ZBLL trainer

[GenTheSnail]

If there are any misoriented pictures

These H cases are all a U2 away from the other cases in their sets.

RLFF: A/O, C/C, O/C
RFLF: C/A, OxC

 

[Cuberstache]

Also, there are 12 cases each for the H sets BBFF and FBFB. There should only be 8 each; more are duplicates.

 

[Roman]

Also, there are 12 cases each for the H sets BBFF and FBFB. There should only be 8 each; more are duplicates.

These H cases are all a U2 away from the other cases in their sets.

RLFF: A/O, C/C, O/C
RFLF: C/A, OxC

Yeah, thanks, I'll fix it on the next update

 

[Roman]

Redundant cases eliminated, now there are only 40 algs for H set

Spoiler

O/C = CxO
C/A = A/O
A/C = O/A
C/O = OxC

This is now fixed, 2 years later. Quarantine works wonders.

 

[Hazel]

I'm really happy to see this trainer is being developed again!
There are 3 things (that I can think of at the moment) that are missing that would make the trainer near perfect in my eyes. If these were added (or even just the first 2), I would have no further reason to continue development of my own ZBLL trainer.

1) An option to go to the next case by pressing the spacebar instead of having to time each case
2) An "indexing" system. This is the only reason I still prefer my own unfinished trainer. Whenever I have trouble with a case, I can click a button or press a key to save that case to a list, and then later I can load the cases in the list so I can practice only the cases I know I have trouble with.
3) A virtual cube via the Giiker i3S like the one in Tao's trainer, preferably one that would automatically go to the next case when the current one is solved.

I feel bad about making big requests for a free software... but there's no obligation of you to add these things, I suppose.

Thank you for all the hard work

 

[Roman]

1) An option to go to the next case by pressing the spacebar instead of having to time each case

Well, you can pretend it's already implemented and just press the spacebar twice This trainer, unlike things like Anki, was meant to be a timer-like tool.

2) An "indexing" system. This is the only reason I still prefer my own unfinished trainer. Whenever I have trouble with a case, I can click a button or press a key to save that case to a list, and then later I can load the cases in the list so I can practice only the cases I know I have trouble with.

Actually this is a nice idea and quite easy to make! "Bookmarked" cases can be saved to a default preset called "bookmarks". I'll add it later.

3) A virtual cube via the Giiker i3S like the one in Tao's trainer, preferably one that would automatically go to the next case when the current one is solved.

Meh, tons of unnecessary work to save one button press. This won't drastically improve the efficiency of training ZBLL. I'm leaving it up to you

Thanks for your suggestions! The trainer has quite a room for improvement, mainly the ability use it on touch devices as easy as from PC. I am adding new features every once in a while when I feel like doing it.

 

[Hazel]

Meh, tons of unnecessary work to save one button press. This won't drastically improve the efficiency of training ZBLL. I'm leaving it up to you

A virtual could save having to scramble the cube every time, practically doubling how fast you can train cases. I understand though

 

[Roman]

A virtual could save having to scramble the cube every time, practically doubling how fast you can train cases. I understand though

Then I didn't understand what "virtual cube" mean. Do you mean (1) not using physical cube at all and control virtual cube with keyboard, or (2) controlling the virtual cube on the screen using Giiker, while Giiker state does not match the virtual cube state? Or none of that?

Note that in neither of these cases you practise real world ZBLL recognition

 

[Hazel]

Then I didn't understand what "virtual cube" mean. Do you mean (1) not using physical cube at all and control virtual cube with keyboard, or (2) controlling the virtual cube on the screen using Giiker, while Giiker state does not match the virtual cube state? Or none of that?

Note that in neither of these cases you practise real world ZBLL recognition

I mean the latter - there's a virtual cube that changes when you turn the Giiker, and you would recognize the case from the cube on-screen, and scrambles would be instantaneous to save tons of time.

 

[Username: Username:]

Could you make an option so that if you don't know the alg, you can press a button near the case's image that can show the algorithm?
like, you request the alg from algdb for that particular case. if you can't add it, it's ok, I think that's a big request.

 

[Roman]

Updates

1) No more "Not secure" warnings - images are loaded over https.
2) Added images caching - during practicing, you should see an image for previous case instantly, without blinking
3) Star/bookmarks system: you can mark cases with little "star" button during practicing; starred cases will appear in the "starred" preset @Aerma

Spoiler: star screenshots

Could you make an option so that if you don't know the alg, you can press a button near the case's image that can show the algorithm?
like, you request the alg from algdb for that particular case. if you can't add it, it's ok, I think that's a big request.

Yep, but I will need some assistance for that. If anyone is ready to make a spreadsheet that will map algdb notation to Swanson notation, please PM me. This should take an hour of manual work, but then I can easily add all algs to zbll trainer.

 

[ProStar]

Yep, but I will need some assistance for that. If anyone is ready to make a spreadsheet that will map algdb notation to Swanson notation, please PM me. This should take an hour of manual work, but then I can easily add all algs to zbll trainer.

This would be a spreadsheet of AlgDB's top ZBLL algorithm for each case except with cases labeled with Swanson notation, correct?

 

[xyzzy]

Yep, but I will need some assistance for that. If anyone is ready to make a spreadsheet that will map algdb notation to Swanson notation, please PM me. This should take an hour of manual work, but then I can easily add all algs to zbll trainer.

It might be easier (and much less error prone) to directly scrape all the top algs from AlgDb; then for each alg, determine which case it solves, and find the alg within the trainer's database that solves the same case (you need to test all four AUFs).

The only problem is that AlgDb might itself have erroneous algs (iirc some of the G perms were wrong when I last checked), but what can you do…

 

[Username: Username:]

It might be easier (and much less error prone) to directly scrape all the top algs from AlgDb; then for each alg, determine which case it solves, and find the alg within the trainer's database that solves the same case (you need to test all four AUFs).

The only problem is that AlgDb might itself have erroneous algs (iirc some of the G perms were wrong when I last checked), but what can you do…

Can't you take the algs from JuJu's ZBLL doc and for the remaining algorithms, use alg.db's algs?

 

[ProStar]

Can't you take the algs from JuJu's ZBLL doc and for the remaining algorithms, use alg.db's algs?

I may be wrong(because I don't know a lot about ZBLL), but doesn't JuJu's doc have outdated algs?

 

[Deleted member 51471]

I may be wrong(because I don't know a lot about ZBLL), but doesn't JuJu's doc have outdated algs?

Nope, it's the best doc rn.

 

[Roman]

It might be easier (and much less error prone) to directly scrape all the top algs from AlgDb; then for each alg, determine which case it solves, and find the alg within the trainer's database that solves the same case (you need to test all four AUFs).

The only problem is that AlgDb might itself have erroneous algs (iirc some of the G perms were wrong when I last checked), but what can you do…

Excellent idea, thanks! I have my solver (in C++) so I can do that. The only thing left is to find where to get all the good algs from. I have 3 sources:
1) algdb
2) JuJu's doc
3) Swanson's list

Do you think they'd be enough or are there any other good ones?
PS erroneous algs won't be a problem because the solver will indicate that no alg from the trainer algset can solve this erroneous alg.
PPS I have to actually test 16 times against each alg: first do one of four pre-AUF so that the alg starts from the right side, and then do AUF so that the cube state is solved.

 

[ProStar]

Excellent idea, thanks! I have my solver (in C++) so I can do that. The only thing left is to find where to get all the good algs from. I have 3 sources:
1) algdb
2) JuJu's doc
3) Swanson's list

Do you think they'd be enough or are there any other good ones?
PS erroneous algs won't be a problem because the solver will indicate that no alg from the trainer algset can solve this erroneous alg.
PPS I have to actually test 16 times against each alg: first do one of four pre-AUF so that the alg starts from the right side, and then do AUF so that the cube state is solved.

From what I know those lists will work great(although I don't know a lot )

 

[Hazel]

Updates

Hey, this is really cool! Thank you for implementing my suggestion and thanks for continuing to update this amazing tool after all this time

 

[trangium]

Sometimes I can get some information about the ZBLL based on the scramble, even after you removed all of the 2gen and short scrambles. For example, one scramble I got was F U2 F' U' F U' F' L' U' L U' L' U2 L, a combo of two Sunes, which means I know it's a 2GLL. Instead of going through all the algs and systematically removing those that give too much information away, I have another idea: wide moves. For example, just by looking at the scramble, I would have no idea that l F2 R' F' l D l' F' R F' l' was just R U2 R' U' R U R' U' R U' R'. This wouldn't be the default, but it could be a setting for people like me that can deduce a lot about the ZBLL just by looking at the scramble.