Ahbanks presumably donated the fishnet compute which processed one or more moves of your game. But they shouldn't be mentioned in "published by", so that's just a bug.
@Kiyomomi said [^](/forum/redirect/post/IZQpPacg)
> I am using Lichess Local Engine v0.0.8 (https://github.com/fitztrev/lichess-tauri)
I don't think that engine harness is maintained anymore and it's certainly not compatible with local analysis. There have been updates to the external engine protocol since 0.0.8 was published 2 years ago.
You can use this one: https://github.com/jonbgamble/engine-host if you don't mind [installing go](https://go.dev/dl/) to build it.
Or if you're comfortable with python, command lines, and environment variables, you can use this one: https://github.com/lichess-org/external-engine
@gamblej i am trying with: https://github.com/lichess-org/external-engine
... and it works normally until it receives a request to analyze move n+1
@gamblej i am trying with: https://github.com/lichess-org/external-engine
... and it works normally until it receives a request to analyze move n+1
I'm having the exact same issue. In my case, I'm using a custom Python script via the Lichess External Engine API to run my engine, but I get the same TypeError: Cannot read properties of undefined (reading 'moves') error right after analyzing the final move.
Also, I found another bug: even when using Lichess's built-in engines, if I analyze a game and select "Copy PGN with annotations," the exported PGN is missing all move annotations (like !?, !, ?) and textual comments.
Thanks for the reports. It will be sorted.
@TarouChess Just so we're clear, you're talking about this button?
@TarouChess Just so we're clear, you're talking about this button?
@gamblej said [^](/forum/redirect/post/qA0v5ZkL)
> @TarouChess 念のため確認ですが、このボタンのことですよね? image.png (https://image.lichess1.org/display?fmt=png&h=0&op=resize&path=hj4anOBxtPZ0.png&w=335&sig=4c4dd814e46430e4d4acbb4469a21ac4eca55c9e)
Yes.
**200.000.000 nodes** is what it says after I do analysis with 'standard' and it is super slow. So I used 'seconds' instead and get around 5 million nodes per move. Speed is OK, but it did make me wish my PC was faster.
Why does 'standard' use 200 million nodes per move? Is that a bug? I thought "standard" is 1 million. Default setting takes so much time that I doubt this option will ever be used by people who don't try other setting assuming their PCs are too weak. Is this a problem only for people with non-AVX CPUs?
Can we at least get option to choose nodes per move, or, even better, DEPTH? That would shorten the time needed when game goes for 70+ moves.
Every move is calculated. I thought it would use cloud for opening and tables for endgame. When I use SF11 to analyze **move by move** it uses cloud for first maybe 10 moves.
Stockfish 11 HCE + Win10 / Firefox / non-AVX CPU
Nodes as a strength measure are not comparable across engines and Stockfish 11 HCE is a few order of magnitudes weaker than Stockfish 19.
It would have been easier to just lie to folks and pretend 1,000,000 nodes was standard strength regardless of engine. Instead, a bit of napkin math calculates how many nodes per move your engine would need to compete head to head with Stockfish 19 at 1,000,000 nodes.
That's why you see 200x for Stockfish 11 HCE. The "value" of nodes computed by pre-NNUE engines are much weaker.
That 200x factor is still a placeholder value (more accurate than 1x but still a guess). Only Stockfish 19 1MB has been truly measured vs Fishnet at this point - it needs roughly 7x more nodes.
General advice - if your browser won't let you run Stockfish 19 94MB, stick to timed analysis.