Why the restrictions on computer analysis with local computer evaluation on?

@IanJamesHudson I don't review 100 games per day, lol. I just like statistics and want all my games analyzed, that's all. Why are you arguing against that? Why this anti-improvement attitude?

Sometimes I notice a mistake or blunder after having done a full computer analysis without having noticed it myself (nor did the guy I was playing against), point is, I just want all my games analyzed, and this restriction is preventing me from doing that fast enough.

Not arguing against. Very average fun player here so computer analysis not something I use.
Do look back manually over the odd game now and then. But rarely bullet or blitz games purely because I generally already know when I blundered or opponent did.
My point was simply if your looking to improve your game then play and review games without time pressure. 1 min games are great fun I love them. But its not serious chess at most peoples level and relies on a variety of other factors including a fair amount of luck sometimes.

I don't think there's any limitation to the local analysis. You might have found a bug. Can you please describe how to reproduce it, and maybe share a screenshot?

@EliasAlucard download your games as @lorb suggested and run stockfish on your local machine.
If you want to burn your CPU, there are tools for that :)
For example github.com/rpdelaney/python-chess-annotator

Guys, just to be clear, downloading the pgn file and running it through some Stockfish app offline, isn't helpful. I want all my Lichess games analyzed in my online account, here on Lichess. It's difficult to keep up with the restricted times per day you can analyze matches. I play more than I'm allowed to analyze.

What I'm suggesting is that the Lichess Android app should offer us to run a computer analysis using our smartphones' CPUs, if the distributed Fishnet resources aren't enough.

@thibault I have local evaluation enabled, still get the "unauthorized access" message after requesting a computer analysis.

Maybe an option in fishnet to prioritize own games...

Why is so much effort being expended to talk @EliasAlucard out of his feature request, rather than engaging with it and explaining why it can or can't be implemented? The thought that if the client can analyze individual positions, it can analyze entire games like the server does seems very reasonable to me -- why not allow this? Perhaps there's a technical reason or it would require more programming effort than is apparent to the naked eye -- but no such argument has been made, just lots of criticism of how Elias wants to spend his time...

@joriki Yeah, I don't get that attitude either. This is a typical mainstream mentality though, you can see the exact same anti-improvement / anti-features attitude among a lot of Android users these days, who argue AGAINST root and custom-ROMs, as a case in point.

All I'm saying is that the Android app should offer the ability to do a full computer analysis using our own smartphone hardware, so that we won't have to be limited by this 25 analyses per day restriction.

@joriki "lots of criticism of how Elias wants to spend his time..."

1. Unlimited client size analysis needs not just your resources for Stockfish, but resources on Lichess servers for communication with your client to get analysis lines and store these lines in the database. If this is unlimited it will be over used. Thibault said somewhere that everything have to be limited on server side. Number of study games, number of chat lines, number of moves of a single game, etc-etc.

2. This is not just about @EliasAlucard time spend, but Lichess developers time spend also. There is a video where Thibault talks about adding new features. 26:40 in www.youtube.com/watch?v=LZgyVadkgmI

@gbtami

Many thanks for your reply!

About 2.: I'm aware that this is also about time spent by developers -- I'm a developer myself, and I didn't want to imply that any feature should be implemented just because one user requests it. I just thought the focus was wrong -- instead of developers listening to users (not just Elias but everyone, e.g. one might ask whether others would also find this useful) about what they want, developers were telling a user why he shouldn't want something -- not just once, but insisting and arguing several times after he'd already confirmed that he wants it nevertheless.

Your argument in 1. is exactly the sort of argument I'd been missing in the thread previously. I'm not quite convinced by it, though. It seems to me that in relation to all the real-time communication that needs to happen between client and server over the course of one game, a single transfer of analysis results isn't much communication. Even if a large fraction of games were analysed, I doubt that it would increase the use of communication resources per game by more than a few percent.

Another option would be to store client-side analyses on the client, thus really avoiding the use of any lichess resources per game -- but I guess that would be a bit more development effort than reusing the existing server-side analysis functionality.