spectacular Komodo handicap game

[S.Taylor]

The following handicap game was perhaps the most spectacular one I have ever seen. It was one of an unofficial series played on chess.com between Komodo 11.3.1 (on my i9 16 core machine) and my son Raymond Kaufman, who is an IM. These games were played to test out automated play on chess.com as well as to test Komodo 11.3.1 and to try out various handicaps for future use in official games. Although they were unofficial, Ray took them seriously and used his time. Time limit for this game was 15 min plus 10 sec per move. Handicap was Knight (g1) for tempo, meaning White gets to play his first move (e4), Black "takes" White's knight off, and it is then White's move. Philidor used to give this handicap in the 1790s to players too strong for knight odds but not strong enough for pawn and two or three moves. I think it works for this purpose.

[pgn][Event "Knight for tempo handicap"]
[Site "chess.com"]
[Date "2018.03.07"]
[Round "?"]
[White "Komodo 11.3.1"]
[Black "Kaufman, Raymond"]
[Result "1-0"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/4P3/8/PPPP1PPP/RNBQKB1R w KQkq - 0 1"]
[PlyCount "57"]
[SourceDate "2018.03.07"]
[SourceVersionDate "2018.03.07"]

{[#]} 1. d4 d6 2. Be3 Nf6 3. Nc3 c6 4. f3 e5 5. Qd2 h5 6. h3 Nbd7 7. O-O-O Be7
8. Rg1 h4 9. Kb1 Nh5 10. Bc4 O-O 11. Ne2 exd4 12. Bxd4 Ne5 13. Bb3 Bf6 14. f4
Ng6 15. g4 hxg3 16. f5 Ne5 17. Nxg3 Nf3 18. Qd3 Nxg3 19. Qxf3 Bxd4 20. Rxd4 Qf6
21. Rxg3 d5 22. Qf2 Re8 23. a4 dxe4 24. Rg6 Qxf5 25. Qg1 Qe5 26. Rxe4 Qxe4 27.
Rxg7+ Kh8 28. Qg5 Bf5 29. Rxf7 1-0 [/pgn]

My belief is that you cannot really judge a sub 2700 human when tactics go out of hand vs a top engine. Certainly not in a rapid tc. U need to analize it later with engine to see if real inroads have been made that can truly outwit material handicaps.

[lkaufman]

The following handicap game was perhaps the most spectacular one I have ever seen. It was one of an unofficial series played on chess.com between Komodo 11.3.1 (on my i9 16 core machine) and my son Raymond Kaufman, who is an IM. These games were played to test out automated play on chess.com as well as to test Komodo 11.3.1 and to try out various handicaps for future use in official games. Although they were unofficial, Ray took them seriously and used his time. Time limit for this game was 15 min plus 10 sec per move. Handicap was Knight (g1) for tempo, meaning White gets to play his first move (e4), Black "takes" White's knight off, and it is then White's move. Philidor used to give this handicap in the 1790s to players too strong for knight odds but not strong enough for pawn and two or three moves. I think it works for this purpose.

[pgn][Event "Knight for tempo handicap"]
[Site "chess.com"]
[Date "2018.03.07"]
[Round "?"]
[White "Komodo 11.3.1"]
[Black "Kaufman, Raymond"]
[Result "1-0"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/4P3/8/PPPP1PPP/RNBQKB1R w KQkq - 0 1"]
[PlyCount "57"]
[SourceDate "2018.03.07"]
[SourceVersionDate "2018.03.07"]

{[#]} 1. d4 d6 2. Be3 Nf6 3. Nc3 c6 4. f3 e5 5. Qd2 h5 6. h3 Nbd7 7. O-O-O Be7
8. Rg1 h4 9. Kb1 Nh5 10. Bc4 O-O 11. Ne2 exd4 12. Bxd4 Ne5 13. Bb3 Bf6 14. f4
Ng6 15. g4 hxg3 16. f5 Ne5 17. Nxg3 Nf3 18. Qd3 Nxg3 19. Qxf3 Bxd4 20. Rxd4 Qf6
21. Rxg3 d5 22. Qf2 Re8 23. a4 dxe4 24. Rg6 Qxf5 25. Qg1 Qe5 26. Rxe4 Qxe4 27.
Rxg7+ Kh8 28. Qg5 Bf5 29. Rxf7 1-0 [/pgn]

My belief is that you cannot really judge a sub 2700 human when tactics go out of hand vs a top engine. Certainly not in a rapid tc. U need to analize it later with engine to see if real inroads have been made that can truly outwit material handicaps.

You can't really judge any player, human or engine, by just one or two games, you need a decent sample. One thing I have noticed over the years about handicap play in chess, regardless of whether the players are humans, engines, or one of each, is that for any given pair of opponents there is some handicap that is roughly balanced, and if you raise or lower the handicap by any meaningful amount the benefitting player becomes a big favorite. For example, at five min plus five sec the balanced handicap between Komodo (on my i9) and me is between knight odds and knight plus White odds. If I play at anything less than a knight, such as two Black pawns, knight for pawn, exchange and pawn, pawn (f7) and 3 moves, etc. I almost always lose, but at rook odds I would almost always win. So I think that a player's strength can be measured pretty accurately by handicap play vs. Komodo, provided that he plays enough games under the same conditions as others have been well tested at.