My "official" request to top engine programmers

[Leo]

When Brain-fish comes out in its full commercial version, all our problems will be solved.

[Rodolfo Leoni]

It seems like you are a fun of infinite analysis, you can also send those positions from infinite analysis automatically to idea, so your infinite analysis sessions will not be lost, it is kept there in idea tree. You don't have to repeat your analysis again and again, you can simply load a particular idea session that you have done in the past, it can be alekhine idea or sicilian idea or nimzo-indian idea. Then start from there if you continue to work on it.

I just think infinite analysis is a better tool, but an engine with PHs should be necessary to keep track ov variation scores.

idea can save variation score if you want to.

I see IDeA evolved a lot in these years. With a PH engine and multipv mode it could be useful. Does it show engine PV(s) output while on analysis, or just its tree?

You can see the ia analysis output, and it will save automatically the position and score of the variation into the idea tree. After mini-maxing the tree, you will have a good idea of what lines may lead to a good score, bad score, equal score and interesting score. You can walk into the leaf from the position you are evaluating and may extend it further if you feel there are still interesting lines to pursue. Your analysis effort is not wasted because they are saved in the tree.

I'll consider the... IDeA ... of buying Aquarium, but I should absolutely be sure it fits the requirements I need. A Komodo or Stockfish version with Critter PHs features would be much better anyway. I'd do a lot of "advertising" on it.

[Ferdy]

It seems like you are a fun of infinite analysis, you can also send those positions from infinite analysis automatically to idea, so your infinite analysis sessions will not be lost, it is kept there in idea tree. You don't have to repeat your analysis again and again, you can simply load a particular idea session that you have done in the past, it can be alekhine idea or sicilian idea or nimzo-indian idea. Then start from there if you continue to work on it.

I just think infinite analysis is a better tool, but an engine with PHs should be necessary to keep track ov variation scores.

idea can save variation score if you want to.

I see IDeA evolved a lot in these years. With a PH engine and multipv mode it could be useful. Does it show engine PV(s) output while on analysis, or just its tree?

You can see the ia analysis output, and it will save automatically the position and score of the variation into the idea tree. After mini-maxing the tree, you will have a good idea of what lines may lead to a good score, bad score, equal score and interesting score. You can walk into the leaf from the position you are evaluating and may extend it further if you feel there are still interesting lines to pursue. Your analysis effort is not wasted because they are saved in the tree.

I'll consider the... IDeA ... of buying Aquarium, but I should absolutely be sure it fits the requirements I need. A Komodo or Stockfish version with Critter PHs features would be much better anyway. I'd do a lot of "advertising" on it.

You can use those engines in idea, so aq gui with idea tree is much much better than those engines alone.

[Rodolfo Leoni]

............
Now, I get it that adding any new feature would take resources away from improving the engine's base strength, but such new features could be bundled into a 'Pro' version of Komodo, perhaps more tweakable, that could be sold at a premium. Then the extra effort would pay off and many customers would be very satisfied.

I never bought a chess software since the old times of CM 2100 on Commodore 64, but I'd surely buy such a Komodo version! And many other people too, I guess.

Houdini, another commercial program, used to have a 'Learning' feature, but in Houdini 5 it was taken OUT (!) How do you figure that? This is the face of 'progress' in computer chess nowadays: expect very little else other than strict Elo gains - Elo is everything!

CL

I remember when Robert Houdart developed a learning system on Houdini. It was just after I published Critter session file test here.

The fact he excluded the option could be because most customers just look for engine strenght. Few chess players know about the benefits of a PH system for their analyses. So first step should be to find a new name for it... Something like "Advanced analysis process" or similar. Just matter of marketing.

I want to add a note: it takes a very little effort to build a PH system. Few hours, and it doesn't need new patches, if bug free. Just one evening for a skilled programmer.

[leavenfish]

Chess engine programmers are just that it seems - tweakers of parameters hoping to get a little more elo here and there by stumbling upon something...while hardware gets faster and faster.

To expect much more of them is perhaps not right.

Perhaps it is the shells in which they are housed which could stand some work by those who think about ways to put engines to use. I've seen so many free or cheap programs...but nothing to really make use of engines much beyond exploiting their number crunching abilities.

[Dann Corbit]

Chess engine programmers are just that it seems - tweakers of parameters hoping to get a little more elo here and there by stumbling upon something...while hardware gets faster and faster.

You will find that all the really big innovations in strength are due to search improvements rather than evaluation improvements. Most of those search improvements involve very interesting and sometimes surprising ideas.
Rybka's drop into qsearch was shocking to one or more top programmers.

To expect much more of them is perhaps not right.

Perhaps it is the shells in which they are housed which could stand some work by those who think about ways to put engines to use. I've seen so many free or cheap programs...but nothing to really make use of engines much beyond exploiting their number crunching abilities.

Here you may be onto something. I think that the really big innovation in chess for humans will have little or nothing to do with engine strength. What is really needed is a better way to display information with fresh thinking ideas.

My friend Les Fernandez wrote an engine that creates heat maps for positions on the board by piece. There are some very surprising things about the images. The program Wilhelm had some novel display techniques for endgame algorithms that helped to visualize how to win in an endgame. Chess engines are so strong now, that many of their moves are inscrutable to all but GM ability chess players without performing laborious analysis. However, a clever use of computer power could turn chess programs into much better chess teachers. I think that graphical display of information could be much improved and that conversational description of the reasons for moves could be much improved.

Why not use computers to help us understand the game better and play better. Just having a computer automaton play online to create an Elo number and not even looking at the games seems a tragic waste to me.

[Dann Corbit]

When Brain-fish comes out in its full commercial version, all our problems will be solved.

Brainfish is nothing more than Stockfish with a really good book attached.

[Leo]

Aquarium Rybka GUI looks dead. Can you still buy it?

[Leo]

Aquarium Rybka GUI looks dead. Can you still buy it?

I found it. 2017 version is available.

[leavenfish]

Chess engine programmers are just that it seems - tweakers of parameters hoping to get a little more elo here and there by stumbling upon something...while hardware gets faster and faster.

You will find that all the really big innovations in strength are due to search improvements rather than evaluation improvements. Most of those search improvements involve very interesting and sometimes surprising ideas.
Rybka's drop into qsearch was shocking to one or more top programmers.

To expect much more of them is perhaps not right.

Perhaps it is the shells in which they are housed which could stand some work by those who think about ways to put engines to use. I've seen so many free or cheap programs...but nothing to really make use of engines much beyond exploiting their number crunching abilities.

Here you may be onto something. I think that the really big innovation in chess for humans will have little or nothing to do with engine strength. What is really needed is a better way to display information with fresh thinking ideas.

My friend Les Fernandez wrote an engine that creates heat maps for positions on the board by piece. There are some very surprising things about the images. The program Wilhelm had some novel display techniques for endgame algorithms that helped to visualize how to win in an endgame. Chess engines are so strong now, that many of their moves are inscrutable to all but GM ability chess players without performing laborious analysis. However, a clever use of computer power could turn chess programs into much better chess teachers. I think that graphical display of information could be much improved and that conversational description of the reasons for moves could be much improved.

Why not use computers to help us understand the game better and play better. Just having a computer automaton play online to create an Elo number and not even looking at the games seems a tragic waste to me.

I will buy that, about search improvements.
I however almost think programmers are just stumbling along not knowing how to do this (much) better and often going 2 steps forward, 1.8 steps back....etc. It's like me using Excel at work...I've gotten to be pretty good at using it...but by slow trial and error as those teaching classes could not fathom how to use Excel to make my job easier.

Yes, interesting ways to make use of the calculation abilities in a good engine would be nice. Some of those you describe seem very interesting...wish programmers would spend more of their time looking at that as it is that which would be useful to humans...far more than getting an engine to increase it's elo a bit.