How to solve twisty puzzles (using Gelatinbrain's 1.4.14 as an example)

[coaster1235]

I finally found the patience to solve Starminx II (726 moves, 1h44min) :D

Indeed it wasn't hard, but the setup moves for the star tips made my solve highly inefficient. My new favourite puzle <3

[skaterfugater]

great tutorial.

I actually wanted to go to the gym already and there is not much to keep me from but now I have been watching this for over half an hour and forgot time.

I'll watch the rest later.

Alaskajoe

[coaster1235]

Encouraged by watching this, I decided to try my odds against a Starminx II (1.4.2), which as you said is this puzzle simplified, but the edges are driving me crazy. I solved them pretty much intuitively up to the last face, and the required setup moves are driving me absolutely crazy. (The edges I'd like to cycle all happen to reside in the same face). How do you do them?

[bmenrigh]

@coaster1235 Hey watch 15:50 through 17:17 and towards the end of that range at around 16:40 I demo two setup moves for the edge cycle that cycle 3 non-adjacent edges on a single face. You can use that to solve all of the edges on the face. If you have 3 edges adjacent to each other you will have to use the cycle twice to do it or add 3 more setup moves on top of the suggested 2 and you can cycle 3 adjacent edges. It's a huge pain, I know...

[bmenrigh]

@coaster1235 Also, starting at about 17:20 I talk about having a single edge flipped. you should watch out for that and take care of it now rather than having to take care of it after you get all of the edges solved.

[Katja1123581321]

Wait, I just checked. You can move edges [1,1] without moving the centers. However, like you said, edges are hard to solve using [1,1]'s, so additional algorithms would be a good idea.

[bmenrigh]

@Katja1123581321 Gah! your right! I feel so dumb for missing that. Yes, I think that is a pretty reasonable alternative then. Solve centers (just turn them into place) then solve edges [1,1] then proceed as normal. The extra setup moves needed to solve the edges while preserving the centers might be worse in the long run but I'm not sure. I'd prefer to solve edges first just because of how awful the edges are. Thank you for your help :-)

[Katja1123581321]

@bmenrigh I guess depending on how much you struggle to do the edges, you may actually end up using less moves solving it the way you suggested in the video. I'm not sure about this though, as solving the centers intuitively can be done in like 10 moves.

I would have solved much more of the edge turning puzzles if it wasn't for how badly I struggle to solve the edges on them.

[Katja1123581321]

Wow, how could I have missed that? In that case, I think your solution is very efficient!

[Katja1123581321]

I would probably solve the centers first. Since turning any face only moves 2 centers at a time, using commutators just isn't necessary :D 

[bmenrigh]

@Katja1123581321 Hey Katja, I agree that it is much easier to solve the centers if you solve them first. The problem with doing that though is that the shortest commutator to move the edges without moving the centers is 8 moves [[1:1],1] and 8 moves for 30 pieces (the edges) is a lot worse that 8 moves for 12 pieces (the centers). Plus, the edges are much harder to setup and solve so you can solve the first half of the edges intuitively and then fall back on [1,1] commutators later.