening in 4%. Open games punish the exit hardest.

Mistake rate on the exit, the eight riskiest openings
  • King's Gambit Accepted30%
  • King's Gambit Declined25%
  • Elephant Gambit24%
  • Scotch Game23%
  • Italian Game23%
  • Petrov's Defense22%
  • Four Knights Game21%
  • Ponziani Opening20%

Rating 1400–1799, half-moves 8 to 24, loss of 10 points or more.

Mistake rate on the exit, the eight safest openings
  • Polish Opening4%
  • Dutch Defense6%
  • Old Indian Defense6%
  • Mieses Opening6%
  • Bird Opening7%
  • Horwitz Defense7%
  • Semi-Slav Defense7%
  • Rat Defense8%

What the numbers show

  • The riskiest exits are in open games: King's Gambit Accepted, King's Gambit Declined, Elephant Gambit, Scotch Game. There you leave theory late, at a median of half-move 12 to 15, and then pay the most.
  • Closed and flank openings are safest: Polish Opening, Dutch Defense, Old Indian Defense, Mieses Opening. The mistake rate on the exit is below 8% there.
  • The risk belongs to the opening, not only to the exit: where the exit is risky, the theory before it is too (correlation 0.97) and so is the way after it (correlation 0.86).
  • Compared with inside theory, the jump is smallest in the risky openings, a median of 3.2 times, because the theory there is already sharp, and largest in the safe ones, 8.5 times.
  • Theory reaches furthest in the Italian Game (median half-move 15) and is shortest in the Van't Kruijs Opening (half-move 7).
The eight riskiest and the eight safest openings
OpeningTheory endsinside theoryexitafter
King's Gambit Accepted1210.7%29.6%17.9%
King's Gambit Declined119.1%25.1%17.1%
Elephant Gambit117.0%24.1%14.9%
Scotch Game158.1%23.2%14.4%
Italian Game155.9%22.7%14.0%
Petrov's Defense138.1%22.0%13.6%
Four Knights Game155.7%20.5%13.9%
Ponziani Opening135.7%20.4%13.5%
Polish Opening80.1%4.3%8.6%
Dutch Defense110.5%5.6%10.8%
Old Indian Defense110.7%6.3%9.5%
Mieses Opening70.9%6.4%11.8%
Bird Opening90.8%6.7%11.4%
Horwitz Defense100.9%6.9%10.5%
Semi-Slav Defense140.4%7.3%8.5%
Rat Defense102.0%7.8%11.5%

Rating 1400–1799. Theory ends: median half-move.

Sources

  1. [1]E. B. Wilson (1927): Probable inference, the law of succession, and statistical inference. Journal of the American Statistical Association 22(158), pp. 209–212, doi:10.1080/01621459.1927.10502953. jstor.org/stable/2276774 (accessed 2 October 2026).

Data: public Lichess games from January to August 2026, 74.2 million games with an engine rating. A mistake costs at least 10 points of winning chances. Lichess analysed these games on request, so they are not a random sample of all games. Groups by the rating of the player who moves, ranges are 95 percent Wilson intervals [1].

Try it on your own games

Enter your Lichess name and see where your knowledge ends.

Try it on your own games