How To Design The Best Speedcubing Trainer With Learning Intelligence? [Part 2]

[Goldsmith]

This time the questions were:
1. Do the suggested beginner values sound right?
2. Any feedback on the Algorithm Driller Idea?
3. Anything seemed wrong or maybe something was missing with the design?
4. To what extent should AUFs be treated as new cases?
5. What should the user experience be when all the cases have been fully learned?
6. Should we count mirror angles as new cases as well?

Relevant links:

The first part of the series: https://youtu.be/wDByXXVsz7o
The App: https://www.sc-trainer.com/
Announcement Post On Speedsolving: https://www.speedsolving.com/threads/new-speedcubing-training-software.80152/
Announcement Post On Reddit (identical to the one on speedsolving): https://www.reddit.com/r/Cubers/comments/mrc7lx/new_speedcubing_training_software/
The Flow Diagram: https://docs.google.com/drawings/d/1D30UXPR183mv7EoG-T7kcUUQJNj_Ed4IvX1l8HGTsL0/edit?usp=sharing

 

[abunickabhi]

Quite a well constructed flow diagram. Looking forward to the next video, discussing more details.

2 seconds recognition and 2.5 TPS is fine defaults, I would prefer 3 second recog and 4 TPS, but thats my experience with beginners.

Algorithm Driller is a good idea. The more friendly it is, the more useful beginners will find it, along with the pros.

I think mirror cases should be shown as new cases, but a label or hyperlink showing its mirrored case will be useful.

 

[Goldsmith]

Thanks for the input @abunickabhi, I'll definitely take that into account!

And yes a good balance for mirror cases is probably showing them but with good communication, that definitely sounds like the right way to go