Varying count of mistakes and blunders (end of a game)

Sort:
Avatar of YanivBorg

Why does the mistakes and blunders count change at the end of the game when we refresh the page? The count is also sometimes different on the analysis page from the page on which we play the game (shown on the right panel below the algebraic moves). Does anyone else also see that?

Avatar of YanivBorg

Why does the mistakes and blunders count change at the end of the game when we refresh the page? It also sometimes is different on the analysis page (different count of mistakes and blunders) from the page on which we play the game (on the right panel below the moves).

Avatar of YanivBorg

Blunder Count 7Blunder count zero
Here is an example. On analysis it shows 7 blunders and then on review it is zero blunders. How can this be logical?

Avatar of Martin_Stahl
YanivBorg wrote:

Why does the mistakes and blunders count change at the end of the game when we refresh the page? The count is also sometimes different on the analysis page from the page on which we play the game (shown on the right panel below the algebraic moves). Does anyone else also see that?

The post game analysis is very low depth and will almost always be different than a full review. My understanding is that's actually going to be phased out and replaced with something else.

Avatar of YanivBorg

@Martin_Stahl thanks. I did not notice a variation in the first few weeks. I heard on a broadcast that the misses and blunders calculation has changed. Maybe they are not in sync. The first score seems to follow the old model while the evaluation is updated to the new model or vice versa. I was hoping the helpdesk would respond to clarify.

Avatar of Martin_Stahl
YanivBorg wrote:

@Martin_Stahl thanks. I did not notice a variation in the first few weeks. I heard on a broadcast that the misses and blunders calculation has changed. Maybe they are not in sync. The first score seems to follow the old model while the evaluation is updated to the new model or vice versa. I was hoping the helpdesk would respond to clarify.

The immediate post game check is extremely low depth and runs in the browser, and can be influenced by client load. In many cases, it won't match a full review which is run server side and will go a lot deeper in most positions.

The two are not meant to match and it's not possible to unless the browser analysis runs deeper and takes longer or the site decided to run those server side, which isn't something that's likely due to the number of games and load it would cause.

Avatar of YanivBorg

@Martin_Stahl of all the games I played today, it matched for one but not for the rest. Maybe it matches when mistakes are very obvious or the game is simpler. When there are complications it probably does a lot more on the server side so the second, refreshed count is more reliable than the default (which runs online?).