Graham Banks wrote:The improvement I'd most like to see is watching the engines think as it's their move.
I think this is a very nice suggestion. Ideally the evolution of the pv could be watched on the little pv boards, perhaps with some delayed updates to make it comprehensible.
I believe there was a suggestion elsewhere (maybe by Ray or Graham?) that the winning threshold 6.5 be increased. I agree with this; with draws so common and with the engines seeing so far, it's very statisfying to get a few more moves.
The problem is 4 straight hours of moves eval'd to 0.0 by both engines, which seems to occur often.
The TCEC draw rule should prevent that. Five moves with no capture or pawn advance will result in a draw if the absolute values of the scores are less or equal to 0.05. Stockfish seems too optimistic to let that happen much.
For exciting games and less draws:
1. less time - 60'/15" or 90'/30"- or without increment
2. every engine on maximum 8 cores
3. Jeroen Noomen test suites, Silver test suite, Arasan test suite, critical lines and sharp variations of many openings
4. the final should be played on a single core for each engine and should include a 5 minutes blitz match without increment
FriedmannC wrote:For exciting games and less draws:
1. less time - 60'/15" or 90'/30"- or without increment
2. every engine on maximum 8 cores
3. Jeroen Noomen test suites, Silver test suite, Arasan test suite, critical lines and sharp variations of many openings
4. the final should be played on a single core for each engine and should include a 5 minutes blitz match without increment
1. Is good because less draws and MUCH more games.
2. and 4. Is the wrong way to decrease the draw rate.
3. Is stupid because games will be played without opening phase and this is bad because 2 engines can have the same elo but one of them could be 150 elo stronger in the opening phase and couldn't take advantage of his strength. It will also increase the draw rate compared to: the 20 possible opening moves.
Graham Banks wrote:The chatroom could be tidied up a bit too - too much crap goes on in there.
Right, never seen such a dense concentration of trolls in a single place.
This is the reason why I decided to not participate anymore - not that it would matter of course;
there are much better new engines like Andscacs and Nirvana.
Also I don't like the role of a "doormat"
Laser is the third new entrant announced for Season 9 of TCEC. It is a brand new engine, yet it is adding ELO points on a daily basis. In TCEC it will participate with an important update
Hai wrote:Laser is the third new entrant announced for Season 9 of TCEC. It is a brand new engine, yet it is adding ELO points on a daily basis. In TCEC it will participate with an important update