TalkChess.com

Krzysztof Grzelak wrote: ↑Sat Mar 16, 2024 6:07 pm

RubiChess wrote: ↑Sun Mar 10, 2024 2:55 pm Problem is that todays engines playing on todays hardware and using a book until position is dead draw is boring as hell. Indeed this kind of tournament seems outdated.

Regards.

Then please write what to do to make it not boring chess.

Krzysztof

TCEC and CCC already have everything needed. Strong and equal hardware for the engines (besides the Leela ratio "problem"), no use of opening books for the engines, biased opening positions leading to interesting games and a chat system for getting in touch with other viewers.

That might all be interesting for the viewers. But it is even more boring for the particiapants, which are not even involved.

hgm wrote: ↑Mon Mar 18, 2024 9:32 am That might all be interesting for the viewers. But it is even more boring for the particiapants, which are not even involved.

They are involved at least at TCEC. Admins are in contact with developers giving feedback about testing, settings of engines and possible problems.

At ics I never did anything more than connecting to server and joining the tournament. Would't call that "being more involved".

Not all participants are so desinterested. And at least you were responsible for setting up your own hardware, and making sure everything would run properly on it.

Krzysztof Grzelak wrote: ↑Sat Mar 16, 2024 6:07 pm

RubiChess wrote: ↑Sun Mar 10, 2024 2:55 pm Problem is that todays engines playing on todays hardware and using a book until position is dead draw is boring as hell. Indeed this kind of tournament seems outdated.

Regards.

Then please write what to do to make it not boring chess.

Krzysztof

Here are some of my wishes to make it less boring. So many people are running all sorts of engine matches and reporting here, making 1000s of games available for download. But are we learning anything from so much data that is generated? Is our understanding of chess or strategic games or AI or computing improving? (I am not an expert, and you may tell me that our knowledge is in fact improving; I want to know.) Are we generating knowledge that is transferable to another domain (other than improving our programming skills)? Or are these just sports events that we watch and come back home and tell our mom X won against Y by 13 points margin? Or sometimes not even watching the event, but just running it on our machines and passing the results through a statistical summariser to tell us who won by how many points? (Maybe I am too cynical. Also, nothing wrong with liking sports events of that kind, or if someone's goal is to beat Stockfish.) So my wish is for improved understanding (ideally) but at least engines that can play good game on low resource, engines that can demonstrate knowledge ( - can we get back hand crafted evaluations?), engines that can help us improve our own knowledge as to how to play better. Or purely to marvel, for entertainment, show us some games where an engine makes positional sacrifices without crunching billions of nodes.