How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

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.
Post Reply
Chessqueen
Posts: 1445
Joined: Wed Sep 05, 2018 12:16 am
Full name: George Pichard

How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by Chessqueen » Thu Feb 11, 2021 3:59 pm

What is your opinion can the 1450 engine get a draw ?

jefk
Posts: 503
Joined: Sun Jul 25, 2010 8:07 pm
Location: the Netherlands
Contact:

Re: How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by jefk » Thu Feb 11, 2021 4:57 pm

depends on how the 1450 engine is programmed;
it it's like a human, there sometimes will be tactical
mistakes, and then maybe Sf even with rook handicap
still can win, at least occasionally.
With my interest in handicap chess (as extra optional variant eg.
on the servers(*) ) i would be interested to do more research in
material odds vs rating differences; meanwhile mr LK
apparently also is writing a book containing such subjects
:)
(*) it existed during the time of Morphy, but for some or
the other reason, it disappeared; maybe due to the lack
of a consistent 'system' (set of handicap vs elo rules (**).
Nowadays, with Elo and with computers it's a good time
to reconsider such a variant imho (and i continue
to 'work' on it occasionally, also for the Zh variant)

(**) PS, although the input by LK is very useful in this area,
exact calibration won't be necessary for a working system
(not exactly Elo base), because the differences will cancel out,
let's say such an SF with 'only' rook odds would start to
play to a 'low' rated player (1450 is lower club level i
would suggest), and then easily win most of the time;
then it will soon be promoted to a higher class, and
get a higher handicap (eg rook/knight etc). Could be
combined with an indicative 'Elo' (difference) as i suggested
in the Fairchess system, but we shouldn't take such Elo
as serious in normal chess, for the time being, ofcourse.
It's a matter of popularity, if someone wants a better
name than Fairchess, fine with me. A fork from sourceforge could do,
and i still would get some credits which would be 'fair' imho ..
8-)
https://sourceforge.net/projects/fairchess/
PS nope i'm not on github, for the time being sourceforge
is adequate for my purposes, and github a bit complicated
(may change later if i would be crazy enough to work
on handicapped (zh or normal) engines for winboard or so
:lol:
Last edited by jefk on Thu Feb 11, 2021 5:25 pm, edited 1 time in total.
jefk

lkaufman
Posts: 4529
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by lkaufman » Thu Feb 11, 2021 5:15 pm

Chessqueen wrote:
Thu Feb 11, 2021 3:59 pm
What is your opinion can the 1450 engine get a draw ?
I was able to give queen odds with about even results to MacHack in 1967, and it got a USCF rating of about 1530. A few years ago I beat Novag Constellation in a game at rook odds played here, it was rated in the 1800s USCF (maybe 1700s FIDE). So I think it is obvious that the odds-giver will win in your hypothetical pairing. Dragon can give knight odds to engines with ratings like 2200-2300 (depending on time control), so rook odds should be about right for maybe 1900-2000 engines in Rapid games.
Komodo rules!

jefk
Posts: 503
Joined: Sun Jul 25, 2010 8:07 pm
Location: the Netherlands
Contact:

Re: How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by jefk » Thu Feb 11, 2021 5:44 pm

as for names (for reviving chess with material odds as handicap)
now also found this:
https://www.chessvariants.com/other.dir/oddschess.html

gives a list of options, but for the rest not worked out
into a consistent system. Nevertheless i should mention him
(Roger Coop ? never heard of him) i guess in the description of 'fair-chess'
(as well as LK regarding a further need of calibrating rating
diffs vs material odds)
:wink:
had a quick look into the possible subvariants in the fairchess
system, and it doesn't go beyond (>)800 rating diff
so also doesn't include a rook handicap; but i could
make that ofcourse (although ideally, players should play
against each other with relatively little handicaps otherwise
the game would be a bit odd indeed (especially the start)
:lol:
jefk

mhouppin
Posts: 83
Joined: Wed Feb 12, 2020 4:00 pm
Full name: Morgan Houppin

Re: How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by mhouppin » Thu Feb 11, 2021 5:49 pm

FWIW, I ran a minimatch between SF dev and Stash 10.0 (roughly 1650 Elo at CCRL 40/4), time control 3'+2", 1 thread, 256 MB hash, 5-man Syzygy tablebases for SF, I used starting position with one of the 4 rooks removed. Here are the games:



It would seem Rook odds are still too tactical for low rated engines to handle. I may ran another minimatch with a more recent version of Stash to get an idea of where the turning point is.

Chessqueen
Posts: 1445
Joined: Wed Sep 05, 2018 12:16 am
Full name: George Pichard

Re: How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by Chessqueen » Thu Feb 11, 2021 5:53 pm

jefk wrote:
Thu Feb 11, 2021 5:44 pm
as for names (for reviving chess with material odds as handicap)
now also found this:
https://www.chessvariants.com/other.dir/oddschess.html

gives a list of options, but for the rest not worked out
into a consistent system. but i should mention him i guess
in the description of 'fair-chess' (as well as LK regarding
a further need of calibrating rating diffs vs material odds
:wink:

Well, I meant to say with a Queen Odds TC games in 60'+3" or any time control less than what I specify, the reason why I want to know is that GM Nakamura gave the Queen Odds to a 1600 rated player and loss horribly on youtube.com , but we all know that Stockfish and Komodo are at least 500 Elo stronger than Nakamura, so I would like to know what do you believe. Probably a 1550 to 1600 rated engine would be more appropriated, but lets try first against a lower rated engine around 1480 like TTE or any other engine close to 1450 to 1500. :roll:

Chessqueen
Posts: 1445
Joined: Wed Sep 05, 2018 12:16 am
Full name: George Pichard

Re: How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by Chessqueen » Thu Feb 11, 2021 5:54 pm

mhouppin wrote:
Thu Feb 11, 2021 5:49 pm
FWIW, I ran a minimatch between SF dev and Stash 10.0 (roughly 1650 Elo at CCRL 40/4), time control 3'+2", 1 thread, 256 MB hash, 5-man Syzygy tablebases for SF, I used starting position with one of the 4 rooks removed. Here are the games:



It would seem Rook odds are still too tactical for low rated engines to handle. I may ran another minimatch with a more recent version of Stash to get an idea of where the turning point is.
Great Now try a Queen Odds, I believe that a Queen is too much to give away versus Stash 10.0 or TTE :roll: :mrgreen:

jefk
Posts: 503
Joined: Sun Jul 25, 2010 8:07 pm
Location: the Netherlands
Contact:

Re: How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by jefk » Thu Feb 11, 2021 6:09 pm

cq
so I would like to know what do you believe.
as for my current 'beliefs', don't have much about such
material odds (*), and you probably better ask LK about such odds;

Currently i only can suggest that games between players with very large
rating difference can give strange results (and one game probably isn't
enough to give an indication). With such games, and i tried it some times
on my Fair-Chess zillions app with >800 rating diff, you either win in
the beginning due to your large material advantage, or you get
crushed later in a transition between middle-game and endgame.
Not much fun for either player, but with lower material odds,
it might be a good learning experience (for lower rated
players) to get into an endgame and see how it goes (before
being crushed, as often happens in 'normal' chess. :!:
(*) in this little zillions app i only gave a start, a suggestion
for material odds vs rating differences, with the idea
that it might work if things cancel out after many games
apparently in Shogi they have a working handicap system,
and this shouldn't be impossible for chess imho>
PS ok one little word about playing without a queen,
apparently once a certain Robert J Fischer made a remark
about knight odds, but then Nakamura found out that
once you go beyond that, it's not so easy. 8-)
see also eg.
https://www.reddit.com/r/chess/comments ... _handicap/
There is some more info which i found, maybe more about that later
jefk

jefk
Posts: 503
Joined: Sun Jul 25, 2010 8:07 pm
Location: the Netherlands
Contact:

Re: How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by jefk » Thu Feb 11, 2021 6:59 pm

a bit more info about chess handicaps
https://en.wikipedia.org/wiki/Handicap_(chess)
here's a bit more from the historical perspective:
http://www.edochess.ca/batgirl/odds-giving.html

Uri Blass claimed in this forum he can beat most
chess players with queen plus double rook odds.
(but then includes unrated players). well there are|
players who hardly know the rules, so it might be true.

As for ratings vs material odds, it's a better idea to look
at rating ratio (vs difference) so Nakamura vs 1400 is
2.0 ratio, but that's also for a 1200 vs 600 player. Queens odds?
maybe , but that is approx 8.5 points and to get a more
balanced game maybe other material odds also can be
applied (eg. rook plus bishop odds instead of queen).
A matter of trial and error, but then again, accuracy isn't
so important i guess, because i think of a self-correcting
system (of rating ratio's). More simple would be just
differences for rating 'classes' as superGM, GM, IM, NM
A,B,C etc. players
jefk

Chessqueen
Posts: 1445
Joined: Wed Sep 05, 2018 12:16 am
Full name: George Pichard

Re: How many of you believe that Stockfish 12 or Komodo can give a rook Odds to a 1450 engine ?

Post by Chessqueen » Thu Feb 11, 2021 7:34 pm

jefk wrote:
Thu Feb 11, 2021 6:59 pm
a bit more info about chess handicaps
https://en.wikipedia.org/wiki/Handicap_(chess)
here's a bit more from the historical perspective:
http://www.edochess.ca/batgirl/odds-giving.html

Uri Blass claimed in this forum he can beat most
chess players with queen plus double rook odds.
(but then includes unrated players). well there are|
players who hardly know the rules, so it might be true.

As for ratings vs material odds, it's a better idea to look
at rating ratio (vs difference) so Nakamura vs 1400 is
2.0 ratio, but that's also for a 1200 vs 600 player. Queens odds?
maybe , but that is approx 8.5 points and to get a more
balanced game maybe other material odds also can be
applied (eg. rook plus bishop odds instead of queen).
A matter of trial and error, but then again, accuracy isn't
so important i guess, because i think of a self-correcting
system (of rating ratio's). More simple would be just
differences for rating 'classes' as superGM, GM, IM, NM
A,B,C etc. players
I just found out that Stockfish latest version can NOT beat TTE rated around 1480 with a Queen Odds at TC 60'+1" Probably the latest version of Komodo can, since Larry Kaufman gave special instructions to play with Odds, but it has to be tested :roll:

Post Reply