Couple of points. First, nowhere in my data do you see a conclusion that 2x more time = 4x more Elo. The number of games does not support that. The issue was whether much longer games would get us into that +40 to +100 range several have quoted. The answer is "no". I did run about 1K really long games, but did not bother reporting the results because 1K games is not a lot of information. But had I posted that, it showed no improvement over the +18 (it was +16 when I stopped). So we are probably looking at more sampling noise than actual gain with so few games in either of the longish tests. I do not believe anyone would seriously believe that 2x longer =4x better stuff. Because 2x longer == 2x faster. I am not sure what speed stockfish searches on a cluster node, but Crafty is in the 2-3M range. I have seen speeds of 100M on good hardware. that could be 50x faster for stockfish. which would be 128x more elo if I did my math right, which is absolutely nonsense...Ralph Stoesser wrote:I said nothing about a mystery, nor a government cover up. That originated from your humor. I said (not literally) that it was no surprise to me that he does not want to look deeper into the scaling issue. Or should I say non-issue?Don wrote:My mistake, yes it was something Ralph said. I take back all those bad things I said about you.Daniel Shawul wrote:What a stupid thing to say! If you can't read properly what people posted, just stop posting. I never ever said anything like that..
What I'm saying is that all you do is talk. That makes me believe it won't matter how we do our testing, you would find it flawed and always be able to produce some reason why it's not the way you think it should be.
You claim that Bob mysteriously stopped his test, etc. Of course it was suspicious since it did not match what you expected.
Recheck the threads and maybe you will find something like that in Ralph's post.
I have been running the singular test on Komodo and the results are not so hot after all. It shows as a small net improvement only.
He stated in this thread that he's mainly interested in myth debunking, so it seemed natural that he will not look deeper into the issue "does ttSE scale with longer TC and if yes, how much", after the Elo gain had quadrupled compared to the 5+5 results.
Nothing to get angry about. Finally he can do what he want with his cluster (as long as the government agents do not abort his tests to free up cpu time for their ww4 simulation).
BTW this was _not_ so much about "debunking myths". It was about validating my results. I looked at my code at length, and could not get to even +15, much less +40 to +100. So I could either look at my code until I turn blue, or I could do a very simple change to stockfish to turn ttSE off and see what that does to it. That was the driving force to me. Was my code working as it should, or was the improvement less than what has been claimed? I think the author of Houdini claimed +100 in fact. Which certainly seems wrong, in light of this testing.