Add date of last FSRS optimization to optimizer

Is your feature request related to a problem? Please describe.
One issue I had to deal with in the past was overfitting due to constant optimizing of the algo. In my case, my memory gets worse as workload increases, which when doing daily parameter optimizations, will slowly increase my workload infinitely. I fixed this by switching my optimizations to a monthly basis which I think allowed my review history to gather enough “new” data to make a proper assessment. The recent update which made it so that parameters are not updated if the evaluation is worse than the previous one may have partially fixed this issue, but I believe it is still possible.

Aside from the above problem, this will also help in determining if your parameters are very outdated.

Describe the solution you’d like
A simple date timestamp under the optimize button showing the last time you optimized and saved new parameters. Clicking the optimize button may also trigger a popup if you are attempting to optimize within a certain timeframe (a month, say) that warns the user against frequent optimization.

Describe alternatives you’ve considered

That’s been suggested before as well.

Use the “optimize all” action, and then a reminder will appear in the deck options after a month has elapsed.

1 Like

I don’t think that this is documented anywhere in the manual.

Also, why should this reminder appear only when “optimize all” is used? Why not when the user optimizes different presets manually?

Anyway, a text showing the date of last optimization and the number of reviews at that time (see: FSRS - Research suggests the minimum limit can be reduced to 16 · Issue #3094 · ankitects/anki · GitHub) would be useful.


Yeah, it’s not documented. And I also agree that it should appear when using the regular “Optimize” as well.

The date is not stored separately for each preset. If you have fancy requirements that involve optimizing presets individually, it sounds like you’d be better off writing an add-on to do what you want.