Open Caleb-Petersen opened 5 years ago
I have deleted the double export and moved the ratings to the per-game and per-move exports. I also made it so that the per-game export has two rows per-game for the statistical analysis side of things (a more accurate reflection of which rating goes with which stats)
I've added a few features, most notably the excel benchmarks
Added graceful cancel that allows the user to save all the analysis completed thus far. Also uncommented the closing of the engine threads.
Added writing of undecided position stats (per player per game) to a spreadsheet with their ratings with the goal being to be able to do statistical analysis to give more precise estimates of the rating of a suspect. If benchmarks with a large sample of clean games are created, given a set of T1/T2/T3/CP loss from a large sample it can be mapped to an expected rating.