SkewbSkills: offline / online Tool to practice Skewb L2L Algs and First Layer

[HugaCuba]

For those of you that are interested in practicing the first layer on skewb or learning / drilling NS or Advanced algs for the last two layers, I created an offline tool (.exe) to help with that: SkewbSkills. It features a timer and you can select whatever types of algs you want to practice, or how many moves the scramble for the first layer shall consist of. You can also check / uncheck whether you want to become colour neutral and work on every colour (and not only on white, for example).



I'm a complete beginner in programming, and my experiences with other platforms than windows are close to zero. Therefore, the only way to use it is by unzipping the folder and starting the .exe (no problem on windows, but of course you need some third-party application on other platforms). But the code and every file is publicly available, so if you want to use the program anywhere else than on windows, using a python interpreter is another option. You can download the repository here.

For those of you that only want to download the program with its files, simply grab your latest version here. This page also gives instructions on how to use your own scramble sequences for the alg trainer, how to use the application and how to update to the latest version.

If you have further questions, want to request a feature, or give some python / pyqt advice, this would be very helpful.

Best regards and happy skewb practice,
HugaCuba (Annika Stein)

 

[u Cube]

This is a good idea! If I finally get the guts to learn ns I may use this!

 

[connorlacrosse]

Late post but I have found this to be extremely helpful in drilling NS/KK algs. Thanks!

 

[TomTheCuber101]

This is really helpful, I’ve been looking for an advanced alg trainer for ages!

 

[HugaCuba]

Glad you like it! Have a look at the new website featuring SkewbSkillsWeb etc.: SkewbPage

 

[connorlacrosse]

Glad you like it! Have a look at the new website featuring SkewbSkillsWeb etc.: SkewbPage

Ya I managed to find it on accident already. Works very well thank you!

 

[Cody_Caston]

It’s good but the scrambling is bad because its giving me one move scrambles (web version)

 

[Cody_Caston]

Never mind, there were scramble options

 

[HugaCuba]

Never mind, there were scramble options

Actually if you don't know about them, you are totally right: I should make things clearer and give hints how to individualize the scramble sequences to your own liking. Your experience is very much appreciated and will be used to make the site better / easier to understand, thanks.

 

[Cody_Caston]

Glad i helped

 

[FakeMMAP]

Sorry for the necroposting, but I have a couple suggestions:

-The FL trainer, as far as I understand, is basically a scramble generator that generates scrambles with n moves. I think a better idea (which need not just replace what exists already, it could also just be an added feature) is to generate a random scramble such that there exists a layer with n moves. The goal of the solver would then be to find that layer and to find the optimal solution to that layer (maybe with the option to make it so that the layer is always white/yellow/...). There should also be an option that tells you the solution in case the solver gives up.

-There's one more thing that can be added: being able to input a specific set of FL cases and have the trainer choose a scramble with a FL that can be solved to one of the specific solution given in the input. Again there can be the option to make it so that you only train on a specific color, or that you have to find the right layer too. To input the FL solutions I was thinking of either having a place to write all the FL solutions in Rubikskewb notation, or have them imported from a file. Importantly there should not only be the option of providing custom layers (which is also important), but also to choose from all the layers with at least 2 corners solved, with a style similar to the way you can choose the NS cases in your L2L trainer. A good start to classify the cases is to divide them in 1. 3 solved corners, 2. 2 adjacent solved corners, 3. 2 diagonally opposite solved corners. You could also incorporate the previous idea of n-move layers to select all layers than can be done in n moves (though I only see this being practical for n=3, mayyyybe n=4, so basically to the classification you would add 4. 3-movers and 5. 4-movers). One could get ambitious and put the possibility of choosing n-move layers such that ...(e.g. 2 adjacent solved corners, 2 opposite solved corners,...), but I feel this is taking it too far, at least for now.

The trainer is great. I don't see much wrong with the L2L trainer (though a link to the skewb resources sheet to reference the id of the cases would be great), and the existing FL trainer is good for some techniques to learn to do efficient layers. I would really like to see the second feature I discussed though (If you're wondering it would help practice specific 1-look cases).

 

[FakeMMAP]

Sorry for the necroposting, but I have a couple suggestions:

-The FL trainer, as far as I understand, is basically a scramble generator that generates scrambles with n moves. I think a better idea (which need not just replace what exists already, it could also just be an added feature) is to generate a random scramble such that there exists a layer with n moves. The goal of the solver would then be to find that layer and to find the optimal solution to that layer (maybe with the option to make it so that the layer is always white/yellow/...). There should also be an option that tells you the solution in case the solver gives up.

-There's one more thing that can be added: being able to input a specific set of FL cases and have the trainer choose a scramble with a FL that can be solved to one of the specific solution given in the input. Again there can be the option to make it so that you only train on a specific color, or that you have to find the right layer too. To input the FL solutions I was thinking of either having a place to write all the FL solutions in Rubikskewb notation, or have them imported from a file. Importantly there should not only be the option of providing custom layers (which is also important), but also to choose from all the layers with at least 2 corners solved, with a style similar to the way you can choose the NS cases in your L2L trainer. A good start to classify the cases is to divide them in 1. 3 solved corners, 2. 2 adjacent solved corners, 3. 2 diagonally opposite solved corners. You could also incorporate the previous idea of n-move layers to select all layers than can be done in n moves (though I only see this being practical for n=3, mayyyybe n=4, so basically to the classification you would add 4. 3-movers and 5. 4-movers). One could get ambitious and put the possibility of choosing n-move layers such that ...(e.g. 2 adjacent solved corners, 2 opposite solved corners,...), but I feel this is taking it too far, at least for now.

The trainer is great. I don't see much wrong with the L2L trainer (though a link to the skewb resources sheet to reference the id of the cases would be great), and the existing FL trainer is good for some techniques to learn to do efficient layers. I would really like to see the second feature I discussed though (If you're wondering it would help practice specific 1-look cases).

For some reason I can't edit, so I'll post in a reply: I saw that the newest version has a 1-look trainer, which pretty much solves suggestion 1. But it'd be really great to implement the possibility of choosing FLs.

 

[NigelTheCuber]

Just found this, it seems really helpful. Thanks!