How to know if an algorithm is good?

The only thing that should matter is average execution time. Time each alg for an average of 12, or even better an untrimmed average of 10, and pick the fastest one. Keep in mind that the algs will all get better as they merge more into your muscle memory. Generally, though, the fastest one right from the start will be the fastest one overall.

Don't go by how the alg feels. This is important. It can certainly happen that an alg that feels "cooler," or "faster," can actually be slower to execute than another alg that doesn't feel as nice.

While this is the most important part, I think good flow is important too. It doesn't matter if an algorithm is faster if you lockup trying to do it in solves.
 
Ranzha, I use that Y perm, and I my PB is 1.5 or lower :P I can constantly sub2 that :P Do you do the B2 with your right ring finger and do two separate pulls? That's the only easy I I was able to do it. I love how nice that R2 U' R2 U R U flows though, if only you could show me a better way to do that B2 :P

I had similar feelings, so I added a rotation:

(R2 U' R2 U' R2 U) y x' r U r' D2 r U' r'

Depends if you like D2's or not I suppose.

OP: You never know until you try. It helps to already know a bunch of macros (like the sexy move) that you like and can spot easily. And if part of alg feels good and the rest of it sucks, trying added a rotation at the spot where it begins to suck.
 
Are there some common rules, wich move sequnces are good/ bad like R'F :) UF' :( ?

Personal everyone can try out what he likes, but when generating moves for others this would be even more usefull?

How important are alignment considerations like headlights to the left in PLL, or starting a Sune-COLL with the oriented Corner at UFL ?

I for example changed the "trigger stickers" I watch in Sune cases in C?LL, wich is a painfull process. I've seen many sites wich just advocate alg sets vor S-cases to sart from a different (U turn) angel zo get easier Lookup.
 
Last edited:
Back
Top