[Flame]

This is abit beyond what we suppose to suggest here, yet i think its a good idea.
As i understand Gib is working on a loop begining from the higher card he can play, for example when he suppose to follow a spade lead holding KJ75 he will first check what happend if he play the K then the J then the 7 then the 5.
If there was no time limit it wouldnt matter which direction we pool but as it is, the looping from above is usually right, but sometimes its much better to loop from the buttom, this is when we are going to lose the trick, this is obvious from a human, when someone leads an ace i will loop from the bottom, Gib sould do the same, this will save him tim and mistakes that happend due to not considering all options while he lose an unnecesary high card.
So i suggest you will make a simple preloop procedure checking who is going to win this trick and get into the loop from the right side.

[barmar]

It would be nice if it used some heuristics to tell that all the choices are close enough that it shouldn't waste time analyzing. I've see GIB stop to think about which card to play from a holding like 642.

[1eyedjack]

Perhaps I am just being stupid but I don't understand why it would make any difference whether you would loop from the bottom or from the top. GIB may happen to arrive at the correct answer earlier in the analysis using one loop rather than the other, but it would not be sure that it has arrived at the correct answer until it has completed the analysis, only after which it would display the conclusion, and the complete analysis would take the same time whichever direction the loop went. Or wouldn't it?