New analysis board is crap - Chess Forums

TheAgeOfSmiles

Jun 21, 2019

0

#1

I just came back from my chess vaccation.

I don't know why you would change the analysis board. The old design was way more clear and the post-game analysis was far easier to handle, the lines easier to read and so on.

There is a saying in software-engineering: Never change a running system. It was perfect the way it was.

notmtwain

Jun 21, 2019

0

#2

TheAgeOfSmiles wrote:

I just came back from my chess vaccation.

I don't know why you would change the analysis board. The old design was way more clear and the post-game analysis was far easier to handle, the lines easier to read and so on.

There is a saying in software-engineering: Never change a running system. It was perfect the way it was.

You went away for 6 months?

It's really much better. It just takes a little time to get used to. It runs much faster

They are still working out some kinks but it's a big improvement overall.

9thBlunder

Jun 22, 2019

0

#3

notmtwain wrote:

TheAgeOfSmiles wrote:

I just came back from my chess vaccation.

I don't know why you would change the analysis board. The old design was way more clear and the post-game analysis was far easier to handle, the lines easier to read and so on.

There is a saying in software-engineering: Never change a running system. It was perfect the way it was.

You went away for 6 months?

It's really much better. It just takes a little time to get used to. It runs much faster

They are still working out some kinks but it's a big improvement overall.

Stop lying. The new analysis board has been around for a couple of months only and the consensus is that it sucks. The OP is right, however, there's already a very long thread showing why the new analysis board sucks, primarily it's inaccuracies. Ever since the change, the new analysis board i inflates the accuracy of my moves, often star that I'm playing stronger than I really am. I use chessbase for analyses nowadays since the change. It's been 2 months and they still haven't fixed it.

fschmitz422

Jun 23, 2019

0

#4

notmtwain wrote:

It's really much better. It just takes a little time to get used to. It runs much faster

They are still working out some kinks but it's a big improvement overall.

Sorry, but that's ridiculous.

The story of the development and deployment of this "analysis board" is a complete disaster.

Just now, both the client side lines AND the lines computed by the server on the fly server have ceased to work. Seeing only wait cursors when I try to evaluate alternative moves.

Even months after deployment, it's still in an alpha state. Imho, the head of the project team should be fired.

Toire

Jun 23, 2019

0

#5

Ignore the apologists, it is an undeniable fact that the new "Analysis" feature is an unmitigated disaster.

notmtwain

Jun 23, 2019

0

#6

Toire wrote:

Ignore the apologists, it is an undeniable fact that the new "Analysis" feature is an unmitigated disaster.

It is no disaster. I have seen that it does produce some strange results occasionally but overall it is quite sufficient as a blunder checker and works a lot faster than the old one.

notmtwain

Jun 23, 2019

0

#7

fschmitz422 wrote:

notmtwain wrote:

It's really much better. It just takes a little time to get used to. It runs much faster

They are still working out some kinks but it's a big improvement overall.

Sorry, but that's ridiculous.

The story of the development and deployment of this "analysis board" is a complete disaster.

Just now, both the client side lines AND the lines computed by the server on the fly server have ceased to work. Seeing only wait cursors when I try to evaluate alternative moves.

Even months after deployment, it's still in an alpha state. Imho, the head of the project team should be fired.

I just tried it and it seems to be running fine.

Compare it to what they had previously, when you had to wait 10 minutes for the analysis to run. Yes, there are some kinks to work out but overall it is a big improvement.

fschmitz422

Jun 23, 2019

0

#8

notmtwain wrote:

It is no disaster. I have seen that it does produce some strange results occasionally but overall it is quite sufficient as a blunder checker and works a lot faster than the old one.

It IS a disaster, and last time I checked (about 30 mins ago), it didn't even show lines.

I don't care if d=18 or d=22 or whatever, or about fancy features, but is has to be stable, responsive and halfway bug-free, and it is NONE of that.

So will you please stop telling fairy tales now?

jdcannon

Jun 25, 2019

0

#9

Ouch this was painful to read

I use the new analysis all the time and I don't quite seem to experience the same issues.

I have heard two main complaints in this thread:

1. It doesn't show lines. (it actually does)

You may just need to turn them on.

2.) The only complaint is that its inaccurate.

On the whole, despite being much much faster, the analysis is far deeper than what you used to get with the old analysis. If you have positions that its misevaluating I would really like you to send them to me so we can research what happened.

Thanks all for the feedback!

fschmitz422

Jun 25, 2019

0

#10

jdcannon wrote:

1. It doesn't show lines. (it actually does)

2.) The only complaint is that its inaccurate.

In the meantime I've found out that it only stops to show lines after some prolongated period of time. Once I restart Firefox, it will again show them. This indicates a memory leak in the DOM tree handling or something similar.

It is simply not true that the only complaint is that it is inaccurate. I've myself posted quite a few issues in the other thread, and I've some more on my personal list, but I see no point in writing more bug reports as long as the main issues are not taken care of. For me, these are not so much related to accuracy (it's been some time now that I've seen obviously erroneous evaluations, and I don't really care about d=x?), but to stability and responsiveness (the UI is really sluggish, most of all the move animations)

To express something favorable for a change: I like the "Retry" feature. It definitely makes sense.

Regards

jdcannon

Jun 25, 2019

0

#11

Sorry, I meant to type, "only other complaint in this thread"

I will investigate lines not showing in firefox. I have not experienced that before.

fschmitz422

Jun 25, 2019

0

#12

Just one more examply why I consider the software to be still in alpha state: I just opened a game report, and: No sounds, until I manually entered an alternative move.

It was maybe a bit harsh to suggest the project head to be fired, but they definitely have to do something about quality control. Sometimes I wonder if they do any testing at all.

jdcannon

Jun 25, 2019

0

#13

What game did you open the report for and get no sounds?

fschmitz422

Jun 25, 2019

0

#14

I really don't think it depends on the game, but since you asked:

https://www.chess.com/analysis/game/live/3812090476

fschmitz422

Jun 25, 2019

0

#15

@jdcannon :

Sounds are off (until the first mouse click), if you right after opening a game report start navigating through moves list with the keyboard (arrow keys) while still on the "Report" tab.

jdcannon

Jun 25, 2019

0

#16

That is actually not a decision we made then Its a relatively new browser feature for Chrome, FF, and Safari, (probably edge too) that you cannot play any sounds until a member interacts with the page in some way.

fschmitz422

Jun 25, 2019

0

#17

Ah, I wasn't aware of that.

jdcannon

Jun 25, 2019

0

#18

The idea from the browser's point of view is to stop websites from playing annoying sound ads as soon as you go to a page or by popping a new page and playing ads there. By forcing you to interact with the page, (even a click) before playing sound you can assume that someone actually wants to see the content there.

TheAgeOfSmiles

Jul 9, 2019

0

#19

So, I am happy that I am not the only one missing the old analysis. I came back from a chess break due to very limited free time, and it was very frustrating.

I played a few games, and I can't make much sense of it.
Take this image for example:

According to the lines in the top right of the image, o-o is White's third best option (why is it not on third position, btw?). White actually chose that move and it was marked as an error as you can see in the move list. Is everything aside from the very best move now an error?!
Also, the top right shows that White is at best +1.73 points ahead. The black-white bar on the left-hand side suggests something else...

Seriously, they should just restore the old analysis system. It's simpler, it's easier to handle, and it is more consistent.

Toire

Jul 9, 2019

0

#20

TheAgeOfSmiles wrote:

So, I am happy that I am not the only one missing the old analysis. I came back from a chess break due to very limited free time, and it was very frustrating.

I played a few games, and I can't make much sense of it.
Take this image for example:

According to the lines in the top right of the image, o-o is White's third best option (why is it not on third position, btw?). White actually chose that move and it was marked as an error as you can see in the move list. Is everything aside from the very best move now an error?!
Also, the top right shows that White is at best +1.73 points ahead. The black-white bar on the left-hand side suggests something else...

Seriously, they should just restore the old analysis system. It's simpler, it's easier to handle, and it is more consistent.

I just looked at the game and the analysis at that point was okay...The move 6. 0-0 was correctly marked as an inaccuracy as 6.Bxc6 was obviously the best move.