My analysis-board-joke of the day (so far) :
Not only that Rb7 is considered a blunder (there sure are faster mates, but a "blunder"?! - the old analysis board was more clever in such positions), and that the server side evaluation is given as 3.7 (which is just nuts), but observe the best-move arrow for the g3 pawn!
@fschmitz422 my word. You beat me there on the bizarre stakes! That is quite a feat for a pawn to make! I think I need to start moving my pawns like this. Or maybe it's a new piece just launched....a knawn. 😂
My analysis-board-joke of the day (so far) :
Not only that Rb7 is considered a blunder (there sure are faster mates, but a "blunder"?! - the old analysis board was more clever in such positions), and that the server side evaluation is given as 3.7 (which is just nuts), but observe the best-move arrow for the g3 pawn!
Wow. What browser are you using here, please? Thank you!
@giancz91 sorry for not responding to your concerns the first time you posted these. Some time has passed and many things have changed. Our server-side analysis (run at the first of the game) is set to run at depth 20. During book moves that depth can vary. During endgame that depth can vary. I do not know the conditions that were present when you performed that analysis, but if you run that game through Analysis now, you will see the depths expected. 35 for book moves, 20 for non-book, 1 for your crushing mate on move 12. Fun game, by the way!
My analysis-board-joke of the day (so far) :
Not only that Rb7 is considered a blunder (there sure are faster mates, but a "blunder"?! - the old analysis board was more clever in such positions), and that the server side evaluation is given as 3.7 (which is just nuts), but observe the best-move arrow for the g3 pawn!
Wow. What browser are you using here, please? Thank you!
Firefox. Updated.
After posting I noticed that all arrows were out of sync with the eval/selected move display. - Possibly related: The server side analysis had stalled for a while.
@dallin one of my chess.com friends has raised a very valid issue with the new system. Once the game analysis is done there is no way to re run the analysis over again if one wants/needs to. Why would we want to redo an analysis? - in order to run the analysis with a more up to date or "fixed" engine. i.e. a game is analysed at a point in time where there is a problem with engine. The analysis therefore has problems. The engine gets fixed. There is no way to now go and rerun the game analyses for these games with the fixed engine. Another scenario: a game is analysed now using the analysis engine that is running Stockfish 10. In the future your dev upgrades the engine to say Stockfish 11. One can't go and rerun an old analysis with the new improved engine.
Is there a way you can create a "re-analyse" button that enables one to re run a fresh analysis scan to account for the issues raised above?
can't y'all just give us the best of both worlds? give us the new and modern gui with the option we once had of running a maximum analysis? I honestly never minded having to wait as long as 20 minutes to analyze a game. at least it was accurate and helpful. now the analysis mode is just prettier but crappy and unreliable.
@9thBlunder when I used to do the full analysis in the old system it didn't even take me 20 minutes. Mine used to be finished in on average 7 minutes. Now I would be QUITE HAPPY to wait like I did for the 7 or so minutes to get a reliable analysis. So I agree 100% with you.
Now @dallin actually said a while back that they were working on a system for premium members where we could choose the initial game analysis strength - see post 21 where Dallin said:
Consistency is great, but what we really value is Accuracy. We want things to be consistently accurate. For those users who do not feel like depth=20 is deep enough, we are adding options for our premium members to change the depth of their Full Game Analysis.
Dallin then gave even more details on the proposed sysytem in post 33 where he communicated what they wanted to do as follows:
1. Allow premium members to select depth
Testing this all right now, but our options will likely be:
10 (for users who just want a basic blunder check)
18 (for users who want something fairly accurate, but fast)
20 (our default setting, and a good balance of accuracy and speed - we can share some excellent studies on the accuracy of d=20)
22 (very accurate, but takes longer)
26 (extremely accurate, but would need to be done on the client, and will take several minutes)
30 (insanely accurate, done on client, and you can have lunch while you wait.)
However I have never heard or seen anything done on this since. @dallin what happened to these solutions relating to the full game analysis strength options you put forward that were in testing?
@flashlight002, Thank you so much for your thoughtful reply. I'm glad they're taking it into consideration.
I still have a problem with them downgrading the services I paid for. It's disrespectful to users. One of the selling points of my premium membership was that I would be able to do a "deep" analysis. It was constantly advertised to me when I was using it for free. Now, all of a sudden, they change their mind and are giving me a downgraded product? If they want to get creative, they should do it with their alpha and beta tests. It feels like the rug was pulled from under me. I love this site, so I don't naturally want to leave, but they really need to understand that what they did was wrong and they're pushing me away the longer it takes for them to fix it. I don't imagine it's very difficult to return the services we paid for while they get their ish together?
The new analysis is done in 20 seconds. It is much much faster than the old ones.
Those who use it to look over their games find it very useful.
Those who want to use it to solve opening lines to 100 pli find it less so.
@flashlight002, Thank you so much for your thoughtful reply. I'm glad they're taking it into consideration.
I still have a problem with them downgrading the services I paid for. It's disrespectful to users. One of the selling points of my premium membership was that I would be able to do a "deep" analysis. It was constantly advertised to me when I was using it for free. Now, all of a sudden, they change their mind and are giving me a downgraded product? If they want to get creative, they should do it with their alpha and beta tests. It feels like the rug was pulled from under me. I love this site, so I don't naturally want to leave, but they really need to understand that what they did was wrong and they're pushing me away the longer it takes for them to fix it. I don't imagine it's very difficult to return the services we paid for while they get their ish together?
They didn't downgrade anything. The analysis is much stronger and faster than previously. Yes, there are some bugs but it is far above the old one in quality.
They resolved the problems with moving to V3. They added Puzzle Rush. They upgraded analysis. They added a lot of streaming services and exciting tournaments.
I will be surprised if they don't raise prices very soon.
@9thBlunder when I used to do the full analysis in the old system it didn't even take me 20 minutes. Mine used to be finished in on average 7 minutes. Now I would be QUITE HAPPY to wait like I did for the 7 or so minutes to get a reliable analysis. So I agree 100% with you.
Now @dallin actually said a while back that they were working on a system for premium members where we could choose the initial game analysis strength - see post 21 where Dallin said:
Consistency is great, but what we really value is Accuracy. We want things to be consistently accurate. For those users who do not feel like depth=20 is deep enough, we are adding options for our premium members to change the depth of their Full Game Analysis.
Dallin then gave even more details on the proposed sysytem in post 33 where he communicated what they wanted to do as follows:
1. Allow premium members to select depth
Testing this all right now, but our options will likely be:
10 (for users who just want a basic blunder check)
18 (for users who want something fairly accurate, but fast)
20 (our default setting, and a good balance of accuracy and speed - we can share some excellent studies on the accuracy of d=20)
22 (very accurate, but takes longer)
26 (extremely accurate, but would need to be done on the client, and will take several minutes)
30 (insanely accurate, done on client, and you can have lunch while you wait.)
However I have never heard or seen anything done on this since. @dallin what happened to these solutions relating to the full game analysis strength options you put forward that were in testing?
You are a 900 rated daily chess player. What do you need 30 pli for?
If you just like to hear the fan working on your computer and see the lights blinking, there are better solutions.
@9thBlunder when I used to do the full analysis in the old system it didn't even take me 20 minutes. Mine used to be finished in on average 7 minutes. Now I would be QUITE HAPPY to wait like I did for the 7 or so minutes to get a reliable analysis. So I agree 100% with you.
Now @dallin actually said a while back that they were working on a system for premium members where we could choose the initial game analysis strength - see post 21 where Dallin said:
Consistency is great, but what we really value is Accuracy. We want things to be consistently accurate. For those users who do not feel like depth=20 is deep enough, we are adding options for our premium members to change the depth of their Full Game Analysis.
Dallin then gave even more details on the proposed sysytem in post 33 where he communicated what they wanted to do as follows:
1. Allow premium members to select depth
Testing this all right now, but our options will likely be:
10 (for users who just want a basic blunder check)
18 (for users who want something fairly accurate, but fast)
20 (our default setting, and a good balance of accuracy and speed - we can share some excellent studies on the accuracy of d=20)
22 (very accurate, but takes longer)
26 (extremely accurate, but would need to be done on the client, and will take several minutes)
30 (insanely accurate, done on client, and you can have lunch while you wait.)
However I have never heard or seen anything done on this since. @dallin what happened to these solutions relating to the full game analysis strength options you put forward that were in testing?
You are a 900 rated daily chess player. What do you need 30 pli for?
If you just like to hear the fan working on your computer and see the lights blinking, there are better solutions.
If for you the important is that the analysis works very fast, and you don't care it makes lots of errors, just don't comment here. It's good for you, no? Keep using it and learn to play mistakes. Stay out of here. Thanks.
Play this move in official tournaments, notmtwain, it's very strong! Please, do it!
@giancz91 sorry for not responding to your concerns the first time you posted these. Some time has passed and many things have changed. Our server-side analysis (run at the first of the game) is set to run at depth 20. During book moves that depth can vary. During endgame that depth can vary. I do not know the conditions that were present when you performed that analysis, but if you run that game through Analysis now, you will see the depths expected. 35 for book moves, 20 for non-book, 1 for your crushing mate on move 12. Fun game, by the way!
Thanks for your answer, @dallin!
I repeat again, I'm sorry, my criticisms were... heavy. And yes, I haven't been the nicest guy. It's just I like Chess.com, and I want it to work well. The lack of a good analysis is not a problem for me, since I can easily do it outside of Chess.com (as I'm doing now). I complain because 1) Chess.com is a site I like very much, and I would like to do the analysis here 2) I'm a fan of Chess.com and want it to go well, and I believe you're losing a lot of credibility with that engine.
Are you saying since the time of my analysis you fixed many things? I hope so.
I hope you'll end up fixing it, keep up with the good work.
@9thBlunder when I used to do the full analysis in the old system it didn't even take me 20 minutes. Mine used to be finished in on average 7 minutes. Now I would be QUITE HAPPY to wait like I did for the 7 or so minutes to get a reliable analysis. So I agree 100% with you.
Now @dallin actually said a while back that they were working on a system for premium members where we could choose the initial game analysis strength - see post 21 where Dallin said:
Consistency is great, but what we really value is Accuracy. We want things to be consistently accurate. For those users who do not feel like depth=20 is deep enough, we are adding options for our premium members to change the depth of their Full Game Analysis.
Dallin then gave even more details on the proposed sysytem in post 33 where he communicated what they wanted to do as follows:
1. Allow premium members to select depth
Testing this all right now, but our options will likely be:
10 (for users who just want a basic blunder check)
18 (for users who want something fairly accurate, but fast)
20 (our default setting, and a good balance of accuracy and speed - we can share some excellent studies on the accuracy of d=20)
22 (very accurate, but takes longer)
26 (extremely accurate, but would need to be done on the client, and will take several minutes)
30 (insanely accurate, done on client, and you can have lunch while you wait.)
However I have never heard or seen anything done on this since. @dallin what happened to these solutions relating to the full game analysis strength options you put forward that were in testing?
You are a 900 rated daily chess player. What do you need 30 pli for?
If you just like to hear the fan working on your computer and see the lights blinking, there are better solutions.
If for you the important is that the analysis works very fast, and you don't care it makes lots of errors, just don't comment here. It's good for you, no? Keep using it and learn to play mistakes. Stay out of here. Thanks.
I do care if it makes a lot of errors but when the people complaining have 900 daily chess ratings, it is hard to take the need for error free analysis at 30 pli seriously.
I found that example you posted. You must have a malfunctioning computer.
@notmtwain the purpose of this forum thread is to bring constructive feedback to the dev team of real problems and bugs being experienced with the new analysis system with the view to assisting the dev team iron out these issues. If you have taken the time to read this entire thread and all the posts you will see irrefutable evidence of bugs, problems and very inaccurate results being reported all the time.
Chess.com has acknowledged there are issues. And they are working to fix them.
The new analysis system is a great new addition to the site with some very clever features that will assist a player of my level in LEARNING TO BE A BETTER PLAYER. But the engine, as @giancz91 so rightly and correctly points out needs to return accurate, logical and bug free results and moves otherwise I and others will be learning absolute nonsense moves!! And the engine, I am afraid to tell you, has been dishing up nonsense suggestions and moves many times. So I don't know how you can call this system "stronger". The points @ giancz91 raises are right on the money! I agree 100% with him.
Right now the analysis is fast yes. But it should not come at the expense of reduced reliability and accuracy and even more importantly dependability. I am sorry but you can't argue with all the hard evidence, screenshots and descriptions of problems documented in this forum thread that show the new engine is not accurate or reliable!
And lastly...on a personal note:
In all my posts I have been nothing but a gentleman and polite and respectful to others in this forum. I do not appreciate your sarcastic and almost belittling comments that bring up my current ratings etc. vis a vis the use of this engine. It is not appreciated at all!! I am a player LEARNING TO GET BETTER and I know I will. My goal is 1500 and I will reach it. But that is besides the point. My ratings have absolutely NOTHING to do with the performance of the new analysis engine in any way or its features.
What is more, yourcomments re me wanting to do 30 ply searches makes no sense. If you read the post properly I am quoting a proposal from Dallin at chess.com who is the head of product development, on the addition of different levels of search options to match different people's needs and requirements. I didn't say I personally was demanding 30 ply searches. The proposal lists a range of depth choices one could choose from - from 10 to 30. I personally think it is a very good idea. And quite frankly...if I feel like I want to perform a 30 ply search (if the feature should become available) that is my right and not for you to sarcastically decide for me thank you very much! I certainly wouldn't prescribe to you what to do with how you play or learn the game...unless you asked me. And I certainly didn't ask or warrant such a sarcastic and belittling tone from you quite frankly.
In closing, I have not once said I don't approve of the good work chess.com is doing to build a great site. Nor have I stated any negative views on any other features. We love this site and all it has to offer. All we are doing is showing patent bugs that are occurring with the new analysis system and communicating with chess.com staff on ideas to make the analysis system better, consistently accurate and more reliable (the very words used by Dallin from chess.com in this thread). As such we are helping and contributing of our time and energy to help improve the site. Nothing more...nothing less.
I am sorry flashlight002. I had read the first several weeks of this thread but was chagrinned to see all the chess.com bashing still going on. I know that the developers are working hard. I overreacted to the negative comments.
Again, I apologize to you.
@notmtwain thank you for replying. Apology accepted.
My views are always that we as a playing community work constructively with each other and with chess.com staff when there are problems to be addresses and fixed. I have seen a willingness by chess.com to fix things when they need fixing. Plus guys like Dallin have kept us up to speed on where they are with fixing things and what they are doing ...which I certainly appreciated. Bashing people or chess.com achieves nothing other than upsetting people...an unproductive outcome all round. It is most unfortunate that the new analysis system is experiencing all these issues. I am absolutely holding thumbs that the dev team can crack the issues and fix the problems and bugs. Given that the analysis engine is such an important tool I am sure they will.
@dallin here is another bizarre example for your dev guys to look at. This is from a game I finished and analysed last night. Move 41...Qc8 is marked as a blunder. But if one analyses with "show lines" on you can see from the screenshot that the first choice is Qc8 for this move! If your dev guys want my game pgn please let me know.
Chess pgn master scans found nothing wrong with move 41...Qc8
In the "Analyze this" scans Komodo classed the move as an inaccuracy. The Stockfish 10 scan didn't find anything wrong with the move.
So yet another strange behaviour and inaccurate analysis of a move by the chess.com engine to look at and fix. I hope this further example is of use to the dev team.