shrapnel wrote:Took the Plunge ! Got a Discount as I'd bought Houdini 5 earlier.
Figured that whatever my doubts about the strength of Houdini 6, private enterprise should be encouraged.
Also, a computer chess world with only Stockfish and its derivatives would be a very boring one !
Cheers !
That's the right attitude mate!
Indeed, let us know your thoughts - especially regarding playing style vs weaker opps.
david wrote:i ran 1 minute blitz matches, again between houdini 6 and stockfish 8, 8 threads on a ryzen 1800x.
Houdini +136 =547 -117
if anyone can convert into Elo and the corresponding uncertainty limits, i would be grateful.
+8 Elo, +/-14 Elo. These latter are called error margins (or error bars). If you run 4X games, the error bars will be +/-7, so if the ratio of the wins/draws/losses stays exactly the same you will know that H6 is definitely at least one (and perhaps as much as 15) Elo stronger. To narrow the bars to +/-2 you'll need 40X games (i.e. 32,000); around 100,000 for +/-1.
You can run Elostat to get all the stats yourself. You already have it if you have Arena installed, otherwise download it here.
Interesting. I ran some 4 minute blitz games, as before.
Houdini +127 =644 -78 Stockfish 8
so Houdini is net +49 from 849 games, or net 5.8% wins.
this compares with a net +19 from 800 games, or 2.375% net wins at 1 minute time control, and with 1.1765% net wins at a 15 minute time control.
I understand that shows Houdini at +8 Elo, but that is +/- 14 Elo according to Elostat.
So, strange; it appears the difference between Houdini and stockfish 8 varies with time control, and not linearly. However, it isn't clear yet that any of these represent significant differences.
I ran 4000 1 minute games between Houdini 6 and stockfish 8.
Houdini +655 -602 =2743 for a net +53 out of ~4000 games.
Elostat says that this is +4 Elo difference, but the error is +/- 6 Elo.
So i still cannot say it is better than stockfish 8
BrendanJNorman wrote:That's the right attitude mate!
Indeed, let us know your thoughts - especially regarding playing style vs weaker opps.
Well, on my Ryzen System, HT On and using all 16 Threads is probably improving performance/results. Certainly doing no harm.
Testing various Contempt values now.
Not yet tested on my Intel System.
Those who bought the Chessbase release of H6 and are using Ryzen processor should check to see that the installer put the correct versions of the engines into their folders:
The installer puts the UCI engines into the Houdini 6 folder and, at least when I installed, it installed the pext versions of the engine instead of the popc versions. I had to manually go into the installer and pull out the correct versions, copy them to the correct folder, etc.
schack wrote:Those who bought the Chessbase release of H6 and are using Ryzen processor should check to see that the installer put the correct versions of the engines into their folders:
The installer puts the UCI engines into the Houdini 6 folder and, at least when I installed, it installed the pext versions of the engine instead of the popc versions. I had to manually go into the installer and pull out the correct versions, copy them to the correct folder, etc.
Thank you for reporting this. I will immediately inform ChessBase.
I did provide clear instructions for installing the "popc" engine on Ryzen, obviously a mistake was made...
