Tracking a cheater (help requested)

[Albert Silver]

You may have read about the oddball cheating incident recently reported on about a young WGM and her father, who lied and cheated their way into a sub-1700 tournament they won. The thing is: these are rarely isolated incidents, and the chances are heavy that they cheated before too. Indeed I found that her games as a 14-year-old in 2007 where she made a norm, are incredibly suspicious. Even her loss to Caruana stinks of engine use. See here what I mean:

Caruana-Stock, Trieste 2007

[D]3rr1k1/1ppbqpp1/p1np1n1p/4p3/B3P2B/2PPRN1P/PP3PP1/R2Q2K1 b - - 0 15

Here the girl played the extremely odd 15...Kh8, but this alone could have a purpose of course. Caruana, who had just captured with Rxe3, now came back with Re1. If I tell you an engine played the next move, can you guess what it could be for me to say this?

[D]3rr2k/1ppbqpp1/p1np1n1p/4p3/B3P2B/2PP1N1P/PP3PP1/R2QR1K1 b - - 0 16

16...Kg8! Yes, this ridiculous king shuffling STINKS of one of those pointless engine plays when it is not sure how to proceed.

Caruana then played 17.Bc2

[D]3rr1k1/1ppbqpp1/p1np1n1p/4p3/4P2B/2PP1N1P/PPB2PP1/R2QR1K1 b - - 0 1

Lara replied with a mistake 17...g5? which ultimately costs her the game. Several engines are quick to see how to punish g5, but are not so quick to see it is really bad before it is played.

I took Rybka 2.2, which is 2006-2007, and ran it multi-PV in 32-bit mode. After 14 full plies here is what it said:

Analysis by Rybka 2.2 mp 32-bit:

1. = (0.25): 17...g5 18.Bg3 Qf8 19.d4 exd4 20.cxd4 Nh5 21.Bh2 Nf4 22.Bxf4 gxf4 23.Qd2 Qg7
2. +/= (0.28): 17...a5 18.Qd2 Ra8 19.d4 a4 20.Bd3 Na5 21.Bg3 Nh5 22.Bh2 Qf6
3. +/= (0.29): 17...Rb8 18.Qd2 a5 19.d4 a4 20.Bd3 Red8 21.d5 Na5 22.Qe3 b5

The question is whether anyone can locate an engine from this time (Fritz, Shredder, or other) that actually chooses g5 for any number of plies as a primary single choice? Don't bother with obscure engines as these aren't engine geeks we're talking about.

[Robert Flesher]

Hello Albert, this seems to be what you are looking for.


New game
3rr1k1/1ppbqpp1/p1np1n1p/4p3/4P2B/2PP1N1P/PPB2PP1/R2QR1K1 b - - 0 1

Analysis by Zappa Mexico II:

3...Na5 4.b4 Nc6 5.Qd2 Kh7 6.a4 g5 7.Bg3
+/= (0.37) Depth: 10/28 00:00:00 1407kN
3...Na5 4.b4 Nc6 5.Qd2 Kh7 6.a4 g5
+/= (0.37) Depth: 10/28 00:00:01 2486kN
3...Na5 4.b4 Nc6 5.Qd2 Na7 6.d4 Nc8 7.Bb3 Nb6
+/= (0.57) Depth: 11/37 00:00:01 3503kN
3...g5 4.Bg3 b5 5.d4 Rb8 6.a4 exd4 7.axb5 Rxb5 8.cxd4
+/= (0.46) Depth: 11/37 00:00:01 5372kN
3...Kh8 4.d4 Rg8 5.b3 g5 6.Bg3 g4 7.hxg4 Bxg4
+/= (0.42) Depth: 11/37 00:00:02 7198kN
3...a5 4.d4 exd4 5.cxd4 Nb4 6.a3 Nxc2 7.Qxc2
+/= (0.37) Depth: 11/37 00:00:02 7829kN
3...a5 4.d4 exd4 5.cxd4 Nb4 6.a3 Nxc2 7.Qxc2
+/= (0.37) Depth: 11/37 00:00:02 8081kN
3...a5 4.Qd2 Ra8 5.d4 a4 6.d5 Na5 7.b3 g5
+/= (0.49) Depth: 12/37 00:00:03 11823kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2
+/= (0.42) Depth: 12/37 00:00:04 16586kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2
+/= (0.42) Depth: 12/37 00:00:04 18093kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2 g4 9.hxg4 Bxg4
+/= (0.45) Depth: 13/41 00:00:05 20180kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2 g4 9.hxg4 Bxg4
+/= (0.45) Depth: 13/41 00:00:06 25324kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2
+/= (0.45) Depth: 14/41 00:00:08 31437kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2
+/= (0.45) Depth: 14/41 00:00:09 37660kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2 g4 10.hxg4 Bxg4
+/= (0.51) Depth: 15/43 00:00:15 64125kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2 g4 10.hxg4 Bxg4
+/= (0.51) Depth: 15/43 00:00:29 122mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.d5 Ne7 8.Nxe5 dxe5 9.Bxe5 Ng6 10.Bxc7 Rc8 11.Bxa5
+/= (0.51) Depth: 16/44 00:00:35 147mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.d5 Ne7 8.Nxe5 dxe5 9.Bxe5 Ng6 10.Bxc7 Rc8 11.Bxa5
+/= (0.51) Depth: 16/44 00:00:42 176mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 17/48 00:01:05 279mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 17/48 00:01:32 397mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 18/48 00:01:36 413mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 18/48 00:02:09 554mN

(, Microsoft 04.03.2011)

[Robert Flesher]

After 5min on much faster hardware than they would have had.


New game
3rr1k1/1ppbqpp1/p1np1n1p/4p3/4P2B/2PP1N1P/PPB2PP1/R2QR1K1 b - - 0 1

Analysis by Zappa Mexico II:

3...Na5 4.b4 Nc6 5.Qd2 Kh7 6.a4 g5 7.Bg3
+/= (0.37) Depth: 10/28 00:00:00 1407kN
3...Na5 4.b4 Nc6 5.Qd2 Kh7 6.a4 g5
+/= (0.37) Depth: 10/28 00:00:01 2486kN
3...Na5 4.b4 Nc6 5.Qd2 Na7 6.d4 Nc8 7.Bb3 Nb6
+/= (0.57) Depth: 11/37 00:00:01 3503kN
3...g5 4.Bg3 b5 5.d4 Rb8 6.a4 exd4 7.axb5 Rxb5 8.cxd4
+/= (0.46) Depth: 11/37 00:00:01 5372kN
3...Kh8 4.d4 Rg8 5.b3 g5 6.Bg3 g4 7.hxg4 Bxg4
+/= (0.42) Depth: 11/37 00:00:02 7198kN
3...a5 4.d4 exd4 5.cxd4 Nb4 6.a3 Nxc2 7.Qxc2
+/= (0.37) Depth: 11/37 00:00:02 7829kN
3...a5 4.d4 exd4 5.cxd4 Nb4 6.a3 Nxc2 7.Qxc2
+/= (0.37) Depth: 11/37 00:00:02 8081kN
3...a5 4.Qd2 Ra8 5.d4 a4 6.d5 Na5 7.b3 g5
+/= (0.49) Depth: 12/37 00:00:03 11823kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2
+/= (0.42) Depth: 12/37 00:00:04 16586kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2
+/= (0.42) Depth: 12/37 00:00:04 18093kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2 g4 9.hxg4 Bxg4
+/= (0.45) Depth: 13/41 00:00:05 20180kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2 g4 9.hxg4 Bxg4
+/= (0.45) Depth: 13/41 00:00:06 25324kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2
+/= (0.45) Depth: 14/41 00:00:08 31437kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2
+/= (0.45) Depth: 14/41 00:00:09 37660kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2 g4 10.hxg4 Bxg4
+/= (0.51) Depth: 15/43 00:00:15 64125kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2 g4 10.hxg4 Bxg4
+/= (0.51) Depth: 15/43 00:00:29 122mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.d5 Ne7 8.Nxe5 dxe5 9.Bxe5 Ng6 10.Bxc7 Rc8 11.Bxa5
+/= (0.51) Depth: 16/44 00:00:35 147mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.d5 Ne7 8.Nxe5 dxe5 9.Bxe5 Ng6 10.Bxc7 Rc8 11.Bxa5
+/= (0.51) Depth: 16/44 00:00:42 176mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 17/48 00:01:05 279mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 17/48 00:01:32 397mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 18/48 00:01:36 413mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 18/48 00:02:09 554mN
3...g5 4.Nxg5 hxg5 5.Bxg5 Kh8 6.Qf3 Kg7 7.Qg3 Rg8 8.Re3 Kf8 9.Qh4 Rxg5 10.Qxg5 Be6 11.Rf3 Ng8 12.Qh5 f6 13.Ba4 Qf7 14.Qh8 Nce7
+/- (0.83) Depth: 19/49 00:03:30 908mN

(, Microsoft 04.03.2011)

[Robert Flesher]

After 6:23 Zappa finally switches to Nb8. However, taking hardware considerations into the comparision it's likely they would not have reached this depth in a reasonable time frame. So g5? would have been played.




New game
3rr1k1/1ppbqpp1/p1np1n1p/4p3/4P2B/2PP1N1P/PPB2PP1/R2QR1K1 b - - 0 1

Analysis by Zappa Mexico II:

3...Na5 4.b4 Nc6 5.Qd2 Kh7 6.a4 g5 7.Bg3
+/= (0.37) Depth: 10/28 00:00:00 1407kN
3...Na5 4.b4 Nc6 5.Qd2 Kh7 6.a4 g5
+/= (0.37) Depth: 10/28 00:00:01 2486kN
3...Na5 4.b4 Nc6 5.Qd2 Na7 6.d4 Nc8 7.Bb3 Nb6
+/= (0.57) Depth: 11/37 00:00:01 3503kN
3...g5 4.Bg3 b5 5.d4 Rb8 6.a4 exd4 7.axb5 Rxb5 8.cxd4
+/= (0.46) Depth: 11/37 00:00:01 5372kN
3...Kh8 4.d4 Rg8 5.b3 g5 6.Bg3 g4 7.hxg4 Bxg4
+/= (0.42) Depth: 11/37 00:00:02 7198kN
3...a5 4.d4 exd4 5.cxd4 Nb4 6.a3 Nxc2 7.Qxc2
+/= (0.37) Depth: 11/37 00:00:02 7829kN
3...a5 4.d4 exd4 5.cxd4 Nb4 6.a3 Nxc2 7.Qxc2
+/= (0.37) Depth: 11/37 00:00:02 8081kN
3...a5 4.Qd2 Ra8 5.d4 a4 6.d5 Na5 7.b3 g5
+/= (0.49) Depth: 12/37 00:00:03 11823kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2
+/= (0.42) Depth: 12/37 00:00:04 16586kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2
+/= (0.42) Depth: 12/37 00:00:04 18093kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2 g4 9.hxg4 Bxg4
+/= (0.45) Depth: 13/41 00:00:05 20180kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Qd2 Nh5 8.Bh2 g4 9.hxg4 Bxg4
+/= (0.45) Depth: 13/41 00:00:06 25324kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2
+/= (0.45) Depth: 14/41 00:00:08 31437kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2
+/= (0.45) Depth: 14/41 00:00:09 37660kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2 g4 10.hxg4 Bxg4
+/= (0.51) Depth: 15/43 00:00:15 64125kN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.Rc1 Qg6 8.Bd3 Nh5 9.Bh2 g4 10.hxg4 Bxg4
+/= (0.51) Depth: 15/43 00:00:29 122mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.d5 Ne7 8.Nxe5 dxe5 9.Bxe5 Ng6 10.Bxc7 Rc8 11.Bxa5
+/= (0.51) Depth: 16/44 00:00:35 147mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 a5 7.d5 Ne7 8.Nxe5 dxe5 9.Bxe5 Ng6 10.Bxc7 Rc8 11.Bxa5
+/= (0.51) Depth: 16/44 00:00:42 176mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 17/48 00:01:05 279mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 17/48 00:01:32 397mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 18/48 00:01:36 413mN
3...g5 4.Bg3 Qf8 5.d4 Qg7 6.a4 g4 7.hxg4 Bxg4 8.d5 Ne7 9.Qd3 Rb8 10.Nh4 Nh5
+/= (0.51) Depth: 18/48 00:02:09 554mN
3...g5 4.Nxg5 hxg5 5.Bxg5 Kh8 6.Qf3 Kg7 7.Qg3 Rg8 8.Re3 Kf8 9.Qh4 Rxg5 10.Qxg5 Be6 11.Rf3 Ng8 12.Qh5 f6 13.Ba4 Qf7 14.Qh8 Nce7
+/- (0.83) Depth: 19/49 00:03:30 908mN
3...Nb8 4.d4 Bc8 5.Qe2 Nbd7 6.Bg3 Nh5 7.Bh2 Nhf6 8.Rad1 Nh7 9.b4 Ng5 10.Nxg5 hxg5 11.Bb3
+/= (0.58) Depth: 19/50 00:06:24 1687mN

(, Microsoft 04.03.2011)

[Robert Flesher]

In this position Kh8 is also considered second best, which could change with time and search.

[d]3rr1k1/1ppbqpp1/p1np1n1p/4p3/B3P2B/2PPRN1P/PP3PP1/R2Q2K1 b - - 0 1



New game
3rr1k1/1ppbqpp1/p1np1n1p/4p3/B3P2B/2PPRN1P/PP3PP1/R2Q2K1 b - - 0 1

Analysis by Zappa Mexico II:

New game
3rr1k1/1ppbqpp1/p1np1n1p/4p3/B3P2B/2PPRN1P/PP3PP1/R2Q2K1 b - - 0 1

Analysis by Zappa Mexico II:

1. +/= (0.37): 1...Nb8 2.Bb3 c5 3.Bxf6 Qxf6 4.Nh2 Rf8 5.Rf3 Qe7 6.Nf1 Nc6 7.Ne3 Na5 8.Bd5 Rc8 9.Nf5 Bxf5 10.Rxf5
2. +/= (0.38): 1...Kh8 2.d4 b5 3.Bc2 Na5 4.Re1 Nc4 5.b3 Nb6 6.Rb1 a5 7.Bd3 a4 8.d5 Qf8

(, Microsoft 04.03.2011)

[jdart]

I don't think this is evidence of cheating by itself.

You could view .. Kh8 perhaps as inviting a draw by repetition, which White could have accepted by also shuttling the Rook.

Some strong players also repeat moves (<3 times) because it's an easy move that gives them some extra time to think (and gets them 1 move closer to time control, if that is a factor).

[michiguel]

I don't think this is evidence of cheating by itself.

You could view .. Kh8 perhaps as inviting a draw by repetition, which White could have accepted by also shuttling the Rook.

Some strong players also repeat moves (<3 times) because it's an easy move that gives them some extra time to think (and gets them 1 move closer to time control, if that is a factor).

I disagree in this case. It's extremely unlikely a human player will play Kh8 + Kg8 + g5. Particularly Kg8 followed by g5 does not make any sense. I may accept Kh8 and g5 (you make room for the rook to come to g8) but playing Kg8 and g5 is a typical computer weak maneuver.

Miguel

[Mike S.]

17...g5 looks like a natural move to me and the refutation seems quite deep. I don't think that any suspicion should be based on it, for that and also because it's only one move. I'm almost sure you could find such "natural but refuted" moves in very many losses of strong players (including the pre-computer era! ).

More promising would be, if a player often wins with "mysterious" tactics which seem beyond human grasp. But also, this has been seen long before the computer era, and not only from very famous players. Chess books are full of that.

P.S. The Hamilton Chess Club responded to the incident with good sense of humour:

http://hamiltonchess.net/2011/03/02/fem ... es-shield/

[Mike S.]

Kg8 and g5 is a typical computer weak maneuver.

Things must be more complicated here, when apparently Zappa Mexico II choses 17...g5 up to depth 19.

Also, I think playing Kg8-Kh8-Kg8 when it seems (! a judgement may fail) appropriate, can also be considered human play by comps, not just vice versa. Humans were first!

[Robert Flesher]

I don't think this is evidence of cheating by itself.

You could view .. Kh8 perhaps as inviting a draw by repetition, which White could have accepted by also shuttling the Rook.

Some strong players also repeat moves (<3 times) because it's an easy move that gives them some extra time to think (and gets them 1 move closer to time control, if that is a factor).

I never stated or believed this was the cased. I just provided Albert with what he was looking for. If it is chess related and I get to do some analysis, then I am game. I could care less for the implications, the game and the chess was of interest.