What makes a solve a good solve.

[Moreno van Rooijen]

Hi all,

I've been thinking about this matter for a while now. What does make a solve a good solve? Some might say that it mainly is the time. But breaking your PB or getting close isn't about whether it was a good solve or not. It is more about getting lucky with the scramble and how the solve turns out.

So what does determine how good a certain solve was?
I think it has mainly to do with the following points:

• Turning speed
• Look ahead
• Eliminating pauses
• Smooth turning (Without lock-ups)

Now i would like to hear from you guys what you think would make a good solve. And how important do you think the individual parts are? Maybe smooth turning is much more important than the turning speed.

What do you guys think?

 

[CRL0312]

I think there is two ways to consider a “good solve”. The first one is if you get skips, and maybe get a PB. Now I would personally call this a “lucky solve”, but it still technically is a good solve. The other one is where you have full step (no skips) and you get a time faster than your average, this clearly shows improvement and due to it being over your average, it is a good solve.

 

[xyzzy]

My turning speed is garbage anyway, so I mostly like to think of a good solve as one (i) that has a good time and (ii) where I feel like I didn't pause too much.

Spoiler: rant about 'fullstep'

Obsessing over "fullstep" solves or solves that are "too lucky" feels pointless imo. Getting lucky isn't a crime!

This should be especially familiar to Roux solvers, because the only step that truly can be "skipped" in Roux is CMLL—every other step can have solution lengths anything between 0 and 15 moves. It's not like (pure) CFOP, where the cross is the only step without a hard threshold: excluding skips, each F2L pair takes at least 3 moves, OLL takes at least 6 moves, and PLL takes at least 9 moves. (IOW, it makes sense to talk about things like a super lucky 1- or 2-move LSE, but "1-move OLL" and "2-move OLL" are literally nonsense.)

With OLL+PLL, the shortest "fullstep" LL cases would need 15 moves, but using ZBLL can bring this threshold down to just 7 moves which is less than half of OLL+PLL. For someone who knows a large chunk of ZBLL, their notion of "fullstep" would be pretty different from someone who uses only OLL+PLL. To me, OLL skips no longer exist—getting a PLL is the same as getting a random ZBLL that I happen to know.

What about big cubes? On a 6×6×6, you get a full parity skip with probability 1/8. Would you count that as "fullstep"? (I wouldn't, but only because I think "fullstep" is a silly term to begin with.) 1/8 is only a bit higher than 1/12, which is the probability of an EPLL skip (on 3×3×3) when using OLLCP. If you answered "yes" to the previous question, would you consider OLLCP with an EPLL skip to be "fullstep"? If you answered "no", what about an EO skip (on 3×3×3), which has the same 1/8 probability?

 

[greentgoatgal]

I think it depends on what you usually do well and do poorly on. If you have terrible F2L, but one solve you have excellent F2L, that would be a good solve.

 

[pglewis]

For me: guess what I think the time will be before looking at the timer, if I'm surprised at how fast it is it's a good solve.

 

[sqAree]

I like pglewis's view. ^^

For me personally all my pbs are trash solves where I build everything super inefficiently, no matter if I get a skip or not.

 

[Max Cruz]

Getting lucky isn't a crime!

-Mitchell Lane