[Bug-gnubg] How strong players study

[Albert Silver]

The title may sound a bit funny, considering the source, but this is after having spoken in detail with a few very strong players who make extensive use of Snowie to improve their game. The players in question are all professionals BTW. What they do, and that is practically impossible in GNU, except through a great deal of work, is to take a position from an analyzed match they are reviewing and begin modifying the match conditions to see the results. The reason is that at their level, they analyze in great detail the plays and equities according to scores, match length, and cube placement. Having done this, they close the position, and go back to the match to continue. If one does this in GNU, the match is essentially closed, so returning to it is not possible without reloading it. Also, changing the match length or from match to money game is not possible without starting a new match or game, and pasting the position. All this is a lot of work needless to say, making it very impractical. One player also did this to study how small changes affected the evaluation. For example, he would take the same position, and move an inner board point from the 3-pt to the 2-pt. He would do this quite a bit to better understand what was happening and to see what features most influenced the evaluation.

Essentially, what would be needed is a means to analyze a position without disturbing an analyzed match, and be able to easily change the conditions of the contest (match length, match or money, etc.). In fact, when has anyone EVER wanted to modify a position to study and have it screw up the match/game it originated from? If I want to save a position separately, I’ll just go to File à Save à Position.

I would suggest having it done so that players might open the program and immediately set up a position with match conditions and scores without needing to start a new game or match and then modify it from there. I came home from an event with one player, and he immediately set up more than a dozen very different positions from memory from the event to analyze with Snowie. I have been trying to get him to use GNU more, but he balks because of the difficulty of doing the above. He does thing GNU is amazing, and analyzes games with it, but mostly sticks to Snowie when he wants to do real study.

I’m including Neil Kaxaross in this, so that he might provide feedback or commentary on anything I wrote here.

 

                                                                                                                        Albert