Two sided ZBLL recognition: 0.33 mean of 50, 82% accuracy

This was a good run. Usually I get around 70-75% on this.

This is a modified version of my alg trainer which greys out all non white LL stickers after you make the first move. You control the cube with your keyboard, and if you can correctly solve the ZBLL case without the information from the stickers that were greyed out it counts as a success. The recognition time is counted as the time between when the case is shown and when you make the first move.

I do a U move immediately at the start of each case to grey out the stickers when I'm ready.

I don't intend to release this version of the trainer publicly. DM me if you want it.

[blackbird]

ur reddit post about sub 40 3bld being easier than sub 1 min is so true

[piershanson]

I was thinking about learning recognition for pll/zbll/whatever alg-set, and I was wondering if it would be possible to make a mode that weights the likelyhood of an alg appearing based on how often a user sucessfully recalls the alg.

(e.g. maybe there's a correct/incorrect button whenever they get the case, and then the alg trainer is able to store data in the browser about a user's success rate on given algs/given auf's of algs. Then, it gives cases more often when the user fails on them: Kinda like an automatic version of manually selecting which sets you want to learn, with the ability to keep you on your toes with cases that you haven't seen in a while. One possible weighting that might work well is w=e^(failures-successes))

[kennedylewicki]

How do you recognize ZBLL? Or what method do you recognize ZBLL?