Testing New Versions all the time vs a new engine revolution

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.
TShackel
Posts: 313
Joined: Fri Apr 04, 2014 10:09 pm
Location: Neenah, WI, United States

Testing New Versions all the time vs a new engine revolution

Post by TShackel » Sun Jan 25, 2015 7:43 pm

Hi all,

Is anyone else tired of testing new versions of the same programs over and over again? I swear I've bought every Komodo and stokfish version that has come out. And do I like new versions? To an extent, yes. However, you can be assured it will play pretty much the same type of game as it did before, perhaps with some search improvements or speed increases that lead to an increase in elo, but nothing revolutionary like a rewrite of code. So we get used to these same engines, and how they play, with each new version that promises a few elo instead of focusing on improvement in chess style or something interesting.

What I want is a new revolution. What I mean by that is a great new program , well-written, that plays much stronger than stockfish, komodo, and houdini, and also impresses with its strategical and attacking phase of the game, playing more human-like moves, not in the tactical phase, when computers are already strong, but rather positional phases or planning an attack. There was a time when stockfish and houdini were such a revolutionary program. But it seems like that time has reaached a sticking point, and we're used to all the top programs now, and need something revolutionary again.

Don't get me wrong, I like stockfish and think it's a great program. I just hope there's effort from programmers put into improving chess positional and attacking "style" rather than just little elo improvements because of a search technique or a speed increase while the program plays exactly like it did before.

I will give komodo a kudos. I saw an improved king attack ability of komodo in komodo 8.0 compared to earlier versions. So I hope they can keep being revolutionary in finding better human-like positional and attacking moves, rather than what I like to call "dumb elo increase" that keeps the program stagnant and playing the same but it perhaps searches a ply deeper and so it's rating is improved. I'm not sure I want to buy every little new version unless it promises a new change in style and revolution.

I'm ready for "quality improvements" from stockfish and komodo or a new revolutionary engine, that improves the way it plays, rather than just "dumb elo increases" which takes strong chess players on the programming team like larry kaufman and those that I'm sure the stockfish team has a few of.

Is anyone else feeling a sense of stagnation on the rating lists, that we all know pretty much who the top ones are, and the top programs keep playing the same style and the same way?

Any opinions?

Edit: And I admit that this post is prompted a little by the taipan 1.0 sales pitch. I realize this one may not be legit. But I hope for one down theh road that is legit, and delivers what the poster was promising.

Tim.

User avatar
Graham Banks
Posts: 34191
Joined: Sun Feb 26, 2006 9:52 am
Location: Auckland, NZ

Re: Testing New Versions all the time vs a new engine revolu

Post by Graham Banks » Sun Jan 25, 2015 7:52 pm

Most engines play a very attractive, aggressive style of chess these days, so it's hard to get bored watching them play or playing against them.

That's my opinion anyway. :)
gbanksnz at gmail.com

User avatar
velmarin
Posts: 1600
Joined: Mon Feb 21, 2011 8:48 am

Re: Testing New Versions all the time vs a new engine revolu

Post by velmarin » Sun Jan 25, 2015 8:14 pm

The duo Larry and mark seems a good solution.
Komodo has quite advanced as computer program ,
Larry's work after years of experience, talent, knowledge and complete dedication is a tremendous advantage.

Stockfish seems to need someone how larry, but this is an altruistic project.

Do something new, it is very difficult, even talented someone would need years.

Ferdy
Posts: 4270
Joined: Sun Aug 10, 2008 1:15 pm
Location: Philippines

Re: Testing New Versions all the time vs a new engine revolu

Post by Ferdy » Sun Jan 25, 2015 9:08 pm

TShackel wrote:Hi all,

Is anyone else tired of testing new versions of the same programs over and over again?
If the new version has no improvement, I also get tired :).
I swear I've bought every Komodo and stokfish version that has come out.
Where can we buy stokfish?
And do I like new versions? To an extent, yes. However, you can be assured it will play pretty much the same type of game as it did before, perhaps with some search improvements or speed increases that lead to an increase in elo, but nothing revolutionary like a rewrite of code. So we get used to these same engines, and how they play, with each new version that promises a few elo instead of focusing on improvement in chess style or something interesting.
What is something interesting?
What I want is a new revolution. What I mean by that is a great new program , well-written, that plays much stronger than stockfish, komodo, and houdini, and also impresses with its strategical and attacking phase of the game, playing more human-like moves, not in the tactical phase, when computers are already strong, but rather positional phases or planning an attack. There was a time when stockfish and houdini were such a revolutionary program. But it seems like that time has reaached a sticking point, and we're used to all the top programs now, and need something revolutionary again.
Too demanding you think it is that easy to increase elo with style, positional and tactical insights never before seen on planet earth :D .

TShackel
Posts: 313
Joined: Fri Apr 04, 2014 10:09 pm
Location: Neenah, WI, United States

Re: Testing New Versions all the time vs a new engine revolu

Post by TShackel » Sun Jan 25, 2015 9:40 pm

Ferdy wrote:
TShackel wrote:Hi all,

Is anyone else tired of testing new versions of the same programs over and over again?
If the new version has no improvement, I also get tired :).
I swear I've bought every Komodo and stokfish version that has come out.
Where can we buy stokfish?
And do I like new versions? To an extent, yes. However, you can be assured it will play pretty much the same type of game as it did before, perhaps with some search improvements or speed increases that lead to an increase in elo, but nothing revolutionary like a rewrite of code. So we get used to these same engines, and how they play, with each new version that promises a few elo instead of focusing on improvement in chess style or something interesting.
What is something interesting?
What I want is a new revolution. What I mean by that is a great new program , well-written, that plays much stronger than stockfish, komodo, and houdini, and also impresses with its strategical and attacking phase of the game, playing more human-like moves, not in the tactical phase, when computers are already strong, but rather positional phases or planning an attack. There was a time when stockfish and houdini were such a revolutionary program. But it seems like that time has reaached a sticking point, and we're used to all the top programs now, and need something revolutionary again.
Too demanding you think it is that easy to increase elo with style, positional and tactical insights never before seen on planet earth :D .
Good work overanalyzing my post and not really contributing anything. Sorry I didn't elaborate that I bought every komodo version, and got every stockfish version free. You're the devil's advocate I suppose.

By the way, when you say I'm too demanding, well, I think I've forked over enough money to think I should have a consumer voice on how I want to see an engine developed further.

Ferdy
Posts: 4270
Joined: Sun Aug 10, 2008 1:15 pm
Location: Philippines

Re: Testing New Versions all the time vs a new engine revolu

Post by Ferdy » Sun Jan 25, 2015 9:55 pm

TShackel wrote:
Ferdy wrote:
TShackel wrote:Hi all,

Is anyone else tired of testing new versions of the same programs over and over again?
If the new version has no improvement, I also get tired :).
I swear I've bought every Komodo and stokfish version that has come out.
Where can we buy stokfish?
And do I like new versions? To an extent, yes. However, you can be assured it will play pretty much the same type of game as it did before, perhaps with some search improvements or speed increases that lead to an increase in elo, but nothing revolutionary like a rewrite of code. So we get used to these same engines, and how they play, with each new version that promises a few elo instead of focusing on improvement in chess style or something interesting.
What is something interesting?
What I want is a new revolution. What I mean by that is a great new program , well-written, that plays much stronger than stockfish, komodo, and houdini, and also impresses with its strategical and attacking phase of the game, playing more human-like moves, not in the tactical phase, when computers are already strong, but rather positional phases or planning an attack. There was a time when stockfish and houdini were such a revolutionary program. But it seems like that time has reaached a sticking point, and we're used to all the top programs now, and need something revolutionary again.
Too demanding you think it is that easy to increase elo with style, positional and tactical insights never before seen on planet earth :D .
Good work overanalyzing my post and not really contributing anything. Sorry I didn't elaborate that I bought every komodo version, and got every stockfish version free. You're the devil's advocate I suppose.
Have time to review what you posted, be sensible. There are developers that have tried and still are trying to create engine that plays interesting style but most of these engines are not comparable to Komodo and Stockfish in strength. What do you really want with the engines? The playing style of top engines now are generally beyond human comprehension, if you set the engine to be understood by humans, then you should reduce its strength.

TShackel
Posts: 313
Joined: Fri Apr 04, 2014 10:09 pm
Location: Neenah, WI, United States

Re: Testing New Versions all the time vs a new engine revolu

Post by TShackel » Sun Jan 25, 2015 10:17 pm

Ferdy wrote: Have time to review what you posted, be sensible. There are developers that have tried and still are trying to create engine that plays interesting style but most of these engines are not comparable to Komodo and Stockfish in strength. What do you really want with the engines? The playing style of top engines now are generally beyond human comprehension, if you set the engine to be understood by humans, then you should reduce its strength.
I did review what I posted. I think anyone who knows about stockfish would be able to discern that I bought Komodo, and got stockfish free.

Can you give examples of programs who are developing interesting styles? I obviously want to combine good style with strength as well and it will be a huge success when they succeed at doing that.

The desire to increase elo in engine testing is well-founded, but do it by good chess programming in evaluation function and proceed strategically in positions. It takes a good chess player to think of ways to improve the engine's evaluation function so that the evaluation function will improve elo, and not by just increasing speed or search strategies which is empty elo improvements in my book. I guess that's my main point is I want to see elo increased solely by evaluation function improvements which incudes attacking prospects, king safety, mobility, pawn breaks in closed positions, and many others. My challenge for programmers is to increase elo by mainly evaluation function. Only after they focus on that should they turn their attention to search.

When you say the style of engines is beyond human comprehension, you really mean to say tactical calculation is too far-sighted or beyond comprehension of humans. Tactics is not the only way to win games though remember. If one engine outplays another engine by strategy then it's a great win just the same. Secondly, since computers are so good at tactics they'll never leave a piece unprotected and so tactics become less a concern when computers are playing computers (since both computers see virtually all tactics), and strategy becomes how one computer beats another.

You asked me what I want to see out of engines and I'll repeat it again.....Yes increase the elo in testing, but do it by focussing on evaluation changes as the maini focus, and only after you make good evaluation progress, then work on the search.

Tim.

User avatar
cdani
Posts: 2166
Joined: Sat Jan 18, 2014 9:24 am
Location: Andorra
Contact:

Re: Testing New Versions all the time vs a new engine revolu

Post by cdani » Sun Jan 25, 2015 10:45 pm

TShackel wrote:Yes increase the elo in testing, but do it by focussing on evaluation changes as the maini focus, and only after you make good evaluation progress, then work on the search.
Tricking the search is very tempting, and a lot easier. You will see it again in the new version of Andscacs I will publish :-)

Anyway I advocate for some evolution in eval. For example a knight outpost is better if is related to something more; may be a king attack, may be a weak pawn, may be an open line... A pair of bishops the same, better if related to something more, an open or semi open column also...

May be it will be not that complicated to do this I propose, or other ideas. I will try when I reach what I consider some minimums in Andscacs. Or maybe someone take a similar path.

User avatar
Leto
Posts: 2040
Joined: Thu May 04, 2006 1:40 am
Location: Dune

Re: Testing New Versions all the time vs a new engine revolu

Post by Leto » Sun Jan 25, 2015 10:49 pm

I'm not, i'm enjoying every moment. I don't see these elo improvements as dumb, search improvements are very important. Finding the correct move a minute sooner can be the difference between winning or drawing, and sometimes losing.

Search improvements are of course not the only thing they've done, Stockfish for example recently improved its king safety, and for Stockfish 7 they're planning on improving its syzygy implementation which should increase its endgame accuracy.

User avatar
cdani
Posts: 2166
Joined: Sat Jan 18, 2014 9:24 am
Location: Andorra
Contact:

Re: Testing New Versions all the time vs a new engine revolu

Post by cdani » Sun Jan 25, 2015 11:36 pm

Leto wrote:I'm not, i'm enjoying every moment. I don't see these elo improvements as dumb, search improvements are very important. Finding the correct move a minute sooner can be the difference between winning or drawing, and sometimes losing.

Search improvements are of course not the only thing they've done, Stockfish for example recently improved its king safety, and for Stockfish 7 they're planning on improving its syzygy implementation which should increase its endgame accuracy.
Sure. But this is like extruding an existing technology, not like creating a new one. Tim is advocating for the second.

Post Reply