stockfish 10 vs. Mephisto III S Glasgow

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

Moderators: Harvey Williamson, Dann Corbit, hgm

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
User avatar
mclane
Posts: 18351
Joined: Thu Mar 09, 2006 5:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub
Contact:

stockfish 10 vs. Mephisto III S Glasgow

Post by mclane » Fri Apr 05, 2019 9:39 pm

artificial stupidity relying on search versus artificial intelligence of 1984 relying on knowledge.

today we have 2019. thats 35 years of computerchess development.



both 1000 NPMOVE

Mephisto III S was running on a 68000 with 12 Mhz and made 3-5 NPS.
Stockfish 10 limited to 1000 Nodes per move.

mephisto III was developed by thomas nitsche and Elmar Henne and was an attempt of A.I.
It should made very few NPS like a human beeing and still made good moves.

it had 64 KB rom and 16 KB ram. it was 16 Bit and had no hash.
it has an 8 bit brother having 32 KB rom and 8 kb ram doing arround 1-3 NPS.

mephisto III got Level 6 = 40/120 time control.
Stockfish 10 had arena and 1000 Nodes per move.

mephisto III doing arround 3 searches. stockfish arround 7/8 searches.

here the 1st game:

[Event "Computer Schach Partie"]
[Site "ORION8"]
[Date "2019.04.05"]
[Round "?"]
[White "Mephisto III S Glasgow"]
[Black "Stockfish_10_x64_popcnt"]
[Result "1:0"]
[ECO "B22"]
[Opening "Sizilianisch (Alapin 2.c3), 2...Nf6 3.e5 Nd5 4.d4 cxd4 5.Nf3 Nc6 6.Bc4"]
[Time "21:53:52"]
[Variation "(Alapin 2.c3), 2...Nf6 3.e5 Nd5 4.d4"]
[WhiteElo "2400"]
[Termination "normal"]
[PlyCount "50"]
[WhiteType "human"]
[BlackType "program"]

1. e4 c5 2. c3 Nf6 3. e5 Nd5 4. d4 cxd4 5. Qxd4 Nc7 {-1.63/7 0} 6. Bc4 b5
{+0.78/7 0} 7. Bb3 e6 {-0.69/7 0} 8. Nf3 Nc6 {-1.46/6 0} 9. Qe4 h6 {-0.17/6
0} 10. O-O a5 {-0.82/6 0} 11. a4 b4 {-1.43/6 0} 12. Rd1 bxc3 {+0.31/7 0}
13. Nxc3 Rb8 {-1.11/6 0} 14. Bc4 Ba6 {-1.95/6 0} 15. Bxa6 Nxa6 {-1.79/7 0}
16. Qe2 Nc5 {-0.12/7 0} 17. Be3 Nb3 {-0.94/7 0} 18. Rab1 Rb4 {-1.25/7 0}
19. Qc2 Be7 {-0.57/7 0} 20. Nb5 O-O {+0.54/6 0} 21. Rd3 Rxa4 {-3.05/6 0}
22. Rxb3 Nb4 {-2.87/6 0} 23. Qe4 Qb8 {-4.78/6 0} 24. Nc3 d5 {-9.62/6 0} 25.
exd6 Qxd6 {-9.38/7 0 Schwarz gibt auf} 1:0

where is the progress ?
the progress mainly relies on hardware.

when in 1984 Mepihsto III S made arround 3-5 NPS the opponent Brute Force engines such
as kittingers Constellations and SuperConstellations did arround 1.500 NPS vs. 3-5 NPS.

today on a slow PC stockfish 10 does arround millions of NPS. lets say 3.000.000 NPS.

so the factor between 6502 4 mhz from 1984 and todays i7 is nearly 2000 !!!!!!

but still Mephisto III S with its artificial intelligence wins the race.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....

Raphexon
Posts: 341
Joined: Sun Mar 17, 2019 11:00 am
Full name: Henk Drost

Re: stockfish 10 vs. Mephisto III S Glasgow

Post by Raphexon » Fri Apr 05, 2019 9:54 pm

Stockfish latest dev

VS

Herman 2.8

https://ccrl.chessdom.com/ccrl/4040/cgi ... 8%2064-bit

1000 nodes per move and 1 thread each:

Hermann easily stomps Stockfish 130 - 7 - 24.

Here's a link a pgn with 162 games:
http://www.mediafire.com/file/1uvs1jt7s ... s.pgn/file

But if the game had been 10 000 nodes per move, or 100 000. SF would have easily crushed Herman 2.8
Just because SF is rather weak at extremely low node-count doesn't mean its software is worse than Mephisto. It's not optimized for it, that's why a weak engine like Herman can also beat SF at 1000 nodes.

dkappe
Posts: 573
Joined: Tue Aug 21, 2018 5:52 pm
Full name: Dietrich Kappe

Re: stockfish 10 vs. Mephisto III S Glasgow

Post by dkappe » Fri Apr 05, 2019 10:00 pm

You are at the low end of nodes, so you have a fair amount of noise.

I don’t follow the logic on why you picked 1000 nodes. The advances in ab engines is more around search and forward pruning. Sf10 has a trade off in smarts vs speed. As an exercise, see where Mephisto and sf perform about equal over 400 games.

User avatar
mclane
Posts: 18351
Joined: Thu Mar 09, 2006 5:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub
Contact:

Re: stockfish 10 vs. Mephisto III S Glasgow

Post by mclane » Fri Apr 05, 2019 11:06 pm

35 years of software development


...







Or is it 35 years of hardware development ?
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....

koedem
Posts: 91
Joined: Fri Mar 18, 2016 9:45 pm

Re: stockfish 10 vs. Mephisto III S Glasgow

Post by koedem » Fri Apr 05, 2019 11:31 pm

As for progress in engine software, nice reddit post: https://www.reddit.com/r/chess/comments ... velopment/
So you're saying that Mephisto had low nps despite other engines having way higher nps at the time? So essentially it just does way more in it's nodes, that's like calculating the worth of 100000 nodes and then saying "Oh, but I only looked at 1000 positions explicitly, therefore it should only count for 1000 nodes."

Also, if you're saying that the hardware difference is x2000 then why not make those the time controls? Let Mephisto play 40/120 and SF play 40/0.06.

User avatar
MikeB
Posts: 4500
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: stockfish 10 vs. Mephisto III S Glasgow

Post by MikeB » Sat Apr 06, 2019 2:28 am

mclane wrote:
Fri Apr 05, 2019 9:39 pm
artificial stupidity relying on search versus artificial intelligence of 1984 relying on knowledge.

today we have 2019. thats 35 years of computerchess development.



both 1000 NPMOVE

Mephisto III S was running on a 68000 with 12 Mhz and made 3-5 NPS.
Stockfish 10 limited to 1000 Nodes per move.

mephisto III was developed by thomas nitsche and Elmar Henne and was an attempt of A.I.
It should made very few NPS like a human beeing and still made good moves.

it had 64 KB rom and 16 KB ram. it was 16 Bit and had no hash.
it has an 8 bit brother having 32 KB rom and 8 kb ram doing arround 1-3 NPS.

mephisto III got Level 6 = 40/120 time control.
Stockfish 10 had arena and 1000 Nodes per move.

mephisto III doing arround 3 searches. stockfish arround 7/8 searches.

here the 1st game:

[Event "Computer Schach Partie"]
[Site "ORION8"]
[Date "2019.04.05"]
[Round "?"]
[White "Mephisto III S Glasgow"]
[Black "Stockfish_10_x64_popcnt"]
[Result "1:0"]
[ECO "B22"]
[Opening "Sizilianisch (Alapin 2.c3), 2...Nf6 3.e5 Nd5 4.d4 cxd4 5.Nf3 Nc6 6.Bc4"]
[Time "21:53:52"]
[Variation "(Alapin 2.c3), 2...Nf6 3.e5 Nd5 4.d4"]
[WhiteElo "2400"]
[Termination "normal"]
[PlyCount "50"]
[WhiteType "human"]
[BlackType "program"]

1. e4 c5 2. c3 Nf6 3. e5 Nd5 4. d4 cxd4 5. Qxd4 Nc7 {-1.63/7 0} 6. Bc4 b5
{+0.78/7 0} 7. Bb3 e6 {-0.69/7 0} 8. Nf3 Nc6 {-1.46/6 0} 9. Qe4 h6 {-0.17/6
0} 10. O-O a5 {-0.82/6 0} 11. a4 b4 {-1.43/6 0} 12. Rd1 bxc3 {+0.31/7 0}
13. Nxc3 Rb8 {-1.11/6 0} 14. Bc4 Ba6 {-1.95/6 0} 15. Bxa6 Nxa6 {-1.79/7 0}
16. Qe2 Nc5 {-0.12/7 0} 17. Be3 Nb3 {-0.94/7 0} 18. Rab1 Rb4 {-1.25/7 0}
19. Qc2 Be7 {-0.57/7 0} 20. Nb5 O-O {+0.54/6 0} 21. Rd3 Rxa4 {-3.05/6 0}
22. Rxb3 Nb4 {-2.87/6 0} 23. Qe4 Qb8 {-4.78/6 0} 24. Nc3 d5 {-9.62/6 0} 25.
exd6 Qxd6 {-9.38/7 0 Schwarz gibt auf} 1:0

where is the progress ?
the progress mainly relies on hardware.

when in 1984 Mepihsto III S made arround 3-5 NPS the opponent Brute Force engines such
as kittingers Constellations and SuperConstellations did arround 1.500 NPS vs. 3-5 NPS.

today on a slow PC stockfish 10 does arround millions of NPS. lets say 3.000.000 NPS.

so the factor between 6502 4 mhz from 1984 and todays i7 is nearly 2000 !!!!!!

but still Mephisto III S with its artificial intelligence wins the race.
Interesting experiment.

I have performed extensive testing with SF at low nodes - a 1000 total node searched for SF is about the level of a Class B player - a 64 NPS search on the other hand SF will play about GM level or more like higher. in a two minute plus 2 sec game ( if you have SF olay at a 64 NPS setting) SF will play at a GM level ( and see about 400K nodes in total). There is no other engine, that I have seen, that gets that good that fast. You can test any engine from the 80's at 400K node search and they will play nowhere near the level of current SF using 400K nodes. (Note: I said "80's")

McCain X2 ( a stockfish derivative = X2 pronounced "ten two" ) was released this week and it has predefined UCI settings for World Champion to Novice. See screenshot below. Make sure you have the UCI_LimitStrength box checked.

Scroll down on the link below to see the binaries available for download - all flavors available - windows, Linux and macOS. A huge callout to Dann Corbit, Lucas Monge and John Stanback for providing the binaries.

https://github.com/MichaelB7/Stockfish/releases/tag/X2
Attachments
Screenshot 2019-04-05 22.23.25.png
Screenshot 2019-04-05 22.23.25.png (125.11 KiB) Viewed 4124 times
Image

MikeGL
Posts: 964
Joined: Thu Sep 01, 2011 12:49 pm

Re: stockfish 10 vs. Mephisto III S Glasgow

Post by MikeGL » Sat Apr 06, 2019 3:29 am

I think I have seen Mephisto engines ported into UCI engines which can run on x86 machines.
Why not let Mephisto III S Glasgow run in a faster hardware (via emulation) to get higher NPS and adjust SF10 with a similar speed with that of emulated Mephisto. I am sure Mephisto at higher NPS would be crushed by SF10. The point was already made above by one poster, both engines and their algo including their code, were optimised for their hardware.

With your setting, you're telling SF10 to stoop down to Mephisto's low level, how about we change it by letting Mephisto level-up by using faster hardware.
Like a heavyweight boxer doing a diet to fight on lightweight division, how about we make lightweight boxer heavier and let it fight in heavyweight division.
I told my wife that a husband is like a fine wine; he gets better with age. The next day, she locked me in the cellar.

User avatar
mclane
Posts: 18351
Joined: Thu Mar 09, 2006 5:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub
Contact:

Re: stockfish 10 vs. Mephisto III S Glasgow

Post by mclane » Sat Apr 06, 2019 8:34 am

My point was that higher nps is not needed for chess.
Because there are ways to rely on knowledge instead of search.

Why should I now increase nps ???

Of course Stockfish is the master of bean counting.
I have no doubt that it will Win.

But my point was that all those nps are irrelevant if you want to play chess and UNDERSTAND what you are doing.
Stockfish is not understanding what it is doing. It is more a pocket calculator.
It outcomputes all possible moves with these high amount of calculation.

In earlier years fritz was such an example of search concentrated chess program. And in the years of dedicated chess computers
It was mm2 .
What has changed is the speed of the hardware used.

At least factor 2000 from the 6502 days.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....

Frank Brenner
Posts: 32
Joined: Sat Jul 02, 2016 11:47 am

Re: stockfish 10 vs. Mephisto III S Glasgow

Post by Frank Brenner » Sat Apr 06, 2019 11:16 am

I don't think you understand the central facts in the theory of chess programming:

If you measure the strength of Elos starting from 1000 nodes /move and then
always doubles the nodes/move, then the playing strength of Stockfish after 20 doubles increases to about 3500 ELO
while the playing strength of Mephisto 3 converges against a value somewhere between 1700-1900 ELO

Mephisto 3 is a trivial program that today can be programmed by the most talented programmers on a weekend.

jp
Posts: 1412
Joined: Mon Apr 23, 2018 5:54 am

Re: stockfish 10 vs. Mephisto III S Glasgow

Post by jp » Sat Apr 06, 2019 11:28 am

mclane wrote:
Sat Apr 06, 2019 8:34 am
But my point was that all those nps are irrelevant if you want to play chess and UNDERSTAND what you are doing.
You just have strange beliefs about computers "UNDERSTANDING" chess. Mephisto does not understand chess.

Post Reply