Amid cheating accusations I would like to post one other very small thread of substance. The topic concerns the 45th game of Clemens Keck's 3Champs, where Stockfish, for a maybe 10th game in a row in a similar situation, lost to Komodo in the peculiar imbalance of queen versus 3 pieces. I would have gladly commented some other games of the tourney, however this was what actually caught my attention, not lastly because the topic of imbalances was recently discussed here, and there might still be some interest.
Here the game:
[pgn][PlyCount "152"]
[EventType "tourn"]
[Source "Doe"]
[Event "3Champs2013"]
[Date "2013.12.04"]
[Round "15.3"]
[White "Stockfish 241113 64 SSE4.2"]
[Black "Komodo 1142.00 64-bit x12-2"]
[Result "0-1"]
[EventDate "2013.11.28"]
[ECO "D18"]
[TimeControl "5400+15"]
[Annotator "0.50;0.05"]
[MLNrOfMoves "76"]
[MLFlags "000100"]
{Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz 2699 MHz W=36.4 plies; 16.905kN/s; Fritz 14.ctg B=27.8 plies; 17.178kN/s; Fritz 14.ctg}
1. d4 {[%eval 0,0] [%emt 0:00:00]} 1... d5 {[%eval 0,0] [%emt 0:00:00]} 2. c4
{[%eval 0,0] [%emt 0:00:00]} 2... c6 {[%eval 0,0] [%emt 0:00:00]} 3. Nc3
{[%eval 0,0] [%emt 0:00:00]} 3... Nf6 {[%eval 0,0] [%emt 0:00:00]} 4. Nf3
{[%eval 0,0] [%emt 0:00:00]} 4... dxc4 {[%eval 0,0] [%emt 0:00:00]} 5. a4
{[%eval 0,0] [%emt 0:00:00]} 5... Bf5 {[%eval 0,0] [%emt 0:00:00]} 6. e3
{[%eval 50,33] [%emt 0:01:58]} 6... e6 {[%eval 5,23] [%emt 0:01:07]} 7. Bxc4
{[%eval 36,35] [%emt 0:05:08]} 7... Bb4 {[%eval 9,27] [%emt 0:01:14]} 8. O-O
{(Ne5) [%eval 36,35] [%emt 0:00:15]} 8... O-O {[%eval 1,24] [%emt 0:02:20]} 9.
Nh4 {(Qb3) [%eval 34,35] [%emt 0:06:42]} 9... Bg6
{(Bg4) [%eval 0,24] [%emt 0:00:42]} 10. Qb3 {[%eval 56,33] [%emt 0:02:23]} 10...
Qb6 {[%eval 7,27] [%emt 0:00:52]} 11. Nxg6 {[%eval 54,33] [%emt 0:02:59]} 11...
hxg6 {[%eval 4,27] [%emt 0:01:41]} 12. Bd2 {(Rd1) [%eval 54,36] [%emt 0:00:23]}
12... Nbd7 {(c5) [%eval 0,29] [%emt 0:02:48]} 13. Rfd1
{(Ne4) [%eval 54,31] [%emt 0:01:55]} 13... a5
{(Rfc8) [%eval -1,25] [%emt 0:00:59]} 14. Be1
{(f3) [%eval 56,31] [%emt 0:01:21]} 14... Rac8 {[%eval 0,25] [%emt 0:02:48]} 15.
Rab1 {(f3) [%eval 58,31] [%emt 0:00:24]} 15... Qc7
{(e5) [%eval 0,26] [%emt 0:02:37]} 16. Rbc1 {[%eval 54,33] [%emt 0:02:29]} 16...
e5 {(g5) [%eval 0,26] [%emt 0:00:51]} 17. Nb5
{(Bd3) [%eval 68,36] [%emt 0:03:56]} 17... cxb5 {[%eval -10,27] [%emt 0:04:49]}
18. Bxf7+ {[%eval 56,43] [%emt 0:00:00]} 18... Rxf7
{[%eval -12,28] [%emt 0:01:12]} 19. Rxc7 {[%eval 66,41] [%emt 0:00:24]} 19...
Rxc7 {[%eval -11,25] [%emt 0:00:23]} 20. Bxb4 {[%eval 60,40] [%emt 0:07:47]}
20... axb4 {[%eval -9,29] [%emt 0:00:22]} 21. dxe5
{(h3) [%eval 74,39] [%emt 0:00:58]} 21... bxa4 {[%eval -37,25] [%emt 0:01:23]}
22. Qxb4 {[%eval 32,39] [%emt 0:05:38]} 22... Nxe5
{[%eval -36,29] [%emt 0:00:46]} 23. Qxa4 {[%eval 40,39] [%emt 0:01:07]} 23... g5
{(Rc4) [%eval -37,28] [%emt 0:01:49]} 24. h3 {[%eval 58,35] [%emt 0:01:38]}
24... Rfe7 {(Kh7) [%eval -38,26] [%emt 0:00:04]} 25. f3
{[%eval 58,36] [%emt 0:02:19]} 25... Kh7 {[%eval -40,28] [%emt 0:02:46]} 26. Qa3
{(Qb4) [%eval 36,36] [%emt 0:05:54]} 26... Kh6
{(Nc4) [%eval -41,28] [%emt 0:06:12]} 27. e4 {[%eval 35534,37] [%emt 0:08:19]}
27... Ng6 {[%eval -44,29] [%emt 0:02:59]} 28. Qb3
{(Qe3) [%eval 4,39] [%emt 0:00:00]} 28... Rc6 {[%eval -60,24] [%emt 0:00:54]}
29. Qe3 {[%eval 0,39] [%emt 0:00:34]} 29... Nh5
{(Rc2) [%eval -56,26] [%emt 0:00:38]} 30. Kf2 {[%eval 0,36] [%emt 0:01:13]}
30... Rf7 {[%eval -55,25] [%emt 0:00:55]} 31. Rd2
{(Rd8) [%eval 0,38] [%emt 0:00:10]} 31... Ne5
{(Nf6) [%eval -68,25] [%emt 0:02:00]} 32. Rd5 {[%eval 0,36] [%emt 0:01:09]}
32... Re6 {[%eval -62,27] [%emt 0:00:14]} 33. Kg1
{(Qb3) [%eval 0,37] [%emt 0:00:45]} 33... Nf4 {[%eval -89,25] [%emt 0:01:33]}
34. Rc5 {(Kh1) [%eval 0,40] [%emt 0:01:23]} 34... b6
{(Kh7) [%eval -106,26] [%emt 0:01:38]} 35. Rc2
{(Rb5) [%eval -82,35] [%emt 0:01:52]} 35... Rd7 {[%eval -113,25] [%emt 0:01:01]}
36. h4 {[%eval -90,35] [%emt 0:00:01]} 36... Kh5
{[%eval -116,29] [%emt 0:02:26]} 37. Kh2 {[%eval -119,40] [%emt 0:00:00]} 37...
Rd3 {[%eval -115,30] [%emt 0:03:00]} 38. Qc1 {[%eval -133,42] [%emt 0:00:37]}
38... Kxh4 {[%eval -117,30] [%emt 0:01:50]} 39. Rc3
{[%eval -123,43] [%emt 0:00:00]} 39... Kh5 {[%eval -132,31] [%emt 0:02:49]} 40.
Rxd3 {[%eval -151,43] [%emt 0:04:33]} 40... Nfxd3
{[%eval 35534,32] [%emt 0:07:08]} 41. Qc8 {[%eval -171,40] [%emt 0:00:00]} 41...
Rf6 {[%eval -142,29] [%emt 0:00:20]} 42. Qe8+ {[%eval -177,38] [%emt 0:01:43]}
42... Kh6 {[%eval -142,32] [%emt 0:00:22]} 43. Qh8+
{[%eval -173,38] [%emt 0:02:15]} 43... Kg6 {[%eval -142,33] [%emt 0:01:29]} 44.
Kg3 {[%eval -181,38] [%emt 0:01:05]} 44... g4 {[%eval -142,29] [%emt 0:00:59]}
45. fxg4 {[%eval -161,37] [%emt 0:00:52]} 45... Re6
{(Kf7) [%eval -148,28] [%emt 0:00:15]} 46. Qh5+ {[%eval -193,34] [%emt 0:03:52]}
46... Kf6 {[%eval -158,30] [%emt 0:01:30]} 47. Qf5+
{[%eval -197,36] [%emt 0:00:00]} 47... Ke7 {[%eval -167,26] [%emt 0:00:50]} 48.
Qg5+ {[%eval -206,36] [%emt 0:00:00]} 48... Kf7 {[%eval -167,28] [%emt 0:00:19]}
49. Qd8 {(b3) [%eval -214,36] [%emt 0:02:15]} 49... Nc5
{[%eval -165,26] [%emt 0:00:22]} 50. Qd5 {(Qd4) [%eval -218,37] [%emt 0:02:02]}
50... g5 {(Ke7) [%eval -182,29] [%emt 0:00:36]} 51. Kh3
{[%eval -195,36] [%emt 0:00:39]} 51... Ke7 {[%eval -191,30] [%emt 0:00:33]} 52.
Qd4 {[%eval -214,37] [%emt 0:00:13]} 52... Nf7 {[%eval -198,29] [%emt 0:00:38]}
53. e5 {(Qb4) [%eval -260,36] [%emt 0:01:22]} 53... Nxe5
{[%eval -198,29] [%emt 0:00:33]} 54. Kg3 {(Qe3) [%eval -284,35] [%emt 0:00:26]}
54... Nf7 {[%eval -217,29] [%emt 0:01:28]} 55. Qb4
{[%eval -286,39] [%emt 0:00:00]} 55... Nd6 {[%eval -225,30] [%emt 0:00:34]} 56.
Kh2 {(Qd4) [%eval -294,42] [%emt 0:00:06]} 56... Nd7
{[%eval -241,28] [%emt 0:00:27]} 57. Qd2 {(Kg3) [%eval -323,42] [%emt 0:03:40]}
57... Re5 {[%eval -249,27] [%emt 0:00:36]} 58. Qd3
{[%eval -339,36] [%emt 0:00:03]} 58... Ne4 {[%eval -258,29] [%emt 0:00:22]} 59.
Qa3+ {[%eval -355,36] [%emt 0:00:43]} 59... Ke6 {[%eval -258,29] [%emt 0:00:08]}
60. Qa8 {(Qb3+) [%eval -383,37] [%emt 0:03:09]} 60... b5
{[%eval -260,28] [%emt 0:01:41]} 61. Qc8 {(Kg1) [%eval -402,40] [%emt 0:00:00]}
61... Ke7 {[%eval -261,26] [%emt 0:00:17]} 62. Qb7
{(Qa6) [%eval -410,39] [%emt 0:00:43]} 62... Nd6
{[%eval -270,29] [%emt 0:00:17]} 63. Qa7 {[%eval -410,37] [%emt 0:00:23]} 63...
Ke6 {[%eval -270,30] [%emt 0:00:06]} 64. Qa2+ {[%eval -418,37] [%emt 0:00:36]}
64... Rd5 {[%eval -271,30] [%emt 0:00:12]} 65. Kg3
{(Qa7) [%eval -418,38] [%emt 0:00:24]} 65... Nf6
{[%eval -286,26] [%emt 0:00:29]} 66. Qa3 {(Kf3) [%eval -418,32] [%emt 0:00:00]}
66... Nc4 {(Rd4) [%eval -302,29] [%emt 0:01:52]} 67. Qa6+
{[%eval -476,34] [%emt 0:01:09]} 67... Kf7
{(Ke5) [%eval -315,29] [%emt 0:00:10]} 68. b4 {[%eval -434,28] [%emt 0:00:20]}
68... Rd3+ {(Ne5) [%eval -332,30] [%emt 0:01:27]} 69. Kf2
{[%eval -412,28] [%emt 0:00:08]} 69... Nxg4+ {[%eval -332,28] [%emt 0:00:17]}
70. Kf1 {[%eval -408,30] [%emt 0:00:03]} 70... Rd5
{[%eval -332,29] [%emt 0:02:21]} 71. Ke2 {[%eval 35534,38] [%emt 0:00:00]} 71...
Nf6 {(Nge5) [%eval -333,29] [%emt 0:01:23]} 72. Qb7+
{(Qc6) [%eval -470,26] [%emt 0:00:23]} 72... Kg6
{[%eval -350,25] [%emt 0:00:15]} 73. Qb8 {(Qc7) [%eval -539,34] [%emt 0:00:31]}
73... Rd2+ {[%eval -375,30] [%emt 0:00:58]} 74. Ke1
{[%eval -733,37] [%emt 0:00:03]} 74... Ne4 {[%eval -438,27] [%emt 0:00:32]} 75.
Qc8 {(g4) [%eval -909,37] [%emt 0:00:06]} 75... Kf6
{(Rxg2) [%eval -454,29] [%emt 0:00:29]} 76. Qh8+
{(Qe8) [%eval -1141,35] [%emt 0:00:25]} 76... Kf5
{(Ke6) [%eval -516,29] [%emt 0:00:24]} 0-1
[/pgn]
And some diagrams illustrating why 3 pieces are stronger than queen + couple of pawns:
[d]2r2rk1/1pqn1pp1/2p2np1/pN2p3/PbBP4/1Q2P3/1P3PPP/2RRB1K1 b - - 0 17
Stockfish has just played 17.Nb5, a move that involves trading with an ensuing imbalance of Q+2 pawns vs R + 2Ns. Now, in terms of material only, the Q+2Ns would make 11 pawns, while R+2Ns only 10.5 pawns, so Stockfish, which regards this position rather materialistically, might have thought it has an advantage. Komodo knows this is not so.
[d]6k1/1pr2rp1/5n2/4n1p1/Q7/4P3/1P3PPP/3R2K1 w - - 0 24
A critical position: Stockfish sees itself some half a pawn in advantage, while Komodo sees the same margin for itself. Which one is right? Of course, Komodo. Q+2 pawns are considerably weaker than R + 2Ns. You must simply have this imbalance evaluation for Q vs 3 pieces, otherwise relying just on material would be wrong. Obviously, there is not other reason for the Stockfish score, as otherwise the position is pretty simple.
[d]8/1p2r1p1/2r3nk/6pn/4P3/4QP1P/1P4P1/3R2K1 w - - 0 30
The position at move 29. Stockfish already sees fully equal score, while Komodo sticks with its big advantage. Obviously it takes some time for Stockfish to see things changing, some 12 moves from the unfortunate trade, but the developments took their shape right at move 17.
[d]8/3r2p1/1p2r2k/4n1p1/4Pn2/4QP1P/1PR3P1/6K1 w - - 0 36
6 moves on, at move 35, Stockfish already sees itself losing by a large margin, some 80cps. It took 20 moves for the evaluation to change from positive to very negative, but the fate of the game was sealed at move 17.
[d]8/4k3/1p1nr3/2n3p1/1Q4P1/8/1P4PK/8 b - - 0 56
Some moves on, resistance already becomes futile. The 3 pieces are simply too much for the queen.
As was discussed recently in the Elephantiasis thread, the main reason for the predominance of the 3 pieces over the lonely queen could be the fact that the 3 pieces defend each other very well, so the queen can not attack and capture them, but also defend in a group successfully all own pawns, while the queen side can not defend its pawns, when a couple of pieces concentrate their attack on them. This would be true however, in a real chess game, only for the imbalance of Q vs 3 pieces. 4 minor pieces vs 2 rooks + pawns could also exhibit a similar behaviour favouring the 4 pieces over their material values, but, to be honest, I have encountered such imbalances in real chess games just once or twice in my whole life, if ever, so that this imbalance lacks practical importance.
So, if I were an engine developer, I would concentrate specifically only on the imbalance of Q vs 3 pieces where elephantiasis or the additional strength of the more pieces of lower power has effect. You have to have at least 2 pieces in terms of number more than your opponent in real life chess games, in order for the phenomenon to have effect. And that, as said, happens only in the imbalance of Q vs 3 pieces.
I think everytime you have an imbalance of Q+ pawns vs 3 pieces, you might bravely assign some additional evaluation bonus points to the side with more pieces of lower power. And this would be over the bonus for pair of bishops. It would not matter at that what the 3 pieces are: rooks, knights or bishops. I would only exclude from the rule Q vs 2Ns + B, where already the queen might be stronger. In all other cases, however, the rule should be valid.
I would give bonus for:
2Bs + N
2Rs + B, 2Rs + N
2Ns + R, R+B+N
I think in the standard case the bonus could be some 70cps, in the case 2Bs + N 50cps, while with 2 rooks it might amount to even a full pawn. However, a standard bonus of 70cps in all cases would be just fine and would solve some outstanding problems, not only with Stockfish, but with all other engines that do not know the rule. (Are there actually any other engines apart from Komodo that know the rule?)
I must say that, for posting this particular thread and for the illumination that 3 pieces are stronger than a lonely queen + pawns, I should thank Komodo, for the games I saw it playing in different tourneys (hey Larry, you are not angry with me for making public a secret weapon of Komodo, are you?:) ), Harm Mueller with his 7Ns vs 3Qs imbalance (obviously unrealistic imbalance with fully undecided outcome, but an illumination still), and ...
RJ Fischer.
It was a Fischer game that I replayed that finally convinced me that the majority of pieces vs a lonely queen matters. Thank you Robert!
Any comments very much appreciated.
But I would also appreciate very much a possible implementation of this piece of very simple knowledge into a number of chess engines (so that I do not use this blind spot to win a lot of games

Below the Fischer game:
[pgn][Site "Buenos Aires"]
[Date "1970.??.??"]
[Round "5"]
[White "Quinteros, Miguel"]
[Black "Fischer, Robert J"]
[Result "0-1"]
[Opening "English, Symmetrical Variation"]
[ECO "A35"]
[MLNrOfMoves "42"]
[MLFlags "000000"]
1. c4 g6 2. Nc3 Bg7 3. Nf3 c5 4. d4 cxd4 5. Nxd4 Nc6 6. Nc2 Bxc3+ 7. bxc3 Nf6 8.
f3 d6 9. e4 Be6 10. Be2 Rc8 11. Ne3 Qa5 12. Bd2 Ne5 13. Qb3 Nfd7 14. f4 Nc5 15.
Qc2 Nc6 16. O-O Qa4 17. Qb1 Na5 18. e5 dxe5 19. fxe5 O-O 20. Rf4 Nd7 21. Nd5
Rfe8 22. Qe4 b5 23. Rxf7 Kxf7 24. Qh4 Kg7 25. Nxe7 Qc2 26. Qh6+ Kh8 27. Rc1
Qxc1+ 28. Bxc1 Rxe7 29. cxb5 Nc4 30. Qh4 Rf7 31. Qd4 Kg8 32. Bf4 Rc5 33. Bf3
Rxb5 34. h3 Ncxe5 35. Ba8 Rf8 36. Bxe5 Nxe5 37. Qxa7 Bd5 38. Bxd5+ Rxd5 39. Qe3
Ra5 40. Qe2 Rfa8 41. a4 Nf7 42. h4 Rxa4 0-1
[/pgn]
[d]6k1/p2n1r1p/4b1p1/1Pr1P3/2nQ1B2/2P5/P3B1PP/6K1 w - - 0 33
Q + 2Bs + 3 pawns weaker than 2Rs + 2Ns +B.