New Stockfish 1.8 JA release?!

[Hagen]

This version of Stockfish is pretty brutal. I've been playing Chess960 games with it using Shredder Classic 4 and I've been amazed as it's ability to ignore obvious material gain in exchange for long term prospects. I'm guessing this engine is about equal to Rybka 4.

[Damir Desevac]

This engine is by no means close to Rybka 4 in playing strength. In fact I think it is weaker than 1.7.1 version. So before you try and speak too loud about it, why don't you try and test it ? I am doing it right now, and am not impressed by its play. I think the SF team rushed with this version, and should have waited a little, until they have made some decent improvement before releasing it...

[Martin Thoresen]

This engine is by no means close to Rybka 4 in playing strength. In fact I think it is weaker than 1.7.1 version. So before you try and speak too loud about it, why don't you try and test it ? I am doing it right now, and am not impressed by its play. I think the SF team rushed with this version, and should have waited a little, until they have made some decent improvement before releasing it...

For this release diff stats from 1.7.1 says:

39 files changed, 2165 insertions(+), 2434 deletions(-)

So more then 250 lines of code removed form previous version, especially in search.cpp

There are no big new features, mainly code shuffle and cleanups, of course we expect an ELO increase, but this time we are quite confident
it will be smaller then last time

We spent more then 3 weeks just to regression test all the patches, and this has somewhat slowed down the development.

As usual many thanks to Jim for his great job in compiling SF and also we would like to thank people that contributed with patches and tests to this version (hope I have not forgot someone):

- Ralph Stoesser
- Marek Kwiatkowski
- Shaun Brewer
- Konstantin Orfenov
- Frank Quisinsky
- Jörg Oster
- Richard Lloyd
- Vratko Polák
- Øystein Johansen
- J. Wesley Cleveland


A special thanks to David Dahlem to allow us to use one of his great logos as the official one you'll find in the package.


Have fun
SF Team

[yanquis1972]

i know no one cares (or should anyway) but just to update 1.8 dropped to +22 over 1.7 after tje second go-round (200 games total)

+38 =127 -25

[Uly]

Well, I care, since I don't have time to run such matches I have to follow the matches of others.

[Hagen]

In other words, the Stockfish team dropped redundant code from the engine, and "cleaned" it up and therefore according to this the ELO ratings gain will be smaller. But my understanding suggests this means they're actually tightening up the code so this means when Stockfish 1.9 is released...then the ELO gain "could" be quite substantial.

Two steps back, one step forward.

[yanquis1972]

yes, that's a nice way of looking at it. a potential +20 elo just for a cleaned up 1.7 a la naum 4.1, then 1.9 gets the big improvements.

[Uly]

I don't think so, I've read about people that keep thinking "the next version will have a huge jump, the next version will have a huge jump", but it doesn't work like that, it's not magic, it's hard work.

I predict small but steady improvements, and that Stockfish will overtake Rybka 4 in 5 months.

[Robert Houdart]

A 20 Elo improvement at the top is quite substantial, you guys have unrealistic expectations.
Congrats to the Stockfish team for continuing to improve their great engine!

Robert

[yanquis1972]

I don't think so, I've read about people that keep thinking "the next version will have a huge jump, the next version will have a huge jump", but it doesn't work like that, it's not magic, it's hard work.

I predict small but steady improvements, and that Stockfish will overtake Rybka 4 in 5 months.

for some reason i gathered this came from someone in the SF group (don t think i ever saw them on record saying this), which obviously means little, (for one thing, why release 1.7 -- which as i understand it was a big overhaul/makeover -- & expect no substantial elo gain. now 1.8 is apparently a cleaned up version. in other words , not much chess work (as i think of it) has been done yet.)

and robert, i dont think anyone was seriously expecting +40 elo. it' d was a nice dream given the last surprise they gave us. of course 20 elo -- if it's even that -- is fantastic.