It's hard to convey the meaning of that probability on the left board. Basically, it's the winning probability before the game actual move, and so it's the probability of Leela best move for that configuration. What I could do is to display something like:snippy wrote:- The probability on the left goban is of the current position and not of the move actually played, isn't it? I would find it much more usefull to display the latter, so you can directly compare what you did with what the bot figures to be right (especially if you played a move that the bot has not displayed as a variation).
Would you find this more meaningful? Or maybe a different wording?goreviewpartner wrote:Move 2 White to play, in the game, white played q4
Value network black/white win probability before q4: 47.6%/52.4%
Value network black/white win probability after q4: 45.4%/54.6%
That is what the first versions of GRP were doing. Until I realized that more often than not, Leela does not select the variation with the higher win rate. Look carefully, you will noticed that the latest variation (by alphabetical order) may have sightly higher win rate, but much lower play-out number. As I understand it, Leela first feeling about those moves are good, but she has not enough time to investigate the follow up moves to be confident enough. And so she discards those possible variations for other variations that are more thoroughly investigated (slightly lower win-rate, but higher confidence). In the end, the alphabetical order truly reflects Leela's best moves order, it's just that Leela does not only consider the value network win rate to make her final decision. So I don't think I should change this.snippy wrote:- Would it be possible to sort the calculated variations (A,B,C,...) by probability? That would simplify things greatly since you could evaluate the possibilities without hovering the mouse over every move.
IIRC, this is also the behaviour of the early versions of GRP (at least that's my favorite behaviour), but L19 readers asked me change this behaviour (the full story is in this thread). Recently, a reddit user asked me if it's possible to have only the best move in blue as well. I plan to make the 3 ways possible with a setting in fact.snippy wrote:- In the same sense, I would much prefer a coloration where (for example) blue meant better than the actual move and red worse then the actual move, instead of >50% and <50%. The latter is just helpful in an even game situation, where the other is helpful in every part of the game. The colotation could even indicate the probabilities like in leela (redder -> higher prob.)
OK for the extra windows (maybe a popup opens automatically on clicking inside the textual area). Those textual area are a hell to deal with resizeable windows anyway, I am not happy with the current result...snippy wrote: - I can't seem to get the text windows with the probs to scroll down. And the textual output is not that easy to read. Perhaps a extra window with the output in table form would be easier and you could even display all the variations inside.
For the table, I guess you mean something similar with what is used inside the Leela GUI ? I need to consider this carefully, because other bots do not provide the same type of information, but I need a solution that is consistent across all bots. I plan to somehow standardize all those data into extra SGF properties. I started already in fact and it's the way the graphs are displayed. When this is completed, displaying that table should be as easy as displaying the graphs.
No problem! I guess in a few years, there will be plenty of "GRP-like" software, but at the moment GRP is more like an experimental software paving the way for "next generation" software. So it's both interesting and important to try on all those ideasnippy wrote:Again, I really enjoy GRP and find it enourmously helpful and don't want my text to sound like nagging. Just that everything that is good, can still be made better
