https://www.chess.com/blog/Ciangsana_MASTER/

Against Stockfish: Game #14: Responsibilities (+announcement)

22 Jul 2024127 viewsEnglish (US)

Super Eval - Correspondence Game #07

Announcement:
We have been through 14 games in this series. And now it's time for a progress check. I am going to be competing in an over-the-board tournament against real human players in the coming few days. The tournament will be FIDE-rated. So I have to prepare. As this is my anonymous account, I am going to post my preparations here. It is unlikely that I will write articles, but I will for sure share them with studies. Stay tuned to catch the action. And if you are interested, try to guess Who I am!
Okay now let's get back to the article...

Continuation of the Super Eval saga:
Game #07 Analysis:

https://lichess.org/study/embed/2guM490P/dju3k7XE

Lessons from this game:
> Maybe 1...e5 is not the best fit for me.
> It's your game, it's your moves. So be responsible for the decisions you make.

Now I'll give you the Strategy descriptions and the external links. By the way, shout out to the two people who already subscribed to my YouTube channel even without any videos. If it's you, don't forget to give yourselves some credit in the forum!

Super Eval

Super Eval is a system in which you take in the situational conditions, process them to make a verbal evaluation, and choose the move that best suits your evaluation. Let's go through the steps one by one.

Situational Conditions

Situational Conditions in a tournament game can be divided into 3 main categories. They are:
1. Board Situation - Comparison of No. of Hanging pieces, King Safety, Alignments, Material, Controlled Squares, Mobility, and Pawn Structure
2. Clock Situation - Comparison of Time available per move
3. Psychological Situation - Comparison of Tournament pressure, Spectator pressure, Rating anxiety
Evaluating the situational conditions correctly requires a lot of experience. This is more of an artistic approach to chess rather than a theoretical approach.

Verbal Evaluation

We now make a verbal evaluation of the position based on the situational conditions. Basically what's happening here is that instead of saying white is better by +1.05 after this, this, that, and that so white should be ahead by +1.05 in this position, we use features of the current position to describe why white is better by +1.05. We don't need an exact evaluation like +1.05, we just need to get a rough idea about who's better in the current position based on the previously mentioned situational factors. So a normal verbal evaluation might look like this:

"White has two hanging pieces while black only has only one. But the black king is severely weakened and having the move here really benefits the white side. White has a lot more mobility than black but a slightly weakened structure. Overall, I think white should be a little bit better"

Move Selection

Now that we've got the verbal evaluation, it's time to choose a move. More than anything else, this requires a lot of knowledge and experience. This knowledge should be gained by analyzing chess games and reading annotations. Let's go back to our example verbal evaluation.

"White has two hanging pieces while black only has one. But the black king is severely weakened and having the move here really benefits the white side. White has a lot more mobility than black but a slightly weakened structure. Overall, I think white should be a little bit better"

In a position of this type, we can consider the moves with the following features:
> Moves that defend your hanging pieces.
> Moves that control the squares near the opponent's king.
> Moves that check the opponent's king.
> Moves that restrict the opponent's pieces.

After you've selected the most likely best move by matching the moves with the positional features, you have to evaluate the complexity of the position. This evaluation differs from one person to another. But the thing is, when you identify the position as complex, you have to start a process where you glance at all the possible options available for you and your opponent and start calculating the moves that look the most promising. This process has to be carried out carefully. A single miss in a complex position can lead to a complete disaster.

That was the opening and middle game strategy. Now I give you my endgame strategy.

Monte Carlo Tree Search

> First, we take a list of candidate moves and see which moves are most preferred by our intuition.
> Then we make a play-out (a.k.a roll-out) by applying the same principle from our opponent's side and mark the evaluation when a static position is reached.
> Then we take the average of the so-found leaf positions to assign a value to the calculated candidate move.
> Then we repeat the process for all the candidate moves and compare the move values to come to a final conclusion.

When using this method, it's important to broaden your list of candidate moves rather than going deeply into one line. Going deeply into one line can make you miss simple moves available for your opponent that would give them the advantage. In doing so, you also run the risk of missing moves that would have simply granted you a better position. It's kind of like how you should think in complex positions when using Super Eval. But instead of Glancing at everything, you have to glance at all the intuitional moves. This way, you can narrow down your search tree and focus only on the most promising variations.

This is how I am going to go forward. Beating Stockfish would sure be a difficult task. But I believe, at least against the Stockfish on Lichess, it's not impossible. Playing these games and writing articles on them is exhausting. But I have improved a lot through this series. So I guess it's worth it. I hope this series is instructive to all the readers and don't forget to share your thoughts in the forum!