Purpose of correspondence chess in modern era

[jefk]

interesting proposition, 'Montecarlo' :

Yes the time odds i proposed are not balanced but this is only
to show (if i could win) 1) that we are not yet at the 'end'
of the 'drawing curve' (increasing Elo with diminishing returns);
and 2) that White really still has some (often meaningful)
advantage in correspondence chess...

NB the one week max time for me as i suggested is only a
maximum sometimes in ICCF you also can take a week sometimes
(after which you often get in 'double time applying' rule).
And nope, at the moment i'm not playing other (corresp) games,
but this may come again this year, which is why i suggested this
maximum (approx 1 week), depending also on your hardware
(see further below)

With software, the 'engine' the differences are smaller, the
current SFNnue (as i suggest) is good enough i think, although
there are quite some derivatives (and alternatives as Lco
or Kom/Dragon), with impressive names as Blackdiamond,
Cfish extended, etc. So NO, you don't have to menion
the software to me. But again, i ask not to upgrade for
this 'test'game the software nor the Nnue (whereby ,
eg. as to compensate for my slower hardware, will be
able to do so (no big deal as i don't expect big quantum
jumps like the recent ones with the neural networks).

As hardware, not knowing who you ('montecarlo') are, i presume you're
not so rich that you can use whatever hardware (or cloud), just a
decent comp would do. My hardware is quite modest not Bmi,
just Sse, and only 8 threads (but effectively only 4) just a simple
quad a few years old (compromise between bucks and speed).
So yes, preferably i would like to know your hardware otherwise i
will certainly stick to max one week per move for me (whereas in
your case 1 or days for posting the move, after you've run the engine
for one (with fast hardware) or two hours (with 'decent' hardware).

From above, i suspect we can reach a compromise.
So i'll start with 1.d4 (yes i know about the updated
LK repertoire, but nevertheless

Good luck and have fun ;
NB it's just a testgame, no betting contest, and although i
think to have decent chances, my life (proverbial) nor my
'honour is depending on it, and if you maintain a draw you
at least will improve my repertoire again a bit
But if you maintain a draw, which certainly is not impossible
(i give you some 40 pct chance) then i'll probably quit with|
correspondence chess, until a new variant with new
'drawing' rules is setup by the Iccf...
PS in the endgame not 50 move rule (as with ICCF (*) ), and Tb's
may apply (als now the 7 men tables) for adjudication
(*) usually doesn't make much difference i noticed last year(s)

[MonteCarlo]

Sounds good! I'm quite interested in how this goes.

For hardware reference, from start position I get ~8 Mnps for traditional SF, ~4 Mnps for SF NNUE, and ~7 knps for LC0 with 384x30.

I'll keep logs for posting afterwards to keep me honest. I'm not a great chess player, but depending on your reference point, I'm not bad either, so I may have some temptation to fiddle with move choices.

I'll keep that nasty impulse restrained, though (not that I think it would help much, since I'm not an experienced centaur).

I've started a thread in the matches board so the game back and forth doesn't clutter up this thread: http://talkchess.com/forum3/viewtopic.php?f=6&t=76469

Cheers!

[jefk]

Montecarlo:

from start position I get ~8 Mnps for traditional SF, ~4 Mnps for SF NNUE, and ~7 knps for LC0 with 384x30.

that's quite a decent machine indeed, the Gpu for Lco could be bigger,
but nevertheless, your Cpu system is already about 20 times faster than mine,
and i don't use Gpu (once tried Lco with Cpu and although interesting it
was very slow). One hour would be enough for a move i guess
but it's up to you (to use another hour if you like); after one hr
it will respond anyway with 1...Nf6 or 1...d5 (running a
few hrs more won't make a difference btw).

So thx for posting the thread in the 'matches' section, awaiting your move,
cheers as well
jef
PS as for fiddling with the suggest engine move i wouldn't do that anyway considering the speed of your hardware; we didn't talk about book, but
as some wrote in the correspondence chess thread, the latest engines don't need such oldfashioned opening books(*), certainly not with 2 hrs calculation time...
PS2 so the only chance for me for a quick win would be if you
erroneously enter a wrong move (yes it happens in ICCF sometimes)
but then i will allow you to retract it, enter the right move and
continue the game.
(*) even nowadays if they're made from blitz games whatever the beautiful
names eg. Blackmamba vs SugarNNProBMI (or AVX) and the impressive
hardware threadrippers or superclusters i don't think they're so good
(although the statistics may indicate some preferred theory for eg.
the first 20 moves or so).
For the rest no comment here until maybe once i get an opening advantage
in our game in the other (matches) thread; after a few months or so..

[Cornfed]

"Purpose"?
Ultimately: To delude oneself into thinking they are doing something worthwhile.
Then again, I spend much of my time writing poetry so...we pick our poisions.

[jefk]

well you (Cornfed) could also ask what's the purpose of 'normal' chess;
(or poetry, but that's another subject

For professional chess players, chess is their profession
(although only the top, aka superGM's seem to earn enough
(or for the absolute top, quite a lot)
For hobby players, chess is their hobby, whereby using
your brain for analysis is beneficial both for kids/youngsters,
as for elderly; for the adult in between group, it has an element
of competition, but also social element (despite the high difference (*) in
playing ability and an element of distraction; some people like to learn
a new language, to broaden their view of the world; others
like games, or computer games (especially during corona times)

For me the various chess disciplines are connected anyway;
In the test game i'm playing now with Monte-Carlo, the
sort of endgame will also indicate whether the proposed modification
of draw rules has sufficient effect. Stalemate will not occur,
(although with avoiding this the egtb's would change),
nor 3 move repetition. KKN or KKP is unlikely, i may
get a pawn end of middlegame (or double bishop advantage),
but whether that's sufficient in the endgame, may be doubtful.

(*) chess with handicaps roughly according to rating diffence would make
it more fun for lower ranked players); and... computers then also
again could participate in tournaments. Thereby, in online chess, but
possibly also otb, it will hugely increase the possible 'matches' between
players (usually players are matched by a server within a certain
range of rating difference, eg. from -200 to plus 300 or so; with
handicaps, we could (optionally) remove such restrictions, and you
would find another player much quicker (even more useful in ZH, maybe)
For equally rated players, to compensate for the first move advantage
for White, i now found that removing the f2 pawn is ok, but then also
the c7 pawn for Black (for equally rated play (*)); a new variant (certainly for
most humans) indeed but also with the 'advantage' (depending on your view)
that normal opening knowledge won't apply anymore. For the rest we
(people as LK and possibly myself) can establish possible
handicaps eg. for different playing categories
C/B/A/NM/IM/GM (an old idea for me, in which i believe now even
more after discovering that the awesome game of Shogi also has
such a handicap system (how often it's used in practice, i dunno;
quite often, is my impression).
(**) for ZH such correction would be even more important, otherwise
we might even be able to 'solve' ZH as win for White ...
(by extending quality opening books to a huge size and minimax)

[jefk]

while poetry may be considered as art (although you don't make
lots bucks (*) with it, if that would be your 'purpose' , chess also is a mix
of (mind)sport, art, and science ; just like my idea about handicap chess
its an important idea that in chess it's not about inborn talent but
you can learn, and improve your abilities; as eg see here:
https://www.amazon.nl/Chess-Improvement ... 1785835025

Wherebyin the meantime, believe it or not, you will improve your brain :


And while in this forum we usually discuss (drawish) top engine
chess, i wrote a (chess) book for beginners... (**)

Eagerly awaiting your book about poetry, we also could talk about purpose
of life, and/or philosophy etc but this would become a bit off topic

(*) there's more bucks in song writing (than in poetry), especially
if it's in modern style 'music' (like eg socalled gangstah rap etc_
yep we (unfortunately) also have it here in the lowlands; the two
geniuses of this song made it in about one day (well at least the vid)
made many bucks and it's indeed awesome as you can see :


**) (art, science, etc ) as also demonstrated in this the chess opening book:
https://www.lulu.com/en/en/shop/jef-kaa ... 8r8qz.html
(yep it's written by yours truly indeed, but i don't consider this spamming
considering the thread topic of the purpose of (correspondence) chess )
Indeed it's not poetry, but it got five stars on Amazon (already for the
b/w edition, the above one is in colour), and imho can improve your chess
knowledge in a pleasant way, not gazing at a compute screen, or getting
lousy advise at your chess club (once you got at that level...)

[Uri Blass]

This is intriguing to me. I may be willing to play the other side, but there are a few things I'd want to clarify since I'm not familiar with the details of the referenced claim by Uri Blass.

1) Is the claim that a top level engine can draw with an unassisted 2 hours per move, even if the centaur opponent knows the exact engine build, configuration settings, and hardware specs? I'm not sure I'd be interested in defending that claim, as even at the current level, there are enough examples of positions where they make big mistakes that I suspect it could be led down a bad path by a skilled centaur if the centaur knows exactly the build/specs and has the time to predict its responses.

If, on the other hand, the claim is just that under normal correspondence conditions, where you don't know exactly what engine the opponent is using, or what build of that engine it's using, or the configuration, or the level of search being obtained, top engines are enough to draw a game unassisted, that I think is interesting.

1)Engines are not deterministic with more than one core so even if the claim is that a top level engine can draw when the opponent know the exact engine build then you cannot plan one game to refute it.

2)I know that engines are not perfect but I do not think that it is easy for a skilled centaur to get into one of the positions that engines go wrong even if the engine is deterministic and use a single core.

I think that you may need a lot of tries for it before finding a single winning path and it may be interesting to see how many tries.

[jefk]

basically agree with your points,
mr Blass.

But my testgame is not to be meant as an definitive judgment;
the purpose (for me atleast) is mostly is to show (and look at
again at) the White advantage, now in the age of Nnue.

The method in my current game will be trying to gradually build
up more and more positional advantage (possibly leading to a winning
endgame, or not), think of some sort of 'Steinitz' methode; and not
about occasion incorrect (or more likely sub-optimal) engine moves,

No doubt there are correspondence players who can exploit that,
another method is trying to go for a line in which the engine
is 'sucked' into a worse position; besides that this is
getting harder and harder these days, i don't think i'm good
at that (or don't want to spend so much time), and also
would have big difficulties to win in a game for Black
under similar conditions as our current game, btw.
So need for me to play more games than the current one.

So indeed more games would be needed to find am objective
and definitive answer to 'Carldaman''s quesion, cq test-proposal (**);
personally i already think the best engines (with enough Nps
and hrs ofcourse) can always draw when playing White, and a bit
later-on possibly/probably also with Black (in lines as e4 e5
(Berlin or anti-Marshall (or d4 d5).

Meanwhile the Bogo-Kaan-Indian i'm going to play against MC
will make some history of-course, and later may appear in GM
otb (or online) play with 'normal' time conditions and
possibly some chess books (until they abandon 1...Nf6 (*))

(*) after that, i don't know if i would play 1.e4, 1.d4,
or maybe 1.Nf3 in correspondence chess but i would not
have big illusions anymore about winning with White
(at ICCF >2300 level)
(**) maybe some others would be willing to do that,
otherwise the ICCF system which will continue for
some time anyway i think (with the same rules may
give some indications (lower rated players may simply
use an engine, and be able to draw anyway, thus going
to flood the title system (at least to IM level or so)

[MuguJack]

There are 2 recent threads here on correspondence chess with similar content, so I thought I'd post to this one being more recent and on topic.

All the discussion brings up questions which it seems the forum may have settled previously.
I for one, still suffer the classic purpose of correspondence chess in any era, that being that there is no OTB chess within a day's drive of where I live.

But for the purpose of correspondence chess in the age of superior computer chess, there are some organizations that still prohibit the use of computer analysis, US Chess Federation being one of them. In theory that should be an option to play correspondence chess without being resigned to a NN draw.

I guess there's already a consensus that this doesn't work? I went through a couple of games posted on the "Check Is In the Mail" part of the USCF site, and the ones I did check seemed to have a > 80% concurrence with Stockfish's best move (I didn't keep track as that was not my purpose). I further expect any attempt to identify computer cheaters will be fraught with failure and simply not work.

So I'd ask forum members if any have any experience with correspondence chess where computers are banned, does it work? If there's going to be a lot of cheating then there would be no point in avoiding ICCF. I appreciate any comments, this is why I have avoided trying corr. chess to date.

[jefk]

So I'd ask forum members if any have any experience with correspondence chess where computers are banned, does it work?

well it probably works best if the stakes are not so high (eg. not playing for titles, or so) and the players cooperate with the rules.

Years ago i've done some 'daily chess' games on chess.com, with reasonable
good experience; a bit faster than regular correspondence chess, and
you are allowed to use an opening book/database, but once out of
book you have to analyze yourself (chess.com has anyway some sort
of anti-cheating system, so it probably still works fine).

Then there's lichess.org, here's a discussion about comp use in
correspondence
https://lichess.org/forum/general-chess ... on-lichess

Then years ago there was IECG, but that was discontinued,
and then games were continued on this server;
https://www.chess-server.net

(did a quick search on computer play, without results,
probably they still don't like the use of computers).

Then there are quite some more informal sites like this one
https://gameknot.com/
works fine, but i don't know if there's any anti-cheat system...
There may be some other similar site (also for chess variant,
for the latter there's something like pychess.org or so)

Summarizing, personally i still prefer chess.com, whereby lichess.org
may certainly also be worth a try (only played there rapid and a few blitz),
but for the real correspondence chess you may have to look further.

Good luck!