Hiarcs' misevaluation of queen vs three minor pieces?

[mackgra]

Having played a games with Hiarcs 11.1 over the playchess server a situation arose from a french defence where Hiarcs exchanged three of its minor pieces for the opponents queen:


[Event "Rated game, 5m + 0s"]
[Site "Engine Room"]
[Date "2007.10.20"]
[Round "?"]
[White "Macgra, HIARCS 11.1 MP UCI."]
[Black "Donmez, Rybka 2.3.2a mp"]
[Result "0-1"]
[ECO "C11"]
[WhiteElo "2479"]
[BlackElo "2579"]
[Annotator "-0.16;0.20"]
[PlyCount "96"]
[EventDate "2007.10.21"]
[TimeControl "300"]

{Rybka 2.3.2a mp: 15.9 ply; 181kN/s Intel(R) Core(TM)2 CPU T5500 @ 1.
66GHz 1662MHz, (2 threads), hiarcsdev071020.ctg, oreDeviceObjects} 1. e4 {B/0 0
} e6 {B/0 0} 2. d4 {B/0 0} d5 {B/0 0} 3. Nc3 {B/0 0} Nf6 {B/0 0} 4. e5 {B/0 0}
Nfd7 {B/0 0} 5. f4 {B/0 0} c5 {B/0 0} 6. Nf3 {B/0 0} Nc6 {B/0 0} 7. Be3 {B/0 0}
Be7 {B/0 0} 8. Be2 {B/0 0} Qb6 {0.20/13 23} 9. Rb1 {(Vd2) -0.16/13 26} O-O {
(f6) -0.03/13 13} 10. O-O {(o/of1) 0.25/13 12} cxd4 {(f6) 0.00/12 2} 11. Nxd4 {
0.19/14 10} Bc5 {-0.03/15 2} 12. Na4 {(Acb5) 0.10/15 7} Bxd4 {-0.02/13 0} 13.
Nxb6 {-0.11/15 9} Bxe3+ {-0.47/15 2} 14. Kh1 {-0.11/15 0} Nxb6 {
(Bxb6) -0.39/15 22} 15. Qe1 {(Fd3) 0.00/13 12} d4 {(f5) -0.48/12 12} 16. Qh4 {
(c3) -0.14/13 13} Nd5 {(f5) -0.66/14 11} 17. Bd3 {-0.35/13 14} h6 {-0.44/14 14}
18. Be4 {(Pf3) -0.60/13 0} Nce7 {-0.71/11 2} 19. Qg3 {(Vg4) -0.37/12 3} b6 {
(Nf5) -0.68/13 5} 20. c4 {(Pf3) -1.01/14 17} dxc3 {-0.89/11 0} 21. bxc3 {
(Pf3) -1.25/14 11} Ba6 {-1.03/15 2} 22. Rf3 {-1.19/13 3} Bc5 {-0.95/14 5} 23.
Qe1 {(Pe1) -1.39/13 7} Rad8 {-0.98/13 3} 24. a4 {-1.46/13 8} Nf5 {
(Ng6) -1.05/14 5} 25. Ra1 {(h3) -1.43/13 6} Bc4 {-1.12/12 5} 26. h3 {
(Fxf5) -1.53/13 0} f6 {(Kh8) -1.12/12 10} 27. a5 {(Fxf5) -1.68/13 12} b5 {
-1.18/10 2} 28. Bb1 {(exf6) -1.73/13 11} fxe5 {-1.52/13 5} 29. Qxe5 {-1.79/14 0
} Nh4 {(Kh8) -1.61/13 1} 30. Rg3 {-2.00/13 5} Rf6 {-1.84/15 4} 31. Be4 {
(Fd3) -2.61/14 11} Bd6 {-1.66/13 0} 32. Qh5 {(Vd4) -2.89/14 9} Nf5 {-2.06/12 1}
33. Bxf5 {-2.92/13 1} exf5 {-2.31/13 1} 34. Rd1 {-3.03/13 3} Bc7 {
(Bxf4) -2.56/13 0} 35. Qf3 {(Pd4) -3.74/13 10} Bxa5 {-2.62/13 2} 36. Qh5 {
(h4) -3.87/13 4} Bc7 {(Nxf4) -2.52/11 0} 37. Rd4 {(Vf3) -4.28/13 7} a5 {
(Bxf4) -3.43/14 4} 38. Rf3 {(Vd1) -3.97/12 2} a4 {(Kh7) -4.05/9 1} 39. Qh4 {
(g4) -5.46/13 8} a3 {(Re6) -4.64/13 2} 40. Qe1 {-4.97/13 4} a2 {-4.66/15 2} 41.
Rd1 {-4.96/13 0} Rf7 {(Rff8) -5.48/15 18} 42. Ra1 {(Pc1) -5.91/13 4} Re7 {
-6.10/14 16} 43. Qc1 {-8.40/14 16} Rde8 {-6.36/14 0} 44. Qd2 {-8.99/13 9} Bxf4
{(Re2) -7.25/13 1} 45. Rxf4 {(Vf2) -6.46/11 1} Re2 {(Nxf4) -7.32/11 3} 46. Qxe2
{(Vc1) -8.87/14 5} Rxe2 {-7.65/11 2} 47. Rxf5 {-8.87/14 0} Nxc3 {
(Rb2) -7.67/10 1} 48. Rf4 {(Pc5) -9.07/12 1} Rb2 {
Macgra,HIARCS 11.1 MP UCI resigns (Lag: Av=0.39s, max=5.0s) -9.01/13 2} 0-1

On move 12 Hiarcs played Na4 to allow the capture of black's queen for three of its minor pieces, which finished with an evaluation of 0.00 after a search of 13 plies.
This is in contrast to what i believe the evaluation should is, which should be around half a pawn worse for white without even having to calculate anything much.

I believe white needs around an extra pawn to stand a chance against the minor pieces and current GM's would be biased against the queen when it is up against three minors (in the opening anyway, with no minors exchanged).

What is interesting is that in the opening the side with the three extra minors commonly has the two bishops, which i believe is an extra half pawn straight away.

What are your views? i would like to hear what mark says on the subject.
Finally I have seen a similar thing happen some months ago when a beta version of Hiarcs played over the playchess server and came off worse.

[mackgra]

As no reply has been made i will have a stab at partially answering this thread.

An interesting experiment would be to have an engine shootout which is an option under fritz GUI.

Hiarcs could play itself say 100 times and then Hiarcs vs other engines and see what results we get.

Might have a go myself if I get the time.


Graham