Page 1 of 3

Near-random movers

Posted: Wed Feb 14, 2018 3:36 am
by Robert Pope
I think I recall somebody making a series of random-move engines to create a progression of engines on the very low end. e.g. pure random mover, random mover unless mate-in-1 exists, etc. They were built off a solid engine, so they didn't have instability from crashes.

Ring any bells? Does anyone know where I can download them?

Re: Near-random movers

Posted: Wed Feb 14, 2018 5:11 am
by Modern Times
There us a Brutus random mover on CCRL 404. Follow the download link on that page.

http://www.computerchess.org.uk/ccrl/40 ... Brutus_RND

Re: Near-random movers

Posted: Wed Feb 14, 2018 7:38 am
by Werner Taelemans
Robert Pope wrote:Ring any bells? Does anyone know where I can download them?
Take a look at this thread:
http://talkchess.com/forum/viewtopic.ph ... 07&t=62510

Re: Near-random movers

Posted: Wed Feb 14, 2018 2:34 pm
by Robert Pope
Werner Taelemans wrote:
Robert Pope wrote:Ring any bells? Does anyone know where I can download them?
Take a look at this thread:
http://talkchess.com/forum/viewtopic.ph ... 07&t=62510
That's the thread I was thinking about - thanks!

Re: Near-random movers

Posted: Wed Feb 14, 2018 3:33 pm
by AlvaroBegue
I made a version of RuyDos that plays randomly, at the request of someone in these forums.

https://www.dropbox.com/sh/1o3ymy403yy8 ... 0-MUa?dl=0

Enjoy!
Álvaro.

Re: Near-random movers

Posted: Wed Feb 14, 2018 3:52 pm
by flok
Not really random (really not), but with the rating of one: https://vanheusden.com/pos/

Re: Near-random movers

Posted: Wed Feb 14, 2018 8:41 pm
by Volker Annuss
Arminius http://www.nnuss.de/Hermann/Arminius2017-01-01.zip has a random mover built in. When you give him the command line argument --Rnd he is a random mover.

When you give him the command line argument --RndMate he will find every mate in one. When there is no mate in one, he is a random mover.

Re: Near-random movers

Posted: Thu Feb 15, 2018 1:49 am
by Luis Babboni
Soberango still has a non sofisticated evaluation function, in fact just based on material, and when more than one move have the same evaluation and the same priority in move order, chose wich to do it nearly* random.
*: said "nearly random" cause for example in case of bishop, tower or queen, the less distances moves are choosed over biger distances moves. But for example knight moves that not allowed or avoided mate in its horizon distance view and with no capture or capturing the same kind of piece, are choosed by pure random.
This way, and without lost time (or very few at most) cause the way it uses to be random in those cases, near never do the same game.

Re: Near-random movers

Posted: Thu Feb 15, 2018 2:17 am
by Robert Pope
Thanks, I validated my RandMover and RandMater against Arminius. I also made a normal searcher that returned a random number instead of the regular eval. That version scored 600 elo better than a RandomMater. Is that to be expected?

Re: Near-random movers

Posted: Thu Feb 15, 2018 8:40 pm
by Volker Annuss
Robert Pope wrote:I also made a normal searcher that returned a random number instead of the regular eval. That version scored 600 elo better than a RandomMater. Is that to be expected?
Does your version find deep mates? It is known, that a random eval helps the search to find positions with high mobility.