Recent content by ananasclassic

The 4x4x4 WTM/OBTM upper bound can be updated from 55 to 54. The proof uses a new exact 18-move bound for phases 1+2, combined with the published 16- and 20-move bounds. Details: https://www.speedsolving.com/threads/4x4x4-obtm-wtm-upper-bound-improved-from-55-to-54.97372 Artifact...

I have posted a new computer-assisted improvement of the 4x4x4 OBTM upper bound from 55 to 54. The new exact phase 1+2 bound is 18 moves, giving 18 + 16 + 20 = 54. Discussion thread: https://www.speedsolving.com/threads/4x4x4-obtm-wtm-upper-bound-improved-from-55-to-54.97372/ Proof artifact...

I have made public a computer-assisted proof artifact that improves the known OBTM upper bound for the 4x4x4 Rubik’s Cube from 55 to 54: https://github.com/AnanasClassic/4x4x4-obtm-54-research The new exact result is that phases 1 and 2 of the three-phase decomposition can always be completed...

I have also put together a catalogue of my improved Megaminx pattern algorithms. I used Walter Randelshofer’s Megaminx pattern collection as the starting point: https://www.randelshofer.ch/rubik/megaminx/ So far, I have improved 218 out of 334 patterns, which is about 65.3% of the entire...

I also keep trying to solve the pure Megaminx, but I still can’t train a model that can solve it at all, and I don’t understand why. :)

I’m finally back with something interesting: a 44-BTM superflip solution...

Link First two turns can be merged into the one BTM turn, but I don’t remember the exact notation for that.

For context, I had previously found a 47-move superflip solution in FTM/BTM, but my best QTM solution was 54 moves, one move longer than Ivan’s.

That’s a very insightful observation! Do you have any intuition about what could be an antipode for the Megaminx? Could it be a state where all pieces are significantly displaced from their solved positions? For example, on the Rubik’s Cube in QTM, this is the superflip combined with four...

After 20 hours of compute it found a 47-FTM solution It's the same approach, just with a 4 times wider beam

Thanks a lot for the kind words - I really appreciate it, especially coming from you. Either way, I'm glad you find the result interesting.

Yes, thanks, fixed it I meant that I wrote a lot of code while working on this problem, and the training script itself was written by me. There is a certain amount of “secret sauce” that I use for Kaggle competition submissions, but for the 48-FTM superflip I used an almost pure CayleyPy...

I wrote a lot of my own code, but in fact, the approach from the paper generalizes to other puzzles quite easily – you just need to change the set of valid moves, the size of the permutation, and a few hyperparameters, and then run the training pipeline to get a model that solves the given...

I’ll definitely try to solve it. Right now, I need to focus on preparing my bachelor’s thesis — it will actually be about the Megaminx — and for now I’ll be working on more down-to-earth experiments.

Of course In short, I used the CayleyPy approach from this article. For the 54-QTM solution, I used an improved model that predicts distances for all neighbors at once (I used it for a sub-80k submission to a kaggle competition) with a beam size of 2^27. The solution took about 15 hours on a...