how many points will Komodo score vs. GM Lenderman?

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.

How many points will Komodo score vs. GM Lenderman (out of 16)?

Poll ended at Mon May 18, 2020 12:55 am

3.5 or less
0
No votes
4 or 4.5
0
No votes
5 or 5.5
0
No votes
6 or 6.5
0
No votes
7 or 7.5
0
No votes
8 (drawn match)
0
No votes
8.5 or 9
0
No votes
9.5 or 10
0
No votes
10.5 or 11
0
No votes
11.5 or 12
0
No votes
12.5 or more
0
No votes
 
Total votes: 0

Chessqueen
Posts: 893
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: how many points will Komodo score vs. GM Lenderman?

Post by Chessqueen » Mon May 18, 2020 8:21 pm

lkaufman wrote:
Mon May 18, 2020 8:06 pm
Chessqueen wrote:
Mon May 18, 2020 7:49 pm
lkaufman wrote:
Mon May 18, 2020 7:03 pm
Chessqueen wrote:
Mon May 18, 2020 6:04 pm
lkaufman wrote:
Mon May 18, 2020 4:16 pm
https://www.chess.com/news/view/man-vs- ... omodo-2020

Here is the chess.com announcement, with info on how to watch.
That is NOT the direct Link to watch Komodo vs Lenderman on chess.com
Try this, there was already another event on main channel.

https://www.twitch.tv/chesscomevents
Yes it is, but Komodo 14 is NOT doing so great :cry:
No, we're having a tough time. Today is standard Komodo, tomorrow I plan to use MCTS, so we can compare them!
Probably for next match with this type of odds the time control should be no more than 10' + 5" since GM Lenderman is so strong. :roll:

I noticed that a lot of people here saw the poll, but most people did not even bother to guess, I was wondering if you should lower your price of Komodo 14 a little bit lower or $5.00 less than Fat Fritz to be competitive.You will have more volume or number of people ordering at a lower price and the revenue might even be greater for you at the long run https://shop.chessbase.com/en?ref=RF73-K3I1E1BK16

User avatar
Ovyron
Posts: 4010
Joined: Tue Jul 03, 2007 2:30 am

Re: how many points will Komodo score vs. GM Lenderman?

Post by Ovyron » Mon May 18, 2020 8:40 pm

Chessqueen wrote:
Mon May 18, 2020 8:21 pm
I noticed that a lot of people here saw the poll, but most people did not even bother to guess
Polls are broken, nobody was able to cast a vote. People did it in posts.

I didn't because I had no idea for a good guess, and I expected Komodo to do better, but it's possible a Knight is too much no matter what you do.

Chessqueen
Posts: 893
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: how many points will Komodo score vs. GM Lenderman?

Post by Chessqueen » Mon May 18, 2020 8:52 pm

Ovyron wrote:
Mon May 18, 2020 8:40 pm
Chessqueen wrote:
Mon May 18, 2020 8:21 pm
I noticed that a lot of people here saw the poll, but most people did not even bother to guess
Polls are broken, nobody was able to cast a vote. People did it in posts.

I didn't because I had no idea for a good guess, and I expected Komodo to do better, but it's possible a Knight is too much no matter what you do.
You are correct probably in Chess960 with a Knight odds plus a pawn removed from GM Lenderman would have been more challenging for his High level FIDE 2654 or so, and also with a lower time control of 10'+ 5" the match would have created more pressure to GM Lenderman. This time control of 15" + 10' can only be given to lower player in the category of Jorge Sammour FIDE 2458 or lower than 2550 with the same odds. :roll:

PS: Probably AbsoluteFury should ask GM Lenderman if he is willing to take the rest of the games in 10'+5" Time control :mrgreen:

Chessqueen
Posts: 893
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: how many points will Komodo score vs. GM Lenderman?

Post by Chessqueen » Mon May 18, 2020 9:23 pm

lkaufman wrote:
Mon May 18, 2020 2:32 pm
Chessqueen wrote:
Mon May 18, 2020 12:42 pm
Chessqueen wrote:
Mon May 18, 2020 3:25 am
Chessqueen wrote:
Fri May 15, 2020 5:15 am
lkaufman wrote:
Fri May 15, 2020 4:11 am
Chessqueen wrote:
Fri May 15, 2020 3:03 am
lkaufman wrote:
Thu May 14, 2020 5:04 pm
Chessqueen wrote:
Thu May 14, 2020 4:04 pm
lkaufman wrote:
Thu May 14, 2020 12:55 am
Komodo will play GM Alex Lenderman (2634 FIDE, 2642 Rapid) 16 games of Rapid (15' + 10") handicap chess next week on chess.com. All games Komodo plays White with a knight missing (half b1, half g1 except frc). Four games FRC but kings and rooks and castling normal, four games normal but Black cannot castle, and four games Black gives back one pawn (b7,c7,d7,e7, two games each). What will White score? We know from past matches that two White pawns favors Komodo, two Black pawns is close, knight odds favors GM. Latest Komodo on 32 core 3970x Threadripper. If you post your prediction as a reply and get it right (exact score), you can get K14 or sub or next K version 40% off from whatever your price would be otherwise.
Larry you should start the match with the easy games those that are just knight odds either Chess960 or Non castling and as GM Lenderman start beating Komodo he will let his guard down, and that is when you feed Lenderman to your Beast Komodo with the other 8 harder games and Komodo will chew him up little by little and piece by piece. Most people here are NOT taking advantage of this, or they simply try to answer the poll and see that it is not working and do NOT read your entire message. This is what they are missing let me make it clear by making it bigger. If you post your prediction as a reply and get it right (exact score), you can get K14 or sub or next Komodo version 40% off from whatever your price would be otherwise.
The schedule is the same for all days; the first and last game each day are knight for pawn, the two in the middle are FRC and NBC. We're not trying to favor one side or the other by scheduling, just trying to maximize viewer interest.
Not that Carlsen will ever accept to play against an engine Like Komodo, but what would be the maximun odds that Komodo Dragon could give Carlsen at 90' + 30'. Do you believe that Komodo could play Black without its f7 pawn and still draw versus Carlsen ?
Komodo as Black without f7 ("pawn and move") at 90' + 30" vs. Carlsen would be a good match. Most of the games would probably end as draws. I don't know which side would score more wins. Probably assuming there is big money on the line Carlsen would prepare well enough to reach positions where he could try to win with little risk, and he would learn from each game, so I would be more inclined to bet on him to win a long match. At rapid with f7 odds, Komodo as Black without f7 ("pawn and move") at 90' + 30" vs, I would bet on Komodo to win a match.

That is good to know, but since Carlsen is to proud to lose a match against any engine in public at least you can try against MVL or Nakamura at 90' + 30" they will be willing to accept playing against Komodo either as Black without f7 ("pawn and move") at 90' + 30" or Komodo as White without f2 pawn at 90' + 30" at standard time control. Now that would be the greatest achievement to see Komodo winning at Standard time control in the near future.
I do not know why most people here do not take advantage of their chances of winning and getting 40% their purchase of Komodo 14 since the majority of the people here pay 1 dollar to play the lottery where their chances of winning is 10,000 times harder based on the possible combination of 4 numbers. Larry there would have been more participants of this poll if one of the person here would have a chance to get Komodo 14 for FREE regardless if they guess or not in just by participating. Well the best way to watch this match tomorrow will be using Komodo to see what Komodo next move will be against GM Lenderman FIDE rated 2656. :roll:

PS I almost forgot where can we watch this match tomorrow ?
https://www.youtube.com/watch?v=clB8HMCU3Yg
I was observing and noticed that Maxime Vachier-Lagrave might play stronger versus computer engine than Nakamura> he played with even material at 5 minutes games , amazing :roll:
https://www.youtube.com/watch?v=exgVvSauvhQ&t=894s
Not amazing at all; the even material games were played against the crippled levels of Komodo, where it is set to play many hundreds of elo below full strength. That was the point of those games, to show that the Komodo levels below the top one could be beaten by humans.

Sorry I did not noticed that Komodo was playing level handicap. I wonder what odds would GM Lenderman need against Magnus, probably the f2 or f7 pawn. Komodo rating on www.chess.com say 2960 but with Knight odds versus Human GM it is only around 2500 and Knight for pawn probably around 2650 at the most. I strongly believe that Jorge Sammour can take Komodo 14 on your 32 cores with a Knight odds or these same positions that you offered Lenderman and same time control NOT 10'+5" of course not?

lkaufman
Posts: 4081
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: how many points will Komodo score vs. GM Lenderman?

Post by lkaufman » Mon May 18, 2020 10:39 pm

Chessqueen wrote:
Mon May 18, 2020 8:52 pm
Ovyron wrote:
Mon May 18, 2020 8:40 pm
Chessqueen wrote:
Mon May 18, 2020 8:21 pm
I noticed that a lot of people here saw the poll, but most people did not even bother to guess
Polls are broken, nobody was able to cast a vote. People did it in posts.

I didn't because I had no idea for a good guess, and I expected Komodo to do better, but it's possible a Knight is too much no matter what you do.
You are correct probably in Chess960 with a Knight odds plus a pawn removed from GM Lenderman would have been more challenging for his High level FIDE 2654 or so, and also with a lower time control of 10'+ 5" the match would have created more pressure to GM Lenderman. This time control of 15" + 10' can only be given to lower player in the category of Jorge Sammour FIDE 2458 or lower than 2550 with the same odds. :roll:

PS: Probably AbsoluteFury should ask GM Lenderman if he is willing to take the rest of the games in 10'+5" Time control :mrgreen:
Well, we drew the fourth game. It turns out that during the last two games, Nakamura was playing Komodo ON THE SAME MACHINE live on his stream; we didn't even know this was possible. He played ten games even at ten min, no increment, and of course lost ten to zero. So Komodo was playing at effectively half speed for the last two Lenderman games, which is no big deal with 32 cores, maybe fifty elo. But the Naka games show that there's nothing really wrong, it's just a huge handicap.
Komodo rules!

User avatar
Ovyron
Posts: 4010
Joined: Tue Jul 03, 2007 2:30 am

Re: how many points will Komodo score vs. GM Lenderman?

Post by Ovyron » Tue May 19, 2020 1:24 am

That's incredible!

But see, Nakamura is willing to play Komodo, in normal chess, for free, which really puts a question of why the game is being modified with handicaps (it's more interesting for me to see if Nakamura can draw a single game in normal chess than how the knight odds match ends. I follow because it's the only human computer-match available, but apparently not and some "Nakamura is going to play Komodo, do you think he'll draw a single game?" thread of its own would have been interesting.)

lkaufman
Posts: 4081
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: how many points will Komodo score vs. GM Lenderman?

Post by lkaufman » Tue May 19, 2020 1:45 am

Ovyron wrote:
Tue May 19, 2020 1:24 am
That's incredible!

But see, Nakamura is willing to play Komodo, in normal chess, for free, which really puts a question of why the game is being modified with handicaps (it's more interesting for me to see if Nakamura can draw a single game in normal chess than how the knight odds match ends. I follow because it's the only human computer-match available, but apparently not and some "Nakamura is going to play Komodo, do you think he'll draw a single game?" thread of its own would have been interesting.)
Naka did come close to drawing one game, the very first one (in which he played White). It turns out that the reason is that he was playing Komodo with the same settings as Lenderman, which meant a very high Contempt value (130), suitable for the handicap games but way too much for normal chess. It caused Komodo to make an unsound gambit in the opening to open the game, which it would not have done with a setting of 40 or less. After that it had insufficient comp for the pawn and normally would have had to outplay White just to hold a draw, although in a ten min game Naka could not hold even with the advantage. Ironically, Contempt, which is intended to avoid draws, could have caused one in this game due to being set too high for normal chess. I agree that this was interesting, I've wondered myself if Komodo could really go 10 to 0 vs. Naka, especially effectively giving ten to five minute time odds. But most of the games were very one-sided crushes by Komodo; a couple were premature resignations by Nakamura just as things started to go wrong. For a match to be interesting, there has to be some drama, the games can't be too one-sided. Whether that should be done by material odds or by huge time/core count/color/draw odds is a personal preference.
Komodo rules!

Cornfed
Posts: 20
Joined: Sun Apr 26, 2020 9:40 pm
Full name: Brian D. Smith

Re: how many points will Komodo score vs. GM Lenderman?

Post by Cornfed » Tue May 19, 2020 1:54 am

"Odds" is little more than theater for those who might like weird spectacles. My opinion of course.

"Time odds" - now that is the only thing of interest to me because both sides are at least playing the 'same game' with the 'same pieces'...only the machine will always think (?? okay, calculate) faster so 'time needs to be the attempted equalizer'. Not material which changes the game on the board. Why not just not allow the computer to put a single piece/pawn on its 3rd rank...think of the oddities - the human could try to use that to his advantage at various parts of the game. But, again, a gimick.

A Nakamura match of say 15 min/5 sec delay to 2 min for Komodo (pick your difference...depends largely on processor(s) speed at the point), now that would be interesting as both sides are truly playing the same game on the board.

lkaufman
Posts: 4081
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: how many points will Komodo score vs. GM Lenderman?

Post by lkaufman » Tue May 19, 2020 2:03 am

Cornfed wrote:
Tue May 19, 2020 1:54 am
"Odds" is little more than theater for those who might like weird spectacles. My opinion of course.

"Time odds" - now that is the only thing of interest to me because both sides are at least playing the 'same game' with the 'same pieces'...only the machine will always think (?? okay, calculate) faster so 'time needs to be the attempted equalizer'. Not material which changes the game on the board. Why not just not allow the computer to put a single piece/pawn on its 3rd rank...think of the oddities - the human could try to use that to his advantage at various parts of the game. But, again, a gimick.

A Nakamura match of say 15 min/5 sec delay to 2 min for Komodo (pick your difference...depends largely on processor(s) speed at the point), now that would be interesting as both sides are truly playing the same game on the board.
Well, someone on the Stockfish team calculated that for a fair match between latest Stockfish on a good quad vs. Magnus the time odds needed for a fair match (without ponder by SF) would be a ratio of a thousand to one, which means about 500 to 1 for Komodo vs Magnus (or Naka, not much difference in blitz). You can reduce that a lot by giving the human White every game and draw odds, and by a 1 core limit, and by an opening book limit (or none at all). We did one match that way, I'd be fine with doing another. Or Naka could just play Komodo level 24, which is roughly like playing it with game in three seconds.
Komodo rules!

Chessqueen
Posts: 893
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: how many points will Komodo score vs. GM Lenderman?

Post by Chessqueen » Tue May 19, 2020 3:22 am

lkaufman wrote:
Tue May 19, 2020 2:03 am
Cornfed wrote:
Tue May 19, 2020 1:54 am
"Odds" is little more than theater for those who might like weird spectacles. My opinion of course.

"Time odds" - now that is the only thing of interest to me because both sides are at least playing the 'same game' with the 'same pieces'...only the machine will always think (?? okay, calculate) faster so 'time needs to be the attempted equalizer'. Not material which changes the game on the board. Why not just not allow the computer to put a single piece/pawn on its 3rd rank...think of the oddities - the human could try to use that to his advantage at various parts of the game. But, again, a gimick.

A Nakamura match of say 15 min/5 sec delay to 2 min for Komodo (pick your difference...depends largely on processor(s) speed at the point), now that would be interesting as both sides are truly playing the same game on the board.
Well, someone on the Stockfish team calculated that for a fair match between latest Stockfish on a good quad vs. Magnus the time odds needed for a fair match (without ponder by SF) would be a ratio of a thousand to one, which means about 500 to 1 for Komodo vs Magnus (or Naka, not much difference in blitz). You can reduce that a lot by giving the human White every game and draw odds, and by a 1 core limit, and by an opening book limit (or none at all). We did one match that way, I'd be fine with doing another. Or Naka could just play Komodo level 24, which is roughly like playing it with game in three seconds.
Well for tomorrow match you should take your other 16 cores computer just in case Chess.com decides to match it against any other GM and Komodo can have full access to the 32 cores which mean an extra 50 to 60 Elo to Komodo, Plus the MCTS version should be used from now on and hopefully we can expect at least 3 draws tomorrow. :roll:

Post Reply