LINK: https://alg-scramble-generator-2-sided-prac.netlify.app/
I just made an alg trainer of what I always imagined would be the perfect alg trainer. These are the key points:
1. I wanted an alg trainer that allows you to practice predicting AUFs and get good at 2 sided recognition without having to use your will power to force yourself to only look at 2 sides of your real physical cube.
2. I wanted an alg trainer that allows you to paste any alg list that you want to learn, not only the ones that are set by default.
3. I wanted an alg trainer that allows you to retry the algorithms that you did wrong, while not repeating the ones that you already got right.
So here it is the explanation of what I did and how the trainer works (I let AI explain it cause I am lazy):
It is a web app, no install needed. Here is how it works:
Paste any alg list you want (one per line) and hit Generate. The app inverts each algorithm and runs it through a proper solver to produce a real scramble — not just the inverse, an actual random-looking scramble for that case. Built-in presets for PLL and OLL are included, and you can save your own lists.
OPTIONS:
- "1 side / 2 sides / 4 sides / Random side" — generates scrambles for 1, 2, or all 4 pre-AUF angles of the same case, so you drill the alg from every angle.
- "Random AUF" — adds a random AUF after the algorithm, so you also have to predict the post-AUF.
- "Mid scramble" — see below.
HOW RECOGNITION TRAINING WORKS:
In practice mode there is a cube viewer on screen with different view options (U+F only, U+F+R, 2D last layer, etc). You apply the scramble to your physical cube, then look only at the screen viewer — not at the cube — to recognize the case and figure out the AUF. This is how you train 2-sided recognition without relying on the full cube state being visible.
MID SCRAMBLE MODE:
The normal mode still requires some willpower — nothing stops you from glancing at more sides of the cube than you should. Mid scramble solves this: the app generates a scramble where your case is embedded inside a random cube state, so the rest of the cube gives you no useful information. You are forced to recognize purely from the screen viewer. After executing your algorithm the app shows a short completion sequence to fully solve the cube and confirm you got it right.
RETRY WRONG ONLY:
After the session you can retry only the algs you got wrong. No repeating things you already know.
If you want to try the alg trainer and have any doubts, you can ask me here
Would appreciate if you try it out and share it with people who can benefit from it!
I just made an alg trainer of what I always imagined would be the perfect alg trainer. These are the key points:
1. I wanted an alg trainer that allows you to practice predicting AUFs and get good at 2 sided recognition without having to use your will power to force yourself to only look at 2 sides of your real physical cube.
2. I wanted an alg trainer that allows you to paste any alg list that you want to learn, not only the ones that are set by default.
3. I wanted an alg trainer that allows you to retry the algorithms that you did wrong, while not repeating the ones that you already got right.
So here it is the explanation of what I did and how the trainer works (I let AI explain it cause I am lazy):
It is a web app, no install needed. Here is how it works:
Paste any alg list you want (one per line) and hit Generate. The app inverts each algorithm and runs it through a proper solver to produce a real scramble — not just the inverse, an actual random-looking scramble for that case. Built-in presets for PLL and OLL are included, and you can save your own lists.
OPTIONS:
- "1 side / 2 sides / 4 sides / Random side" — generates scrambles for 1, 2, or all 4 pre-AUF angles of the same case, so you drill the alg from every angle.
- "Random AUF" — adds a random AUF after the algorithm, so you also have to predict the post-AUF.
- "Mid scramble" — see below.
HOW RECOGNITION TRAINING WORKS:
In practice mode there is a cube viewer on screen with different view options (U+F only, U+F+R, 2D last layer, etc). You apply the scramble to your physical cube, then look only at the screen viewer — not at the cube — to recognize the case and figure out the AUF. This is how you train 2-sided recognition without relying on the full cube state being visible.
MID SCRAMBLE MODE:
The normal mode still requires some willpower — nothing stops you from glancing at more sides of the cube than you should. Mid scramble solves this: the app generates a scramble where your case is embedded inside a random cube state, so the rest of the cube gives you no useful information. You are forced to recognize purely from the screen viewer. After executing your algorithm the app shows a short completion sequence to fully solve the cube and confirm you got it right.
RETRY WRONG ONLY:
After the session you can retry only the algs you got wrong. No repeating things you already know.
If you want to try the alg trainer and have any doubts, you can ask me here
Would appreciate if you try it out and share it with people who can benefit from it!
