What's lucky?

[Dyste]

In differentiating between a normal solve and a "lucky" solve, I think it would be interesting to see how people define such cases. Obviously, OLL and/or PLL skips are universally considered lucky, but among those, what else?

Do you think a lucky solve is characterized by:
self-formed CE pair
last layer edges oriented
etc.

 

[tegalogic]

I define it as any step skip of the seven (Cross [unlikely], 4 F2L pairs, OLL, PLL) without "working for the skip". So, x-cross would skip an F2L pair, but wouldn't be considered lucky. However, if you just do the cross but accidentally solve an F2L pair, you get a lucky case.
I know 29/57 OLL (need to learn more lol), but an edges-oriented-OLL doesn't count as lucky to me.

 

[Dyste]

Well, that's partially what happened for my 17.xx solve. I'm not sure of the whole solve, but I'm pretty sure I didn't skip anything. Either way, it just seems so unlikely to be a non-lucky case.

 

[Johannes91]

Well, there's some sort of definition here:

A lucky record is any solve of a puzzle where a step of your normal solution is skipped.
There are three exceptions:
1) You worked for it, f.i. by anticipating the next step.
2) The chance to skip the step is 20 percent or more.
3) The skipped step represents 3 percent or less of the total moves of the complete solution.

But it assumes that everybody's using a very static method in every solve, like LBL. Using occasional extended cross already makes it difficult to draw a line between skipping the first ce-pair and easy xcross. (And the rules for BLD solving are even more naive.)

I don't see how I could apply those rules to my solves, and anyway the numbers are quite arbitrary. I simply think that some solves are easier than others and that (single solve) records aren't very meaningful.

 

[Dene]

I think, assuming a normal cross, doing the F2L: If 2 CE pairs are made completely by accident and ready to pop-in, it is a lucky F2L.

 

[hait2]

opinion below ;P

i have no idea how this isn't obvious. everyone has a method which has steps. if you skip a step without intention, you get lucky

oll and pll are obvious. there's really no wondering what's a lucky oll or pll clearly.
just to clarify - every case can be given an algorithm. some are easy (Uperm) some are hard (N-perm).
if you accidentally solve the cross, it's not luck iff your method doesn't involve a step which solves the cross. if you know all OLL, an already solved cross just makes for an easier solve. if your method does involve solving the cross (2look OLL) and you accidentally solve it during f2l, then you got lucky.

cross is obvious. a solved cross is a lucky solve. a partially solved cross (even if it is a 1move one) isn't a lucky solve, it is an easy one. there's a (rather arbitrary, i admit) distinction -- you didn't skip the step, so it is not luck.

then we have CE pairs, 4 steps there. a matched up pair, once again, isn't luck, it is an easy solve. why? because each case can be given an algorithm (for those that don't learn intuitively, 'matching up the pair' doesnt even exist). some are easy (R U' R') some are not as much (U' R U' R' U2 R U' R').
so a matched up pair, as long as its not put in, isn't a lucky solve. it is an easy one.

and once again, oll and pll should be painfully obvious to everyone

and obviously outside of fridrich, similar concepts follow.

edit: and i don't really agree with the rule that if it's 20% or more to skip, it's not luck. it's still luck -.-; ditto for the one where there's an arbitrary (3%) materiality constraint enforced. it's still a step you do, that you didn't because you got lucky. so it should be a lucky solve. but ofc this is just my opinion

 

[Johannes91]

i have no idea how this isn't obvious. everyone has a method which has steps. if you skip a step without intention, you get lucky

What if you know many methods, and choose the best one depending on the scramble? And sometimes you combine a few different methods or do something completely different to get a really easy solve?

It's not obvious to me at all.

 

[Kenneth]

Johannes, that's how you "work for it" =)

Hait2, "cross is obvious. a solved cross is a lucky solve. a partially solved cross (even if it is a 1move one)"

I thought it was at least 3 turns for the cross and at least 3 turns for each pair.

But it's not possible to solve a pair in less than 3 turns anyway, well, OK, you can solve 2 in five (R U' R2 U R), wonder how lucky that is?

 

[hait2]

Johannes, that's how you "work for it" =)

Hait2, "cross is obvious. a solved cross is a lucky solve. a partially solved cross (even if it is a 1move one)"

I thought it was at least 3 turns for the cross and at least 3 turns for each pair.

But it's not possible to solve a pair in less than 3 turns anyway, well, OK, you can solve 2 in five (R U' R2 U R), wonder how lucky that is?

i don't know where you got the 3moves from. it's possible to have a solved cross, and so it is possible to have one which is 1 move away from being solved because it is the first step.

@johannes, it doesn't really matter. say you normally solve fridrich, however you saw a really easy 2x2x2 block, which led to a quite easy f2l, and you saw that the last pair, inserted slightly differently, could give you cross on top, which lead to an easy COLL that you also happened to know.

ok so that's a mix of like 3 methods but the steps are still present. your first step was a 2x2x2, and so if you had an already-solved 2x2x2 you'd skip the step and get lucky. and so on and so forth.

i guess the only thing that's not clear cut is if you know a certain case that leads to a skip 100% of the time. and your solve happened to have that case, which you solved, getting the skip. you didn't intentionally set up the case, but as soon as you saw it (say a certain COLL case) you knew you'd have a PLL skip.

so here, if it's your regular method, it's still luck, even if you know it's coming. but if you altered your method (say you normally do OLL PLL, but you saw the COLL case leading to PLL skip so you did that instead), then it's not luck -- it would be considered 'working' for it.

quite arbitrary =x

 

[Johannes91]

Hait2, "cross is obvious. a solved cross is a lucky solve. a partially solved cross (even if it is a 1move one)"

I thought it was at least 3 turns for the cross and at least 3 turns for each pair.

i don't know where you got the 3moves from.

Maybe from here: http://www.ws.binghamton.edu/fridrich/hints.html#limits.

@johannes

http://en.wikipedia.org/wiki/Shift_key

ok so that's a mix of like 3 methods but the steps are still present.

In that example, yes. But take a look at Ryan's Linear FMC. The solutions don't always have clear steps, at least mine didn't when I participated.

 

[pjk]

I agree with Johannes on this issue so far.

I simply think that some solves are easier than others and that (single solve) records aren't very meaningful.

And I definitely agree with this. Single solves are meaningless IMO as well.

 

[masterofthebass]

Yeah, for personal rankings I always go by averages. Certain people, for instance my brother, have gotten a single solve, ridiculously faster than their averages. My brother got a 8 sec 2x2 solve when his average was 23 seconds.

 

[joey]

Single solves aren't obviosuly a representation of your normal ability. I still think they are important, I don't ever disregard them.

 

[hait2]

Hait2, "cross is obvious. a solved cross is a lucky solve. a partially solved cross (even if it is a 1move one)"

I thought it was at least 3 turns for the cross and at least 3 turns for each pair.

i don't know where you got the 3moves from.

Maybe from here: http://www.ws.binghamton.edu/fridrich/hints.html#limits.

arbitrary. i don't see how skipping a stage is equated with having an easy stage. in fact, there's a clear line between the two. putting minimums on move count is rather stupid. for instance say the cross can be solved by doing R u R'. So solving it as R u R' is lucky whereas R u U R' is not even though it's the same cross. okay..
so then either it is wrong, or we must always consider the optimal solution. but is that valid? if you always consider the optimal solution, then there will be far more lucky solves as people simply don't see optimal solutions. if a fast f2l solution exists that leads to an OLL skip, that doesn't make the solve lucky if the solver doesn't see it. so looking at optimal solutions isn't valid either otherwise people would go around saying "hey but the optimal solution says you got an OLL skip"
it's very easy to manipulate arbitrary lines like that 3move cross one. just add a 4th random move and suddenly it's not luck anymore.

@johannes

http://en.wikipedia.org/wiki/Shift_key

ok so that's a mix of like 3 methods but the steps are still present.

In that example, yes. But take a look at Ryan's Linear FMC. The solutions don't always have clear steps, at least mine didn't when I participated.

you have a method in mind when you're doing it. you're not randomly moving the faces. hence there exist steps.

and yes i agree that single solves aren't very meaningful. an average is far more representative of your current skill level

 

[Johannes91]

The solutions don't always have clear steps

you have a method in mind when you're doing it.
hence there exist steps.

How do you know? o_O

Ok, I have a method in mind and it has one step, F3L. Haven't skipped that yet.

 

[amateurguy]

This seems to be the revival of a thread I posted up before in the Beginners' Central: http://www.speedsolving.com/showthread.php?t=1700

It's still quite unclear really. I learnt all PLLs already so CP skips are not lucky anymore. EO skips I'm not too sure.

It's funny that I can have an EO skip and still have a bad time because everything else was difficult. But my PB (23.11) was non-lucky but it had very easy F2L pairs (RUR' triggers, and two of them were already connected).

 

[jackolanternsoup]

yeah i agree... kinda how i got my pb too... easy f2l and a J-Perm finish..

 

[hait2]

The solutions don't always have clear steps

you have a method in mind when you're doing it.
hence there exist steps.

How do you know? o_O

Ok, I have a method in mind and it has one step, F3L. Haven't skipped that yet.

i know by definition.
sure now it's a matter of deciding how far you have to break it down. after all, fridrich f2l is a step, so even if you accidentally do a triple xcross you didn't skip the f2l step. you did skip the 3 CE pairs, but they're sub-steps of a step, so doing an accidental 1move triple xcross isn't a lucky solve if you don't break it down past the f2l stage.

just looking at current accepted breakdowns of steps (cross, 4 ce pairs, oll, pll.. or 2x2x2, 2x2x3, edges, f2l, ll), it seems people measure it by looks

so, if you're 1looking the cube, then sure, your method is f3l (also known as blind speedsolving). and yes, it's impossible to get lucky there unless you have a solved cube. i don't see your point.

edit: of course this depends on how people define accepted steps. since doing it by looks is seemingly invalid (read rules for lucky solve in bld'ing for instance), there exists a different way of determining which steps exist in your method. maybe you could look at it as instructions you give to another person. who knows

 

[Dyste]

This seems to be the revival of a thread I posted up before in the Beginners' Central: http://www.speedsolving.com/showthread.php?t=1700

It's still quite unclear really. I learnt all PLLs already so CP skips are not lucky anymore. EO skips I'm not too sure.

It's funny that I can have an EO skip and still have a bad time because everything else was difficult. But my PB (23.11) was non-lucky but it had very easy F2L pairs (RUR' triggers, and two of them were already connected).

Sorry. I hadn't really bothered to do a search. I was more intent on witnessing what people had in mind. It was also my intention to see how many people--a seemingly a good amount so far--think that this topic of lucky vs. unlucky solves is defined and set. People were confident that there was no other really rational contrary opinion, but it seeems that it's still as arbitrary--sorry to steal your word hait2--as I had thought.

Edit: Also, did I mention I registered later that month, after that thread started and ended on the same day? xD