Fixing New Analysis

New analysis is VERY inaccurate, old one was much better since it gave you the choice to make deep analysis.

CEO Erik said new analysis runs on dedicated servers and reaches greater pli than the max analysis ever did.

I tried it and it's just terrible, apart its slow, heavy UI it makes a lot of mistakes

Chess.com, at least give us an option to use the old one!

Slow? The old one took a lot longer to run. 

The max analysis took many minutes.

The new one is done in ten seconds.

I said UI is slow, not analysis.

Are you on Chess.com paybook? xD

Just buy (if you can, of course) Houdini or a similar program and do it right! 

I am not on their payroll. Not at all.  

I think the new analysis is a big improvement over the previous version. First off, it runs much much faster than the old one. Secondly, the graphics make the interpretation a lot easier.  I am not quite sold on the CAPS accuracy score but I haven't seen it much previously (since it was only available to diamond subscribers.)

That said, I have noted a few problems with it that I have reported to the staff since it first came out in Beta.  There have been some unusual and questionable results.  The move counts haven't added up,  Overall, however, I think that the accuracy is very good.

Plus,  I see the effort they are making to fix problems with it.

If you want to get a response from them, join the Beta club and post specific problems that they can identify.  For example, if something is inaccurate, post a screenshot showing the inaccuracy along with a link to the game.  I am sure they would be interested.

But they need specifics.

@notmtwain @giancz91 while I think the new GUI for the new analysis engine is very cool and clever and there are some tools that will really help with learning, many of the results from this new analysis engine are HIGHLY inaccurate!! I am getting results where suggeted lines variations contain MULTIPLE blunders, bad moves and inaccuracies. I have made screenshots and even a screen video which I have sent with the game pgn file to customer support and the development team. I urge all users who experience situations where the engine returns very strange results or results with many errors in them, to screenshot or screen video it and send to support immediately together with a pgn of your game. Anyway this is just one example of what I experienced:

I did a retry. It supplied a move variation based on my new move. It appeared under the green bar that appears when one gets a blunder right. It populates a moves variation starting from one's new correct move. I clicked at the end of this variation and I was immediately taken to the analysis section where this new variation was now in the game move list. So I decided to move through each of these new moves to understand the logic of the moves. As I moved through each move the engine was re evaluating these moves. AND IT GOES AND FINDS 2 BLUNDERS IN IT'S OWN RECOMMENDED MOVE VARIATION!!

In a second test on another variation the engine suggested it found 4 blunders when I went through the moves one by one and the engine re evaluated it's moves again on the fly. 4 blunders!! The variations suggested by the engine should all be "best" moves. It should not contain 4 blunders! The variation was 11 moves in depth.

So I have serious concerns about the accuracy of the new analysis engine and system if it can dish up a suggested variation that one assumes consists of "best"moves but actually contains blunders and mistakes! I was quite taken aback to say the least.

So to reiterate: if you see that the new engine is giving you results with blunders, bad moves and inaccuracies in them please report it to customer support so that they are made aware of it and that they can send on to the dev dev team...so that we can get this new analysis tool fixed!!  

The new analysis changes its mind!!! Do the daily puzzle and after it is solved, open the analysis. Follow the "best" suggested line on self-analysis for several moves, then restart and analyze the moves. Stockfish reevaluates some of those "best" moves as inaccuracies or good, then shows the "best" move. It has done this for the last two daily puzzles. Now that is odd! How can its previously suggested best move now become a lesser move?! 

Yes, flashlight, it's very, very inaccurate. I don't remember exactly what happened to me, but I remember it made lots of mistakes, and not only in variations. In my experience it's TOTALLY unreliable. I will use autoanalysis (which they worsened since it now calculates no more than 3 lines) until they give me back the old one.

Those who love it so much will find themselves not improving from it, and learning to make mistakes. It should never be forgotten that analysis is an instrument to improve, and shouldn't be something you blindly trust of. They apparently forgot this, since it seems they blindly trust of what analysis says.

One of these days I'll make some new analysis, and this time I'll make screenshots.

Yes, such things do happen. It's totally inaccurate and unreliable.

I feel it's like it spends few time to calculate, and makes mistakes and changes his mind all the time.

@m_connors yet more concrete proof that there is something really wrong with this new analysis engine! Have you got the ability to make a screen video and send it to support showing this in action? Or even just take a screenshot of the first analysis where it shows a move as a best move and then a second screenshot where that same move is reclassified as an inacuracy or worse. Support needs to see these things so that the development team responsible for creating this new analysis tool see all these problems with this analysis tool, as it needs to be fixed!

@m_connors and @giancz91 when you do make your screenshots of the errors as they happen (so for example first screenshot showing the move classified  by the engine as a best move or part of a variation it is suggesting, and the second screenshot now showing the engine re evaluating the move in its original suggested variation as an inacuracy, bad move or blunder) send your explanations and screenshots to membersupport@chess.com as I see you are paying members. Even better try and make a screen video of it all. That is what I have done. A screen video makes it so easy for support and the dev team to actually see the errors happening! There are a number of free programs for Windows and Android devises to capture everything happening on your screen. 

All I can say is...if more people send in examples or send mails to support explaining what is happening, then the development team have more to go on to fix the problems. I have no idea if chess.com staff or moderators check our posts and report back to the dev team when they see so many people complaining about an issue.    

With the old analysis when I failed to solve Tactics, I could easily play my idea and see what was wrong with it. Is there s way to do that with the new analysis? 

I could 

Use autoanalysis, luckily it's still there

I will make screenshots, I think they'll be enough.

flashlight002@ Agree with all your points. Have experienced similar things.

Did they delete the old thread where these things were discussed?

Thank you, giancz91!

What was confusing me was that I have to make the opponent's move before I can make my move. This was not necessary with the old Tactics.