How to lose a game with B+R advantage.

[Cubeman]

I was playing a PDA game with two engines Naum 4.2 white on fast PPC against Stockfish on fast ipod, and the following position was reached
[fen]Q7/6k1/1B1p4/3Pp3/5ppq/8/1R4K1/8[/fen]
It is black to move and we get this series of moves 1..Qg3+ 2.Kf1 Qd3+ 3.Kf2 Qf3+ 4.Ke1 Qc3+ 5.Rd2 Qc1+ 6.Ke2 f3+ 7.Kd3 Qb1+ 8.Rc2 Qb3+ 9.Rc3 Qb1+10.Kd2 Qb2+ 11.Rc2 Qb4+ 12.Kc1 Qxb6 13.Qa4 Qg1+ to get the following position.
[fen]8/6k1/3p4/3Pp3/Q5p1/5p2/2R5/2K3q1[/fen]
Stockfish has been showing an = score right from the start of the previous position and Naum has thought that it holds a slight advantage, so doing what a typical human does, it thinks to itself that if I move the King to d2 there is probably a perpetual check and does not want to give black the pleasure so decides to play the move 14.Kb2?? and stockfish wastes no time in playing 14..Qd4+ with a large advantage to Black and ended up winning with the connected passed pawns against the rook.It takes Naum about 70 seconds on my fast PDA to know that 14.Kb2 is a mistake so this would make a good test position for a move to avoid.

[RadioSmall]

I was playing a PDA game with two engines Naum 4.2 white on fast PPC against Stockfish on fast ipod, and the following position was reached
[fen]Q7/6k1/1B1p4/3Pp3/5ppq/8/1R4K1/8[/fen]
It is black to move and we get this series of moves 1..Qg3+ 2.Kf1 Qd3+ 3.Kf2 Qf3+ 4.Ke1 Qc3+ 5.Rd2 Qc1+ 6.Ke2 f3+ 7.Kd3 Qb1+ 8.Rc2 Qb3+ 9.Rc3 Qb1+10.Kd2 Qb2+ 11.Rc2 Qb4+ 12.Kc1 Qxb6 13.Qa4 Qg1+ to get the following position.
[fen]8/6k1/3p4/3Pp3/Q5p1/5p2/2R5/2K3q1[/fen]
Stockfish has been showing an = score right from the start of the previous position and Naum has thought that it holds a slight advantage, so doing what a typical human does, it thinks to itself that if I move the King to d2 there is probably a perpetual check and does not want to give black the pleasure so decides to play the move 14.Kb2?? and stockfish wastes no time in playing 14..Qd4+ with a large advantage to Black and ended up winning with the connected passed pawns against the rook.It takes Naum about 70 seconds on my fast PDA to know that 14.Kb2 is a mistake so this would make a good test position for a move to avoid.

Hi Cubeman :
What was the time control??

[Cubeman]

I was using a 30 second/move with ponder on.Quite a few other engines also prefer Kb2?? over the correct Kd2 which draws. Shredder fails while Hiarcs and Tiger take a little bit of time to decide on Kd2. I suppose engines are taught that it is bad strategy to swap down material when behind so they don't expect the Qd4+ which not only swaps down but leaves double isolated pawns.
It would make a good test for some of the dedicated machines out there.

[RadioSmall]

I was using a 30 second/move with ponder on.Quite a few other engines also prefer Kb2?? over the correct Kd2 which draws. Shredder fails while Hiarcs and Tiger take a little bit of time to decide on Kd2. I suppose engines are taught that it is bad strategy to swap down material when behind so they don't expect the Qd4+ which not only swaps down but leaves double isolated pawns.
It would make a good test for some of the dedicated machines out there.

Hi
Palm Hiarcs 13.1 on the palm Tungsten C takes only 4 seconds to reject Kb2? and play Kd2 .I will put this position on the Novag Sapphire and report the results............

[RadioSmall]

I was using a 30 second/move with ponder on.Quite a few other engines also prefer Kb2?? over the correct Kd2 which draws. Shredder fails while Hiarcs and Tiger take a little bit of time to decide on Kd2. I suppose engines are taught that it is bad strategy to swap down material when behind so they don't expect the Qd4+ which not only swaps down but leaves double isolated pawns.
It would make a good test for some of the dedicated machines out there.

Hi Cubeman :
This is not hard for dedicated computers ,at least not for Novag Sapphire.Sapphire does not even consider Kb2?? for a split second and plays Kd2 all the way...
Radio Small