correspondence chess in the age of NNUE

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
Posts: 6347
Joined: Thu Aug 18, 2011 10:04 am

Re: correspondence chess in the age of NNUE

Post by Rebel » Sat Mar 13, 2021 11:53 am

lkaufman wrote:
Thu Jan 21, 2021 5:26 am
Now that strong Stockfish NNUE versions have been out for about six months, there should have been lots of correspondence games played by email that have completed during this interval. Based on results we have seen in both normal chess and chess960 when powerful hardware is used with NNUE engines, unless bad openings are mandated, the draw percentage between roughly equal NNUE engines at long time controls is in the high 90s, basically making the game unplayable. So what is the experience of correspondence players using good hardware and NNUE engines over the past six months? I suppose some games are won due to opponents having inferior hardware or not knowing about NNUE or not caring about the result much, but is anyone winning any games from opponents where these factors are not applicable? Is correspondence chess still playable without mandated openings?
@Larry and @others,

I am currently analyzing interesting openings extracted from Millionbase 3.5 and Correspondence 2, together 5.5 million games. I am almost finished the gambits part and moving to popular openings. My question is, what are the popular openings of correspondence players?
90% of coding is debugging, the other 10% is writing bugs.

Thomas A. Anderson
Posts: 27
Joined: Tue Feb 23, 2016 5:57 pm

Re: correspondence chess in the age of NNUE

Post by Thomas A. Anderson » Sat Mar 13, 2021 11:54 am

cdani wrote:
Fri Mar 12, 2021 7:41 pm
Thomas A. Anderson wrote:
Fri Mar 12, 2021 4:42 pm
cdani wrote:
Tue Mar 09, 2021 7:01 pm
lkaufman wrote:
Tue Mar 09, 2021 3:27 am
... The key is that as long as the chosen opening positions are near the edge of the win/draw line, this won't be the case even with a million to one speedup from current hardware.
I hope those positions can be found/provoked by future engines because they understand chess much better than actual ones, and not because they speculate.
Using tools/engines to find test/borderlin postions for the same tools/engines smells like catch22.
I mean, you know that an engine plays statistically. Statistically is different than perfect. Stockfish cannot say if following 30 straigth quiet moves there is something to win because there is not statistict knowledge inside it to be able to follow such moves. A future engine maybe will be able to, so it will find positions that maybe are not wining, but are wining against current stockfish because it cannot understand them.
That's very possible and I used the AlphaZero games heavily to search for positions that SF didn't judge in the same way. But this means using engines/tools to create challanges for different engines/tools.

User avatar
Posts: 2200
Joined: Sat Jan 18, 2014 9:24 am
Location: Andorra

Re: correspondence chess in the age of NNUE

Post by cdani » Sat Mar 13, 2021 12:15 pm

Rebel wrote:
Sat Mar 13, 2021 11:53 am
My question is, what are the popular openings of correspondence players?
I saw everything. Most strong players do whatever is complicated enough to try that the game lasts enough to prove your rival.

User avatar
Posts: 4423
Joined: Tue Jul 03, 2007 2:30 am

Re: correspondence chess in the age of NNUE

Post by Ovyron » Thu Jul 29, 2021 9:26 pm

lkaufman wrote:
Tue Jan 26, 2021 12:12 am
Let's suppose that the rules of chess were changed in some way to eliminate draws while keeping the chances of White and Black fairly equal.
What about my idea of starting with a countdown? Every time a player moves the countdown goes down by 1. If white manage to force a draw it counts as a win. If the countdown hits 0, black wins. 0% Draws.

The idea comes from a Starcraft's Custom Rule map called "Heaven's Last Stand", where Jesus and the other angels are in Heaven and being attacked by Satan and his forces; Satan can create infinitely new demons to be sent to attack, though this takes time, while Heaven has many outposts that need to be defended.

Due to the nature of the game, the outposts of Heaven eventually fall, so Satan keeps always making progress, until he makes it to the last Stand which is indefensible and hell always wins. 3 Players control hell and 4 control Heaven's units. It uses all the units of the Zerg and Terran factions of the game, and some custom ones, so the only way to balance it is to make games shorter (make heaven weaker) or longer (make Hell weaker.)

What they did is adding a countdown that's about "waiting for God", when it hits 0 God arrives and Heaven wins automatically. It has been specifically designed so that when 7 skilled players play together hell is about to win when the clock is reaching 0, with close encounters being really fun, 0% drawn games, and about 50% Heaven/Hell split wins (Hell has the edge since it's given enough time to beat Heaven, but this is balanced because less experienced players tend to play hell so the skilled heaven players can defend long enough for the countdown to hit 0.)

How were these guys able to solve the balance problem so easily, in a much more complex 7-player game with a real time system by just adding a countdown and chess can't?

I have no idea what the countdown should be, but we could be at a point at which beating black is impossible so you switch white's goal to forcing a draw ASAP, and black just needs to defend and keep the game going longer than that.

Post Reply