Gui's that annotate?

[JohnW]

I enjoy reading Chess Life articles and books that contain annotated chess games. For a while now Fritz has had a feature to analyze and annotate a game but the annotation is pretty limited. Is there any software that does it better? I can't say I have noticed any improvement with Fritz. I am sure it would be pretty difficult but I would like something that could annotate my games but make better and more helpful comments so that it at least feels like a human grandmaster looked it.

Thanks..

[xr_a_y]

this?

https://github.com/fsmosca/chess-artist

[abgursu]

I enjoy reading Chess Life articles and books that contain annotated chess games. For a while now Fritz has had a feature to analyze and annotate a game but the annotation is pretty limited. Is there any software that does it better? I can't say I have noticed any improvement with Fritz. I am sure it would be pretty difficult but I would like something that could annotate my games but make better and more helpful comments so that it at least feels like a human grandmaster looked it.

Thanks..

As far as I know, the best is Chessbase's Tactical Analysis.

[Cornfed]

I enjoy reading Chess Life articles and books that contain annotated chess games. For a while now Fritz has had a feature to analyze and annotate a game but the annotation is pretty limited. Is there any software that does it better? I can't say I have noticed any improvement with Fritz. I am sure it would be pretty difficult but I would like something that could annotate my games but make better and more helpful comments so that it at least feels like a human grandmaster looked it.

Thanks..

As far as I know, the best is Chessbase's Tactical Analysis.

Tactical analysis feature is good for what it is, but I don't think the OP is going to get what he wants. Converting engine output to truly useful analysis seems impossible. Fritz may be best...but it is useful to only really weak players...stock phrases used over and over again. "Bg5 seems hotter" for example. What?????

That said, I use the TA feature to spot things in a group of my games I might play online or OTB to see recurring things I might have missed and it takes you to certain positions to 'solve'. But as they do not have any way to convert that to say an excel spreadsheet where certain types of missed tactics (for example) happen, it is not very good for what I want. IF they would let the user define some engine output it would be better. For example, what constitutes a 'blunder'? It is some built in difference in numeric output....say .75/pawn. For stockfish that would be one thing, for komodo, something else. In a decidedly 'great position', is + 2.75 really any better than + 2.0? Probably not. It's lazy programming really.

Chess Tempo has a way to classify your missed tactical play in a game/games...but it is not very well thought out and I feel like it is a waste of my time mostly. AI isn't really useful for chess analysis beyond basic things IMHO.

[Guenther]

I enjoy reading Chess Life articles and books that contain annotated chess games. For a while now Fritz has had a feature to analyze and annotate a game but the annotation is pretty limited. Is there any software that does it better? I can't say I have noticed any improvement with Fritz. I am sure it would be pretty difficult but I would like something that could annotate my games but make better and more helpful comments so that it at least feels like a human grandmaster looked it.

Thanks..

By accident I found this site a few days ago, which does online commenting on chess pgns, trying to mimick Human explanations via advanced AI
(together with SF).
You can try the demo and register for free and get two free commented pgns per month, otherwise it would cost some fee.
https://decodechess.com/about/

The demo and the all its possibilities look surprisingly good. I have not tried to register yet, but it sure looks interesting
and I will check it soon with some pgns.
https://decodechess.com/first-ai-chess-tutor/

[Cornfed]

I had a few min and tried to use Decode.
Unfortunately, they do not seem to be taking my 'free trial'. And that trial does NOT check thru a whole game, but rather '3 positions' you can load.
And I've not see any example of a game 'decoded'.
I tried to load a Tal-Botvinnik games from their first match...was going to compare the output....but again, you can't do that for free.
Doesn't smell right as they say...

[Cornfed]

Nevermind my post above this...user error.

So, I picked the 'free option' for DECODE and loaded Game 19 of Tal - Botvinnik 1960, to compare output with Tal's thoughts.

Obviously this is unfair on the face of it because Tal - Botvinnik 1960 might just be the best book ever written!

For the OP, this might be a bit better than the Chessbase TA or Fritz options in some respects. The other products do set up problems derived from a game for you to solve. This does not.

Decode is faster. However, while it does give more verbally, it's explanations are limited, just like the other options. It also...just says stuff that really doesn't tell you anything really. After move 12...fe4 it says this is good because:

r1bn1rk1/1pp1q1bp/3ppnp1/p7/2PPp3/P1N2NP1/1PQ2PBP/1RBR2K1 w - - 0 13

1. decoys the white queen to e4 (Qxe4)
2. threatens to play exf3 (or d5)
3. captures the white pawn
4. decoys the white knight to e4 (Nc3xe4)

So...all so obvious and not worth mentioning unless you barely know how the pieces move. Tal says: "Botvinnink takes on e4 improving the placement of his pieces." This of course after ensuing exchanges (doesn't say that part).

Ex: White to move at 19: Tal played 19.Ng5 saying this was even stronger than 19.Ne5 because "White's most important goal is to exchange the Knight on d6 after which White's two bishops will be able to slice thru Black's Queenside pawns" THAT is a chessmasters thinking! Something AI will never really be able to do (?).

Decode does not give any 'reason' behind 19.Ng5. It only says that 19.Ne5 was better because it threatens to take the c4 pawn. Pretty obvious. It does however indicate that 19.Ng5 should be met by 19...Bh6 (instead of Botvinnik's 19...Re8) which Tal does not consider...and gives an 11 ply line with no comments as to why those moves should be played. White still has a slight edge by the way.

r1b2rk1/1pp3bp/3npqp1/p7/2pP1B2/P4NPB/1P2QP1P/1R1R2K1 w - - 0 19

In short...there is no replacement for the thoughts of THE PLAYER in a good book.
Chess AI just isn't there yet.

(sadly my trying to copy in two positions for the examples did not work...not sure now to do that it seems.

[Cornfed]

Too late to edit.

I find there are other options in DECODE: Game info which lists the number of 'mistakes/blunders/inaccuracies'....plans, concepts, piece roles (in a given position)...not impressed with this. But again, to a beginner it could indeed be useful. @$99/yr it's a bit steep. Even a beginner can get more for their money on chess.com or some other site. 1 yr at 50% off currently on offer...otherwise $12/mo. I would rather a beginner/low level players look elsewhere personally.