If you want me to plot the "moving average perfect engine", then for an eval threshold of 0.90 the interval of eval from 0.70 to 1.10 will have a slope, as the expected results are linear interpolation of draws and wins for the average. I plotted it here:syzygy wrote:If each point represents the average result starting from several positions (with approximately the same eval), then the graph for the "perfect engine" most likely won't look like a step function. Some positions that are evaluated as 0.40 will be theoretical draws (and will be drawn by a perfect engine in self-play), some positions will be theoretical wins (and will be won by a perfect engine in self-play). So the "perfect graph" will look like those for SF and K.Laskos wrote:Different middlegame positions (moves 20-30) for an interval of eval (say a shown eval of 0.4 means interval [0.3,0.5]). I can do only with these moving averages because I need to collect some sufficient data, which is easy for many ultra-fast games, but hard for only 100 TCEC games. Games are self-play for short ones done here locally, K-SF games for TCEC Superfinal.syzygy wrote:Sorry if I have missed this, but what are your graphs showing exactly?Laskos wrote:In fact, this evening I got curious about TCEC games versus ultra-fast games in Komodo's dependency of expected outcome of the games and the value of the eval.
Does each point correspond to multiple self-plays from one and the same opening position? Or does each point correspond to self-plays from various opening positions, each with (approximately) the same initial evaluation?
(Of course different points on the same line correspond to different opening positions.)
TCEC level is a bit closer now to "perfect", but "artistic regression" still seems to show that at least 800 ELO points above TCEC would be needed to reach the "perfect engine (moving average)".
