Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Aquarium / Getting to the next level...
- - By ChiefPushesWood (**) Date 2019-04-20 23:30
Hey folks. I've sent a private email to a couple of you asking this. But I thought I'd post here for a general ask to the community.. Hope for some good feedback.

I honestly feel like I've got the mechanics of Aquarium down and I'm fairly confident that I'm not going to be out-analyzed by anyone on the software. At least to the point I'd lose because of that reason.

However, I'm beginning to think that I'm missing some parts that will get me to the next level. In reviewing some of my games, I'm wondering if there are situations that I'm settling for draws where I could have pressed for a win. That kind of thing. Obviously, these kinds of things could potentially lead to losses in the future with better competition. Not to mention, the wins I might be missing out on.

So, are there things I should be concentrating on learning, getting better at that will help propel me to the next level? Endgames? What else? What other kinds of things should I be focusing on. For example, how would you push for a win when you come to the realization that your opponent is also using Aquarium (that becomes quite apparent doesn't it)??

I mean, how much do you keep pushing when every line in Aquarium is 0.00 and you've analyzed for days at 20, 30, 40 ply deep on each one? To this point, when I reach that point, I'm pretty much resigned to the idea that the game is drawn. Is that a fair mindset?

Of course I know there are exceptions, but is that a fairly general rule?

Hope some of you guys can help.

Parent - - By dickie (**) Date 2019-04-21 15:25
It sometimes helps to look at the games of those slightly better than oneself say by 50 – 100 elo points. You can download the games archive from ICCF. Best to look at the last 3 or 4 years. What are the winners doing that you may not? Are they avoiding mainlines that are so deeply analysed now? Do they play deep into the middle and endgame before getting a decisive advantage? CC games are often won by the accumulation of the smallest gains and may take 50 – 100 moves or more. Always make a plan and use the tools you have to help you accomplish it. Aquarium is a great analysis tool, but it can also be quite seductive.
Parent - - By pawnslinger (****) Date 2019-04-21 17:57 Edited 2019-04-21 19:22
I agree with you, dickie.  I especially like to look at the games of previous WCCC winners.  They float right to the top in my Chessbase reference and I always take note when one of he Champs makes a move that I hadn't thought good.  I especially like following games of GM Joop van Oosterom.

I think opening preparation is most important.  As you wrote, "CC games are often won by the accumulation of the smallest gains and may take 50 – 100 moves or more."  This is completely true.  And points out the need for opening preparation.  Even very minor errors in the opening can become very large mistakes after 50 moves, if not sooner.  The openings should be played with a view toward the endgame.  I like to follow the advice of GM S. Reshevsky, one of the greatest students of the opening.  Reshevsky is best remembered for his epic battles with GM R. J. Fischer.
Parent - - By ChiefPushesWood (**) Date 2019-04-22 14:18
Thanks fellas. I appreciate it.

Parent - - By Ghengis-Kann (***) Date 2019-04-22 16:14
It probably wouldn't hurt if you personally were a better chess player.

I'm currently reading a book called "Sacrifice and Initiative in Chess" by Ivan Sokolov, and I recommend you do the same.

Positional sacrifice is exactly the thing that every machine except Alphazero has difficulty seeing, and certainly provides a way to push for a win when more prosaic moves would lead to a draw.
Parent - - By ChiefPushesWood (**) Date 2019-04-22 19:27

>It probably wouldn't hurt if you personally were a better chess player.

Thanks, GK. This is definitely one of the things that I was intimating in my post.  I was hoping some of the more experienced folks would post specific things they've found to be useful in getting over that hump.

I'll check your suggested book out.

Parent - - By Ghengis-Kann (***) Date 2019-04-22 19:52
I haven't read the Game Changer book yet, but based on the videos I have watched online the way Alphazero outplayed Stockfish relied heavily on an amazingly deep understanding of positional sacrifice.

We all know that the value of specific pieces depend on the board position, with the most commonly cited examples being good versus bad bishops, and knights being better than bishops in closed positions. Bishops are also better than knights in minor piece endgames with pawns on both sides of the board.

But it goes way deeper than that, and I believe a better understanding of positional sacrifice is "the next level".
Parent - - By pawnslinger (****) Date 2019-04-22 20:10 Edited 2019-04-23 17:47
I have often recommended a book by Hans Kmoch called "Pawn Power in Chess".

It is one of the greatest strengths of Lc0 - Leela seems to understand pawn structure and its weaknesses much better than AB engines do.  In a game I am currently working on, SF seems completely blind to the idea of connected passed pawns.  It evaluates many positions where opponents have these powerful duos, as equal.  From experience, and from watching Leela punish lesser engines, I do not agree with SF's bright evaluation.

And of course, as a result, SF under values such things when it has a chance to gain them for itself.
Parent - By Ghengis-Kann (***) Date 2019-04-22 21:18
Stockfish is open source, so you could modify it to increase the weight on this.

I printed out the code one time to see if I could learn anything useful for my own squishy brain evaluation function, and remember that in addition to the obvious increase in value as pawns move down the board there is also a Phalanx Bonus, that increases the value of pawns that are side by side. So pushing one of the pawns would give more value for that pawn being more advanced, but lose some value for breaking the phalanx. Intuitively this is easy to understand in terms of pawns on d4 and c4 in many openings. pushing the pawn to d5 provides space, but you definitely give something up by reducing the mobility of your pawn structure.

My personal favorite book on that subject is Chess Structures: a Grandmaster's Guide by Mauricio Flores Rios.
Up Topic Rybka Support & Discussion / Aquarium / Getting to the next level...

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill