[Help Thread] FMC Discussion

[Kryptonite]

...now let's consider that the previous moves in the skeleton were U' R and now we can make insertion R' U R' D2 R U' R' D2 R2. Now doing the commutator a move earlier wouldn't make the same cancellations.

Good example, you're definitely right on this one--it looks like I'll have to get those 9f*s down!

If I'm not mistaken, this could be considered to be a Markov chain so you could make a transition matrix and calculate the probabilities from there.

Looks a bit daunting, though I may have to do that. I am curious though if different solving methods affect the odds of 8f*s appearing, which would be beyond a Markov chain--though it would be an excellent way to have a par for the course.

Sampling appeals to me mostly because I am more familiar with it.

 

[Kryptonite]

Alright, I looked into setting up a Markov chain, but from what I can tell, it is incredible complex.

Instead I've done a quick and dirty analysis. I ran five random scrambles and solved them optimally using Cube Explorer. I picked three corners at random, then ran through all five scrambles continually for 90 total moves. At each move I looked to see if I could find a 8f* cycle, and counted how many I found.

I averaged 45% (+/-10% based on sample size) from this, the caveats being:
1. Humans, of course, don't solve optimally
2. I may have found moves that are identical after cancelations, though I honestly wasn't sure if they should or shouldn't be counted

One interesting place where human solutions differ from optimal solutions is the likelihood of half turns. In optimal solutions half turns occur about 30% of the time (sampled from 40 solutions). Human FM solutions tend to have about 25% (sampled from 50 solutions).

 

[blah]

At each move I looked to see if I could find a 8f* cycle, and counted how many I found.

I may be wrong, but I still think you don't quite know how commutators work :/

 

[Stini]

At each move I looked to see if I could find a 8f* cycle, and counted how many I found.

I may be wrong, but I still think you don't quite know how commutators work :/

Weird, to me it seems like he understands them well enough.

 

[Kryptonite]

At each move I looked to see if I could find a 8f* cycle, and counted how many I found.

I may be wrong, but I still think you don't quite know how commutators work :/

Hmm, while I certainly wouldn't claim to be a master, I think I grasp them firmly enough for this application (thanks for the vote of confidence Stini ).

If I have made errors though, I'm absolutely on board to track them down--just an opportunity to learn more about the cube! If you let me know what you think I'm missing in my last post, I'm more than happy to update my results or clarify them.

 

[sz35]

Hi guys!
I'm pretty new to FMC, and I have question: What's NISS?

 

[guusrs]

Hi guys!
I'm pretty new to FMC, and I have question: What's NISS?

Hi Shai,

Welcome to this thread.
For understanding NISS a little bit have a look at this:

Hi folks,

For advanced FMC users here is a nice example of my NISS-method for FMC (Normal-Inverse-Scramble-Switch):
Scramble: (fmc.mustcube.net #265) F2 B' R' B D B F' R B F L' D2 U2 B2 D U' B' F D2 L' F2 B' R2 B2 R L' F2 U' F' L

my solve: D' B' U B2 D' L F2 L' D R U2 R U2 B' R' U2 RU' R B' R' F2 U' L B' (25)
explanation:
inverse scramble: nice start: B L' U F2 (4)
apply on normal sramble: [pre-moves F2 U' L B'] nice continuation: D' B' U B2 R2
apply on inverse scramble: [pre-moves R2 B2 U' B D]: B L' U F2 (4+[5])
then continuation on inverse scramble was easy:
F2L: R B R' U R' U2 R B (12+[5])
LL: U2 R' U2 R' D' L F2 L' D R2 (22+[5])
pre-move correction: R2 B2 U' B D, 2 moves cancel (25)
Took my 90 minutes including finding a 29-move 1-hour solution

NB. I should reduce executing time otherwise I will get an DNF next WC as well

Gus

and this:

Generally, you use premoves to modify a scramble so you can solve it more easily. The most common approach for this is to solve a "pseudo-block" of some sort, where for instance you might put in a corner-edge pair of the wrong color into a 2x2x2 or 2x2x3 block. Then you find the moves necessary to put that corner-edge pair (or whatever) into it's proper location in the solved cube. If you then apply those moves to the solved cube prior to applying the scramble, you'll find that the pseudo-block becomes an actual block. Then you can add those moves to the end of your final solution and it will be solved.

Guus's solution you mention is using "NISS" - his "normal-inverse-scramble-switch" method, where you solve part of the cube with the normal scramble, then when you get stuck, you switch to the inverse scramble and use it, then when you get stuck again, you switch back to the normal scramble, and so on. I love NISS - it's really fun to do! Anyway, when you switch from normal to inverse scrambles, the moves you used at the beginning of the normal scramble solve become premoves to the inverse scramble (and vice-versa). It takes trying it to see how it really works, and it's really quite ingenious.

The best way to understand these is to look at some of the solves that others have done (like the ones in the weekly competitions) and understand how they work. Examples are much more useful than descriptions - this isn't a particularly good description, but maybe it gives you an idea of what's going on.

Both from this thread

Gus

 

[sz35]

Hi guys!
I'm pretty new to FMC, and I have question: What's NISS?

Hi Shai,

Welcome to this thread.
For understanding NISS a little bit have a look at this:

Hi folks,

For advanced FMC users here is a nice example of my NISS-method for FMC (Normal-Inverse-Scramble-Switch):
Scramble: (fmc.mustcube.net #265) F2 B' R' B D B F' R B F L' D2 U2 B2 D U' B' F D2 L' F2 B' R2 B2 R L' F2 U' F' L

my solve: D' B' U B2 D' L F2 L' D R U2 R U2 B' R' U2 RU' R B' R' F2 U' L B' (25)
explanation:
inverse scramble: nice start: B L' U F2 (4)
apply on normal sramble: [pre-moves F2 U' L B'] nice continuation: D' B' U B2 R2
apply on inverse scramble: [pre-moves R2 B2 U' B D]: B L' U F2 (4+[5])
then continuation on inverse scramble was easy:
F2L: R B R' U R' U2 R B (12+[5])
LL: U2 R' U2 R' D' L F2 L' D R2 (22+[5])
pre-move correction: R2 B2 U' B D, 2 moves cancel (25)
Took my 90 minutes including finding a 29-move 1-hour solution

NB. I should reduce executing time otherwise I will get an DNF next WC as well

Gus

and this:

Generally, you use premoves to modify a scramble so you can solve it more easily. The most common approach for this is to solve a "pseudo-block" of some sort, where for instance you might put in a corner-edge pair of the wrong color into a 2x2x2 or 2x2x3 block. Then you find the moves necessary to put that corner-edge pair (or whatever) into it's proper location in the solved cube. If you then apply those moves to the solved cube prior to applying the scramble, you'll find that the pseudo-block becomes an actual block. Then you can add those moves to the end of your final solution and it will be solved.

Guus's solution you mention is using "NISS" - his "normal-inverse-scramble-switch" method, where you solve part of the cube with the normal scramble, then when you get stuck, you switch to the inverse scramble and use it, then when you get stuck again, you switch back to the normal scramble, and so on. I love NISS - it's really fun to do! Anyway, when you switch from normal to inverse scrambles, the moves you used at the beginning of the normal scramble solve become premoves to the inverse scramble (and vice-versa). It takes trying it to see how it really works, and it's really quite ingenious.

The best way to understand these is to look at some of the solves that others have done (like the ones in the weekly competitions) and understand how they work. Examples are much more useful than descriptions - this isn't a particularly good description, but maybe it gives you an idea of what's going on.

Both from this thread

Gus

Thanks Gus/
I knew about switching to inverse scramble I just didn't know it is called NISS

 

[Kryptonite]

Hey, while we're on the subject, how do you know when NISS will yield good results? I have a decent grasp on how to do it/why we can do it, but when I've done it in my solves it seems like the cube case at the switch is quite similar to the case we switched from. Any general guidelines for when to try a switch?

 

[guusrs]

Hey, while we're on the subject, how do you know when NISS will yield good results? I have a decent grasp on how to do it/why we can do it, but when I've done it in my solves it seems like the cube case at the switch is quite similar to the case we switched from. Any general guidelines for when to try a switch?

It doesn't work al the time, but sometimes it does, so it is worth a try, especially when you're stuck on a good start.

Gus

 

[Stini]

Hey, while we're on the subject, how do you know when NISS will yield good results? I have a decent grasp on how to do it/why we can do it, but when I've done it in my solves it seems like the cube case at the switch is quite similar to the case we switched from. Any general guidelines for when to try a switch?

As far as I know, there's really no guarantee that NISS will help you, so I find that you should avoid overusing it. I usually try NISS only when I get stuck, since it's a bit time consuming. In general I prefer using the normal scramble for practical reasons, since you don't have to invert your solution in the end and I also find it more elegant that the actual solution is understandable without any tricks like inverse scramble or NISS.
There are some cubers though, who seem to use NISS all the time, so their opinion is probably different than mine.

That being said, you can actually see quite a bit about the resulting position beforehand, but explaining it is rather difficult, and it's probably faster and easier to just do NISS instead of trying to figure out the pre-moves without using the inverse scramble. For example I look often for "pseudo pairs", which are something that would become an actual pair with a single pre-move. Recognizing pseudo pairs is easier than it sounds, but you could also check the inverse scramble and notice that there is a separated pair (a pair that you can connect with a single move). So if you see a pseudo pair, you know that there is a separated pair if you switch to the inverse scramble. For me it's just faster to notice pseudo pairs than to switch to the inverse scramble, but obviously this way you can only make the normal scramble easier - you don't actually try if the inverse scramble is easier than the normal scramble (you pretty much just switch to the inverse scramble for one move and then switch back, of course this could be generalized for several moves, but it's quite complicated).

 

[Kryptonite]

It doesn't work al the time, but sometimes it does, so it is worth a try, especially when you're stuck on a good start.

As far as I know, there's really no guarantee that NISS will help you, so I find that you should avoid overusing it.

Very cool, sounds like I wasn't missing anything on NISS then.

In general I prefer using the normal scramble for practical reasons, since you don't have to invert your solution in the end and I also find it more elegant that the actual solution is understandable without any tricks like inverse scramble or NISS.

Got to admit, I love solutions that are completely incomprehensible in their final form--the uninitiated can only assume that you're a genius, haha.

I look often for "pseudo pairs", which are something that would become an actual pair with a single pre-move.

Clever, I'll have to try this. I'll often take some time to predict how a pre-move to fix pseudo-block will affect the cube state to see if it's worth the time or whether I should explore more options, so it sounds like looking for pseudo pairs is taking this one step further.

 

[Cride5]

Hey, while we're on the subject, how do you know when NISS will yield good results?

With my blockbuilding style, I tend to concentrate on smaller sub-goals of building 1x2x2's and pairs, rather than the traditional 2x2x2 -> 2x2x3 -> F2L-1 approach. I then then tie them together (using pseudo blocks if needs be), while also concentrating on EO. Following that approach it's quite common to end up with blocks tangled in such a way that they are difficult to put together. Switching to the inverse will give exactly the same blocks, but arranged around the cube in a completely different way ... and if you're lucky they will be much easier to put together. This 21-mover from the weekly comp is a good example of this strategy: http://www.speedsolving.com/forum/showpost.php?p=412605&postcount=28

... so my short answer to your question would be: When you have some nice blocks but no easy way to put them together.

 

[CharlesOBlack]

hi guys.

I've recently gone back to FMC, and I suck at it. [my solves are usually 45 moves or so.] So I did one just now and wanted to post it, since it's only 37 moves with cancellation ^_^

Scramble: U D R' U2 B2 F R' L2 U L B2 F2 L U' L B' L2 U' R2 L2 F R U' R2 B2 F' D B F' L

B U' R2 (3/3) [2x2x2]

R' U' R F' L' F U2 (7/10) [2x2x3]
U' L U x' z (3/13) [EO]
U R2 U' R' U R2 (6/19) [F2L]

R' U' R U' R' U2 R y' (7/26)

R U R' U' R' F R2 U' R' U' R U R' F' (14/40) [40 - 3 = 37 moves]

I guess I got lucky. Solved with Petrus and T-Perm only.

 

[Kryptonite]

hi guys.

I've recently gone back to FMC, and I suck at it. [my solves are usually 45 moves or so.] So I did one just now and wanted to post it, since it's only 37 moves with cancellation ^_^

Scramble: U D R' U2 B2 F R' L2 U L B2 F2 L U' L B' L2 U' R2 L2 F R U' R2 B2 F' D B F' L

B U' R2 (3/3) [2x2x2]

R' U' R F' L' F U2 (7/10) [2x2x3]
U' L U x' z (3/13) [EO]
U R2 U' R' U R2 (6/19) [F2L]

R' U' R U' R' U2 R y' (7/26)

R U R' U' R' F R2 U' R' U' R U R' F' (14/40) [40 - 3 = 37 moves]

I guess I got lucky. Solved with Petrus and T-Perm only.

I'm pretty new as well, and looking back on my progress, I think I can offer a few tips.

You have a really good start here; it's after where you should be looking to shave a couple moves off. Also, you don't end up with any pairs after your EO. So by adding a single move for a different and suboptimal EO, you might end up with a more favorable case.

Looking back at my solutions, this is pretty much on par with my better solutions before I started using more advanced FM techniques. So from here I recommend you learn how to use a keyhole, insertions, pre-moves, and NISS--probably in that order.

While there's a lot to go through, nearly everything you need to know is in this thread and this one: http://www.speedsolving.com/forum/showthread.php?t=1566 If you can't find anything here, it's somewhere on the net.

The last thing I recommend is to not put cube rotations in your notation--it's easier to catch cancelations and it's simpler for sharing them.

Good luck learning the techniques! I'd recommend taking them on one at a time.

 

[Sebastien]

My Solution from Austrian Open 2010

Scramble: L' D' R L2 U' L2 F' L F' R F' L D' F2 U2 R2 B U2

Inverse Scramble: U2 B' R2 U2 F2 D L' F R' F L' F L2 U L2 R' D L

Double X-Cross: x2 B' R2 . B' U L F U F2 B2
Cube-corner-3-cycle: B R B' R2 B R2 B' D R2 D' F' R F (1 move cancellation)

. = R2 U L2 U' R2 U L2 U' ( 2 moves cancellation)

Inverse Solution: x2 B' U L2 U' R2 U L2 U' B' U L F U F2 B' R B' R2 B R2 B' D R2 D' F' R F
Solution: x2 F' R' F D R2 D' B R2 B' R2 B R' B F2 U' F' L' U' B U L2 U' R2 U L2 U' B

27 moves

After the competition I also found a nice start based on my girlfriends beginning:

Premove: B2

F2L: R U2 R D2 U' B2 R' U' B' U B2 U' B'

13/14 Move F2L!

But the best I found with this was:

Cube minus 3cycle: B L U L' U B' U' B U' B' U (2 moves cancellation)


Comments and suggestions are welcome

 

[Stini]

hi guys.

I've recently gone back to FMC, and I suck at it. [my solves are usually 45 moves or so.] So I did one just now and wanted to post it, since it's only 37 moves with cancellation ^_^

Scramble: U D R' U2 B2 F R' L2 U L B2 F2 L U' L B' L2 U' R2 L2 F R U' R2 B2 F' D B F' L

B U' R2 (3/3) [2x2x2]

R' U' R F' L' F U2 (7/10) [2x2x3]
U' L U x' z (3/13) [EO]
U R2 U' R' U R2 (6/19) [F2L]

R' U' R U' R' U2 R y' (7/26)

R U R' U' R' F R2 U' R' U' R U R' F' (14/40) [40 - 3 = 37 moves]

I guess I got lucky. Solved with Petrus and T-Perm only.

This T-perm is quite useful to know: R2 u R2 u' R2 y' R2 u' R2 u R2
It's just a fairly known F2L-trick done twice (from different angles). It also works for Domino. Nice solve nevertheless.

 

[CharlesOBlack]

I'm pretty new as well, and looking back on my progress, I think I can offer a few tips.

You have a really good start here; it's after where you should be looking to shave a couple moves off. Also, you don't end up with any pairs after your EO. So by adding a single move for a different and suboptimal EO, you might end up with a more favorable case.

Looking back at my solutions, this is pretty much on par with my better solutions before I started using more advanced FM techniques. So from here I recommend you learn how to use a keyhole, insertions, pre-moves, and NISS--probably in that order.

While there's a lot to go through, nearly everything you need to know is in this thread and this one: http://www.speedsolving.com/forum/showthread.php?t=1566 If you can't find anything here, it's somewhere on the net.

The last thing I recommend is to not put cube rotations in your notation--it's easier to catch cancelations and it's simpler for sharing them.

Good luck learning the techniques! I'd recommend taking them on one at a time.

I know how to use insertions, keyhole and pre-moves. I just didn't use them on that solve
Yeah, I know about the cube rotations. I catch cancellations, so I don't really care, but I might change (even though it's hard...) for the sake of sharing them.

This T-perm is quite useful to know: R2 u R2 u' R2 y' R2 u' R2 u R2
It's just a fairly known F2L-trick done twice (from different angles). It also works for Domino. Nice solve nevertheless.

:O...... that is such an easy t-perm, and it's soooo short. That would've made my solve sub-35. XD I'll learn this one by heart for sure.

 

[Kryptonite]

I know how to use insertions, keyhole and pre-moves. I just didn't use them on that solve

Haha, my bad. What is it they say about making assumptions...

 

[CharlesOBlack]

I know how to use insertions, keyhole and pre-moves. I just didn't use them on that solve

Haha, my bad. What is it they say about making assumptions...

No problem.

I still have some trouble with inverses, though. Mind if I ask a few questions?

I have the following scramble:

A B C'

I solve the inverse of the scramble (C B' A') with D E' F.

Thus, the solve for the original scramble is...? I tried both D E' F and F' E D', but they don't seem to work. Help me?