Hi,
What about the following method for tuning evaluation:
- run through, say, 1000 positions
- generate a (semi-)random set of tuning parameters
- get their eval from your program
- get the eval from the must-be-good program (I'm using stockfish)
- compare those 1000 pairs by using https://en.wikipedia.org/wiki/Pearson_c ... oefficient
- coefficient > previous_coefficient? then remember this tuning parameters set
What do you think?
tuning for the uninformed
Moderators: hgm, Rebel, chrisw
-
- Posts: 7216
- Joined: Mon May 27, 2013 10:31 am
Re: tuning for the uninformed
Something is better then nothing.flok wrote:Hi,
What about the following method for tuning evaluation:
- run through, say, 1000 positions
- generate a (semi-)random set of tuning parameters
- get their eval from your program
- get the eval from the must-be-good program (I'm using stockfish)
- compare those 1000 pairs by using https://en.wikipedia.org/wiki/Pearson_c ... oefficient
- coefficient > previous_coefficient? then remember this tuning parameters set
What do you think?
Are these 1000 positions representative for every position that may occur when playing games. Probably not.
Don't you want to create something different than Stockfish. I already have a copy of Stockfish running on my machine.
But something is better than nothing. (Skipper is nothing). Or not. For instance. Why waste time on something which won't make it. Maybe only for generating better ideas.
Last edited by Henk on Thu Nov 23, 2017 12:52 pm, edited 2 times in total.
-
- Posts: 364
- Joined: Thu Jun 05, 2014 2:14 pm
- Location: Iran
- Full name: Mehdi Amini
Re: tuning for the uninformed
You can use Genetic Algorithm too. See this link for instance:flok wrote:Hi,
What about the following method for tuning evaluation:
- run through, say, 1000 positions
- generate a (semi-)random set of tuning parameters
- get their eval from your program
- get the eval from the must-be-good program (I'm using stockfish)
- compare those 1000 pairs by using https://en.wikipedia.org/wiki/Pearson_c ... oefficient
- coefficient > previous_coefficient? then remember this tuning parameters set
What do you think?
https://www.doc.ic.ac.uk/~nd/surprise_9 ... icle1.html
Farewell.
-
- Posts: 7216
- Joined: Mon May 27, 2013 10:31 am
Re: tuning for the uninformed
Yes he is doing hill climbing now or perhaps only sampling.
-
- Posts: 7216
- Joined: Mon May 27, 2013 10:31 am
Re: tuning for the uninformed
Sampling is ok for tuning two parameters or so. Otherwise it is terribly slow.
O wait if you tune it badly it generalizes better. So it will do better on evaluating unseen positions. Somewhere there is an optimum between bad tuning and 'overtuning'.
Other constraint is that tuning should not cost too much time so better use hill climbing with restart. Or genetic algorithm (with restart?)
O wait if you tune it badly it generalizes better. So it will do better on evaluating unseen positions. Somewhere there is an optimum between bad tuning and 'overtuning'.
Other constraint is that tuning should not cost too much time so better use hill climbing with restart. Or genetic algorithm (with restart?)
-
- Posts: 433
- Joined: Fri Jan 16, 2015 4:02 pm
Re: tuning for the uninformed
I tried pearson correlation in the past, but it's not a good measure. Right now my experiments with evolving the eval parameters use texel tuning on a set of 100K positions (so I can train ~6000 models per day).
In terms of the error rate, I got very close to the hand tuned model (stable version). If you want to see the fully automatic trained eval with almost 0 human intervention check [1] or [2]. The evolved version is 100 Elo weaker than the stable version of Zurichess. I need to add back the pawns cache, though.
[1] https://bitbucket.org/brtzsnr/zurichess ... ew-default
[2] https://bitbucket.org/brtzsnr/zurichess ... ew-default
In terms of the error rate, I got very close to the hand tuned model (stable version). If you want to see the fully automatic trained eval with almost 0 human intervention check [1] or [2]. The evolved version is 100 Elo weaker than the stable version of Zurichess. I need to add back the pawns cache, though.
[1] https://bitbucket.org/brtzsnr/zurichess ... ew-default
[2] https://bitbucket.org/brtzsnr/zurichess ... ew-default
zurichess - http://www.zurichess.xyz
Re: tuning for the uninformed
Yeah I looked at texel tuning but it looked rather complicated. This pearson was implemented in an hour during the morning commute
Re: tuning for the uninformed
If anyone is willing to explain the Texel tuning method tht would be great!
Sofar I understand I have to let it play (well, run QS + eval on FENs) millions of games and then do something with the evaluation-value. But what? I don't understand the wiki explanation.
Sofar I understand I have to let it play (well, run QS + eval on FENs) millions of games and then do something with the evaluation-value. But what? I don't understand the wiki explanation.
-
- Posts: 160
- Joined: Sat Jan 28, 2017 1:29 pm
- Location: The Netherlands
Re: tuning for the uninformed
The basic idea is pretty simple: calculate the error of the evaluation when it is compared to the actual outcome of the positions. Lower a particular evaluation parameter and check if the error has improved, if not, higher the parameter, if again not improved, keep the original value. Do this for all parameters until you have reached the lowest error.flok wrote:If anyone is willing to explain the Texel tuning method tht would be great!
Sofar I understand I have to let it play (well, run QS + eval on FENs) millions of games and then do something with the evaluation-value. But what? I don't understand the wiki explanation.