Search found 62 matches

by fierz
Sat May 19, 2018 7:24 am
Forum: Computer Chess Club: General Topics
Topic: Is cutechess-cli's rating interval 1 sigma or 2 sigma?
Replies: 2
Views: 520

Re: Is cutechess-cli's rating interval 1 sigma or 2 sigma?

Thanks! I was worried a bit about the size of my confidence intervals but if they are 95% then it's not as bad as I thought (I usually use 1 sigma by default in anything I do...)
by fierz
Fri May 18, 2018 3:29 pm
Forum: Computer Chess Club: General Topics
Topic: Is cutechess-cli's rating interval 1 sigma or 2 sigma?
Replies: 2
Views: 520

Is cutechess-cli's rating interval 1 sigma or 2 sigma?

I can't see anything on the rating table that cutechess-cli produces in its documentation. Does anyone know if the +- corresponds to a 1 sigma (68%) or 2 sigma (95%) interval?

cheers
Martin
by fierz
Sat May 12, 2018 7:46 pm
Forum: Computer Chess Club: General Topics
Topic: cutechess-cli?
Replies: 11
Views: 2025

Re: cutechess-cli?

So if I go

-games 2
-rounds 50

then I should get the desired behavior?

I will try as soon as this match is finished... testing takes a long time :-(

cheers
Martin
by fierz
Sat May 12, 2018 12:37 pm
Forum: Computer Chess Club: General Topics
Topic: cutechess-cli?
Replies: 11
Views: 2025

Re: cutechess-cli?

Thanks for all the tips! I got it working with concurrency in the GUI, and also with the -ratinginterval N command I get to see more output in the command line interface version. I have one more question: I'm running a gauntlet match of my engine against 9 opponents, using "work1.pgn" (from Ed Schrö...
by fierz
Fri May 11, 2018 3:28 pm
Forum: Computer Chess Club: General Topics
Topic: cutechess-cli?
Replies: 11
Views: 2025

Re: cutechess-cli?

BTW, is cutechess-cli the best tool for playing engine matches, or is there anything else that is easier/better than cutechess-cli? Cute Chess GUI. You get a tournament table dynamically updated after each game, as well as Elo stats. This is accessed from the Tournament menu -> Results. Here is a 6...
by fierz
Thu May 10, 2018 8:40 pm
Forum: Computer Chess Club: General Topics
Topic: cutechess-cli?
Replies: 11
Views: 2025

Re: cutechess-cli?

Can you post the output you get from cutechess-cli and the commandline you are using? for me cutechess is the best for playing tournaments Sure, cutechess command line: d:\programme\cutechess\cutechess-cli -engine conf=Muse -engine conf=Anmon -engine conf=Hermann -engine conf=Ruffian -engine conf=S...
by fierz
Thu May 10, 2018 4:36 pm
Forum: Computer Chess Club: General Topics
Topic: cutechess-cli?
Replies: 11
Views: 2025

Re: cutechess-cli?

Maybe you are affected by this? https://github.com/cutechess/cutechess/issues/366 Actually no, I get output to the console, or with "> output.txt" or similar I get it into a file, but I get a large file with the statistics in the end instead of just statistics in one file, I was wondering if it was...
by fierz
Wed May 09, 2018 7:25 pm
Forum: Computer Chess Club: General Topics
Topic: cutechess-cli?
Replies: 11
Views: 2025

cutechess-cli?

I just tried using cutechess-cli to run a gauntlet match with my engine, and while it seems to work, I was surprised that I couldn't find any kind of final result of the match anywhere. Did I miss an option, or do I really need to parse the pgn file of the match to get statistics of the match?
by fierz
Wed May 02, 2018 4:23 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: MCTS with minmax backup operator?
Replies: 24
Views: 4425

Re: MCTS with minmax backup operator?

Dear Daniel,

thanks, now it makes a lot more sense to me :-)

I wanted to try MCTS/UCT because of its nice property of not needing an eval, just to learn something, so to remain pure I will try the multi-rollout.

best regards
Martin
by fierz
Wed May 02, 2018 4:38 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: MCTS with minmax backup operator?
Replies: 24
Views: 4425

Re: MCTS with minmax backup operator?

Dear Daniel, I'm sorry to ask again (feeling pretty dumb...) but I just don't get it: with only 6 nodes, and adding a seventh node, you explained that the seventh node is added to the previous average directly, i.e. new average = 1/2 ( old average(of six) + new value) instead of new average = 1/7 ( ...