Forums

Why an inaccuracy?

Sort:
Bunneh7

Why is dxe1=Q+ an inaccuracy and dxe1=R+ considered best here? Is there some stalemate possibility or something that I'm not seeing? The predicted outcome the engine talks about is different for the Queen and the Rook promotion. Could I not do the same in either case? When it is a Rook it says 31. Kh2 Qf4+ 32. g3 Qf3 33. Qh7+ Kxh7 34. a5 Rh1#. When Queen it says 31.Kh2 Qf4+ 32. g3 Qexf2+ 33. Qxf2 Qxf2+, etc. Engine quirk or something I'm missing?

ShikshaWithPraveen

Reading too much into. Don't believe every aspect of the computer analysis you see. Sometimes common sense and intuition prevail.

Duckfest

You aren't missing anything: dxe1=Q+ is not an inaccuracy.

This is something that happens because of the way Stockfish evaluations come with some variance and this can be misinterpreted in your Game Review. If you let Stockfish calculate long enough and at higher depth both moves will be shown as equally strong.

My most recent article is about the engine analysis and game review, how they work together and how the feedback for players can be confusing, misleading or even wrong. I'm addressing typical situations where you shouldn't blindly trust the review, including situations exactly like yours. Understanding the mechanisms better makes it easier to interpret the game analysis better.

Compadre_J

Technically, the engine is correct for the reasons it gave.

Queen Promotion is considered an inaccuracy because with best play it prolongs the game.

—————————————

Here is Rook line you gave:

31. Kh2 Qf4+ 32. g3 Qf3 33. Qh7+ Kxh7 34. a5 Rh1#.

The Key part to focus on is on move 33.

White can play any move they want to save themselves from the mate.

It just so happens in this position White doesn’t have any saving move.

The best White can do is prolong the mate by 1 turn with Queen Sac.

———————————————

Here is Queen line you was asking about:

31. Kh2 Qf4+ 32. g3 Qf3 33. Qh7+ Kxh7 34. a5 Qh1#.

You asked can’t White do same mate with the Queen Promotion?

The answer is Yes, but it is considered inferior to play the position like above.

The Key part to focus on is on move 33 again.

We are giving White 1 single move to save themselves.

Again, they just don’t have a saving move in this position.

Now, Watch how the computer wants to play!

31.Kh2 Qf4+ 32. g3 Qexf2+ 33. Qxf2 Qxf2+

Do you see the difference?

———————-

Previously, we was putting the Black Queen on f3 and passing the turn.

This allowed White the chance to have 1 single game saving move.

This position doesn’t have any game saving move, but in other chess games it could be possible.

——————————

By taking the f2 pawn, we force the trade of Queen.

This causes White to be losing forever as now they have no way to recover and they have no saving move.

Do you see how giving your opponent 0 saving moves is better vs. giving your opponent 1 saving move?

But if we do the above it creates an issue right? Black is winning, but it will take a lot longer to mate.

———————————

The engine has difficulty interpreting the above actions.

The way the engine fixes the situation is by under promoting to Rook so the f2 square isn’t double attacked.

magipi
Compadre_J wrote:

Technically, the engine is correct for the reasons it gave.

(...)


The way the engine fixes the situation is by under promoting to Rook so the f2 square isn’t double attacked.

Everything you wrote in that huge wall of text is untrue. It's total fabrication on your part and it has nothing to do with reality.

In reality, the engine instantly identifies that promoting to a rook and to a queen are both mate-in-5. There is absolutely no difference.

Mazetoskylo

As you can see, it makes no difference at all if you promote to queen, or rook: both are mate in 5 in identical fashion. Let's suppose that the computer thinks that promoting to rook is more "flashy", but actually both promotions are the same.

I just feel sorry for you if you wasted your precious time reading all that nonsense in post #4.

borovicka75
Both moves are mate in 5 if 32… Qf3 is played. Game tewiew is probably using too little time to analyse whole game which can lead to strange things like recomending Qxf2 instead of Qf3 with mate in 2. I have no idea if there is a possibility to set up more time for analysis. Otherwise i recomend to check your game with manual analysis.
snowmannen64

I checked. They’re both best

snowmannen64

I used analysis in chess.com

magipi

To answer the original question: chess.com's Game Review calls it an "inaccurracy" because it's bugged. It has nothing to do with the engine, the engine is fine. It's the chess.com script that's bugged.

I'm always astonished that people use Game Review and nobody seems to complain about all the bugs and other issues.

Mazetoskylo
borovicka75 wrote:
 I have no idea if there is a possibility to set up more time for analysis.