Houdini vs Houdini_QP_off

[LudiBuda]

I was curious to see how much quiescent postponement adds to Houdini's strength. This feature was discussed by Larry and Vida a year ago and more recently in the 'forbidden' thread. Result of the test is surprising (more games needed to confirm).
Seems that the only somewhat original idea in Houdini search actually reduces ELO by 10

40 moves in 60 seconds

Houdini vs Houdini_QP_off 67 : 74 with 99 draws


I think I just improved Houdini by accident

[Houdini]

Alex,

Other than demonstrating some naivety about engine testing, your post is unacceptable for the same reasons as in the previous discussion.

You have RE'd and binary modified the Houdini executable to produce this result. It is as unacceptable and illegal as your previous interventions.

Please stop this abuse.

Robert

[LudiBuda]

Don't see what's so naive about it. I just posted the test result and said that more testing is needed to confirm it.
It is only an indication that this particular feature doesn't seem to do much for the elo strength.

[LudiBuda]

40 moves in 60 seconds

Houdini - Critter: 58 - 49 with 85 draws

Houdini_QP_off - Critter: 56 - 44 with 92 draws

Not a big difference, but QP obviously isn't the magic bullet.