Page 1 of 2

Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 11:28 am
by Rebel
Fruit-Family - http://rebel13.nl/text/fruit-family.txt
Rybka3-Family - http://rebel13.nl/text/rybka-family.txt

Today - http://rebel13.nl/text/anno-2019.txt
Pretty much healthy situation, computer chess healed itself.

More at - http://rebel13.nl/rebel13/experiments.html#11

Re: Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 12:24 pm
by Laskos
Rebel wrote:
Thu Mar 14, 2019 11:28 am
Fruit-Family - http://rebel13.nl/text/fruit-family.txt
Rybka3-Family - http://rebel13.nl/text/rybka-family.txt

Today - http://rebel13.nl/text/anno-2019.txt
Pretty much healthy situation, computer chess healed itself.

More at - http://rebel13.nl/rebel13/experiments.html#11
Have you tried to put Sf8 and Sf9 there?
Do you have Houdini 6?

Re: Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 1:58 pm
by velmarin
Rebel wrote:
Thu Mar 14, 2019 11:28 am
Today - http://rebel13.nl/text/anno-2019.txt
Pretty much healthy situation, computer chess healed itself.
The list is very short and mostly open_source engines, some though open_source and although the negative test, are clearly inspired or derived from Stockfish.
Simply the mentality towards this question has changed.
The better or worse, is the question.

Re: Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 2:59 pm
by Guenther
Rebel wrote:
Thu Mar 14, 2019 11:28 am

Today - http://rebel13.nl/text/anno-2019.txt
Pretty much healthy situation, computer chess healed itself.
This description sounds quite naive, especially considering the minimal sample.

Re: Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 3:09 pm
by Rebel
Laskos wrote:
Thu Mar 14, 2019 12:24 pm
Rebel wrote:
Thu Mar 14, 2019 11:28 am
Fruit-Family - http://rebel13.nl/text/fruit-family.txt
Rybka3-Family - http://rebel13.nl/text/rybka-family.txt

Today - http://rebel13.nl/text/anno-2019.txt
Pretty much healthy situation, computer chess healed itself.

More at - http://rebel13.nl/rebel13/experiments.html#11
Have you tried to put Sf8 and Sf9 there?
I have a longer version, not much different.

Code: Select all

Positions  8238   Ands  Ethe  Komo  Lase  Lase  Stoc  Stoc  Stoc  Xiph 
{Andscacs 0.93 } ----- 45.52 45.19 44.40 44.97 46.01 47.44 47.40 46.99 
{Ethereal 11.25} 45.52 ----- 46.61 47.58 50.44 50.64 50.47 50.79 50.67 
{Komodo 9.42   } 45.19 46.61 ----- 45.13 46.69 48.66 49.33 48.86 48.32 
{Laser 1.6     } 44.40 47.58 45.13 ----- 53.13 45.42 47.38 46.35 46.22 
{Laser 1.7     } 44.97 50.44 46.69 53.13 ----- 48.82 49.54 49.16 49.04 
{Stockfish 10  } 46.01 50.64 48.66 45.42 48.82 ----- 56.20 58.22 51.63 
{Stockfish 8   } 47.44 50.47 49.33 47.38 49.54 56.20 ----- 57.19 51.21 
{Stockfish 9   } 47.40 50.79 48.86 46.35 49.16 58.22 57.19 ----- 50.53 
{Xiphos 0.5    } 46.99 50.67 48.32 46.22 49.04 51.63 51.21 50.53 ----- 

 1. Engine {Stockfish 10  } and {Stockfish 9   } show a similarity of 58.22%
 2. Engine {Stockfish 8   } and {Stockfish 9   } show a similarity of 57.19%
 3. Engine {Stockfish 10  } and {Stockfish 8   } show a similarity of 56.20%
 4. Engine {Laser 1.6     } and {Laser 1.7     } show a similarity of 53.13%
 5. Engine {Stockfish 10  } and {Xiphos 0.5    } show a similarity of 51.63%
 6. Engine {Stockfish 8   } and {Xiphos 0.5    } show a similarity of 51.21%
 7. Engine {Ethereal 11.25} and {Stockfish 9   } show a similarity of 50.79%
 8. Engine {Ethereal 11.25} and {Xiphos 0.5    } show a similarity of 50.67%
 9. Engine {Ethereal 11.25} and {Stockfish 10  } show a similarity of 50.64%
10. Engine {Stockfish 9   } and {Xiphos 0.5    } show a similarity of 50.53%
11. Engine {Ethereal 11.25} and {Stockfish 8   } show a similarity of 50.47%
12. Engine {Ethereal 11.25} and {Laser 1.7     } show a similarity of 50.44%
13. Engine {Laser 1.7     } and {Stockfish 8   } show a similarity of 49.54%
14. Engine {Komodo 9.42   } and {Stockfish 8   } show a similarity of 49.33%
15. Engine {Laser 1.7     } and {Stockfish 9   } show a similarity of 49.16%
16. Engine {Laser 1.7     } and {Xiphos 0.5    } show a similarity of 49.04%
17. Engine {Komodo 9.42   } and {Stockfish 9   } show a similarity of 48.86%
18. Engine {Laser 1.7     } and {Stockfish 10  } show a similarity of 48.82%
19. Engine {Komodo 9.42   } and {Stockfish 10  } show a similarity of 48.66%
20. Engine {Komodo 9.42   } and {Xiphos 0.5    } show a similarity of 48.32%
21. Engine {Ethereal 11.25} and {Laser 1.6     } show a similarity of 47.58%
22. Engine {Andscacs 0.93 } and {Stockfish 8   } show a similarity of 47.44%
23. Engine {Andscacs 0.93 } and {Stockfish 9   } show a similarity of 47.40%
24. Engine {Laser 1.6     } and {Stockfish 8   } show a similarity of 47.38%
25. Engine {Andscacs 0.93 } and {Xiphos 0.5    } show a similarity of 46.99%
26. Engine {Komodo 9.42   } and {Laser 1.7     } show a similarity of 46.69%
27. Engine {Ethereal 11.25} and {Komodo 9.42   } show a similarity of 46.61%
28. Engine {Laser 1.6     } and {Stockfish 9   } show a similarity of 46.35%
29. Engine {Laser 1.6     } and {Xiphos 0.5    } show a similarity of 46.22%
30. Engine {Andscacs 0.93 } and {Stockfish 10  } show a similarity of 46.01%
31. Engine {Andscacs 0.93 } and {Ethereal 11.25} show a similarity of 45.52%
32. Engine {Laser 1.6     } and {Stockfish 10  } show a similarity of 45.42%
33. Engine {Andscacs 0.93 } and {Komodo 9.42   } show a similarity of 45.19%
34. Engine {Komodo 9.42   } and {Laser 1.6     } show a similarity of 45.13%
35. Engine {Andscacs 0.93 } and {Laser 1.7     } show a similarity of 44.97%
36. Engine {Andscacs 0.93 } and {Laser 1.6     } show a similarity of 44.40%
Do you have Houdini 6?
Nope.

Re: Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 3:19 pm
by velmarin
Rebel wrote:
Thu Mar 14, 2019 3:09 pm

I have a longer version, not much different.
t is clear that with SIM 3.
A and B: Sim 60%
A and C: Sim 60%
But not always B and C are 60%, sometimes very different.

Therefore, a Stockfish 8 derivative is not similar to a Stockfish 10 derivative in most cases.

Re: Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 3:38 pm
by Rebel
velmarin wrote:
Thu Mar 14, 2019 1:58 pm
Rebel wrote:
Thu Mar 14, 2019 11:28 am
Today - http://rebel13.nl/text/anno-2019.txt
Pretty much healthy situation, computer chess healed itself.
The list is very short and mostly open_source engines, some though open_source and although the negative test, are clearly inspired or derived from Stockfish.
Simply the mentality towards this question has changed.
The better or worse, is the question.
The crazy days you got a public spanking when you as a programmer released a version derived from an open source -- and were open about that -- are fortunately over.

Re: Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 3:40 pm
by Rebel
Guenther wrote:
Thu Mar 14, 2019 2:59 pm
Rebel wrote:
Thu Mar 14, 2019 11:28 am

Today - http://rebel13.nl/text/anno-2019.txt
Pretty much healthy situation, computer chess healed itself.
This description sounds quite naive, especially considering the minimal sample.
It's true for the top-10. I don't worry much about H6.

Re: Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 5:39 pm
by Rebel
Rebel wrote:
Thu Mar 14, 2019 3:40 pm
Today - http://rebel13.nl/text/anno-2019.txt
Added latest versions of Fire, Booot, Fizbo, still sane.

Re: Experiment 11 - Similarity between the top engines back then and now.

Posted: Thu Mar 14, 2019 11:13 pm
by Frank Quisinsky
Hello Ed,

http://www.amateurschach.de/common/feob ... ines-3.png

Based on average of 41.614 opening positions for each of the ten engines.
More of such stats are included in Excel development files for FEOBOS opening book.

Not longer important ... FEOBOS is ready, project is closed.
All what I need for testing engines I have.

More important what you do here!!
All the time interesting to see such results.
Thank you for your work!!

I will study in detail on weekend ...
Not a work for two minutes and the next please!

Best
Frank