[Help Thread] FMC Discussion

[puzzlemaster]

US Nationals Round 2 Scramble:
B2 D' U B2 R2 D2 R F2 D' L2 D2 F L' U R2 B2 L' D2 U'

2x2x2 block: R' B R2 D L U'
2x2x3 Block: D2 F U F U'
Bad Edges: L' F L

Now the luck begins...

x2 y U2 R2 U2 R2 U2 R2 U' R2 U2

8 move corner commutator: x2 D2 R U R' D' R U' R' D'

Messed up cancellations nats...should've been 30 moves...instead i screwed it up and got 33...

Oh well.

 

[irontwig]

bamilan: It's a inverse solve. This will make more sense:
B2 L2 U R2 B' L2 F' U' R U R' F2 R F' U2 L' U2 B' U2 B U2 L U2 R U R U' R2 D U' R' U2 R D' U2 R U2

 

[trying-to-speedcube...]

4x4FMC: 92

Scramble (SiGN notation): f B F2 r' f D' u r U2 D B' f r U' L' D U2 L' u' R' L' B2 f2 F2 u U' D' r' B F' f' U2 R' L' f r2 U' f' u' R2

Solution: (SiGN notation):
F l' F l2 F l b R' b'
R L u' L' u L2 u2
B' d2 F2 B d B' d2 L'

R' D' R2 D2 R' u'
B' L B D' L2 D r
F' R2 F D L D' B' R2 B r'
U B u L U' L' u'

L' y2 x'

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

I meant this solve

It was a NISS-ish approach. Scramble + reduction didn't leave a good 3x3 part, so I used the inverse of the reduction as a premove on the inverse scramble, and then put the inverse of the 3x3 solution on that at the end of my normal solution.

NISS-confuzzlement ftw.

 

[Johannes91]

The main reason why I like STM over HTM is that I think STM provides more useful cancellations and would, overall, leave no move-count restrictions (except the archaic R L move, being counted as one; and R r, as such. I've seen it done with some freaky notation at one time or another).

Why not take it a step further and count all 2-gen sequences, such as Sune, as one move? That'd provide even more cancellations and leave even less "move-count restrictions".

 

[Kryptonite]

The main reason why I like STM over HTM is that I think STM provides more useful cancellations and would, overall, leave no move-count restrictions (except the archaic R L move, being counted as one; and R r, as such. I've seen it done with some freaky notation at one time or another).

Why not take it a step further and count all 2-gen sequences, such as Sune, as one move? That'd provide even more cancellations and leave even less "move-count restrictions".

Haha, Hell, by KTM (Kryptonite Turn Metric), I often solve the cube in a single move.

 

[blah]

US Nationals FMC #2.

(Premoves) Inverse scramble: (L' U2) U R' F U2 B D B2 D' F2 D F D F2 U2 F' D' L2 U

2x2x3: R2 B U' B' L2 B R U (8/8)
Triple x-cross: L' U' F' U (4/12)
Edges: F' D F' . D' F' (5/17)
Undo premoves: L' U2 (2/19)
Insertion #1 at .: F2 D B D' .. F2 D B' D' (5/24)
Insertion #2 at ..: D F' U' F D' F' U F (6/30)

 

[Kryptonite]

Perhaps the best solution that we can have now for the FMC is have FMC for different move counts. One for QTM, HTM, and STM. That way, people can use everything in them to solve a cube and not have unnecessary worries about whether one move to them is two moves to another, and vice versa. Is it really that hard to count a solution in three different metrics? Why only use one metric?

You have a lot of excellent points in this post. All NxNxN cubes are fundamentally connected, yet we can't seem to unify them under a single metric, or even for just the 3x3x3 cube.

I think when it comes down to it, we do have to decide on one. In a lot of ways we should just be able to go, hey, check out my 33 STM solve! And everyone is cool with that. And while cubing has been incredibly collaborative, there's a lot of fun to its competitive roots. And when it comes down to it, QTM, HTM, and STM don't compete easily.

If we really want to develop a way to interchange them, I would propose something like this:

After completing your solution, we solve the cube optimally in each metric. Take the optimal move counts in each metric and divide the corresponding human solution move count, multiply by say 1000 to get your "points." This number is essentially the efficiency of your solve, and the higher points the better. 1000 of course being that you solved the cube optimally. Since this should eliminate metric-bias, take your highest score.

For example, my FMC 305 result (classic) was ( 29/[35]/{28}/{[34]}). Once we have the optimal moves count for each turn metric, we would do this:

QTM = 19 / 35 = 543
HTM = 18 / 29 = 621
STM = 17 / 28 = 607

My score for this solve would be 621, as it's the closest to optimal by any given turn metric.

I'm not suggesting we do this, I'm simply saying that we would need a system of this complexity to accurately compare varying metrics, and who knows if this would/wouldn't give an advantage to a given metric anyway.

EDIT: Supplied correct optimal move counts for the scramble instead of guesses. Here is the multi-optimal solution: F' D F' L U' B L' B' F' R D2 U' F {R L'} D' F' U (19q*, 18f*, 17s*)

 

[Kryptonite]

I like that idea a lot (even though you said that it doesn't necessarily need to be implemented). It kind of reminds me of the SAT/ACT college entrance exams. I think it should be. Maybe in the future others will agree as well.

Haha, we better add in the 200 points for signing your name then! Maybe I'll play around with it some.

 

[qqwref]

Perhaps the best solution that we can have now for the FMC is have FMC for different move counts. One for QTM, HTM, and STM. That way, people can use everything in them to solve a cube and not have unnecessary worries about whether one move to them is two moves to another, and vice versa. Is it really that hard to count a solution in three different metrics? Why only use one metric? I bet this will make FMC much more popular too.
[...]
I applaud all of you who are experts at HTM fewest moves. But, in all honesty, why can't those who are strong in STM (or QTM) also enjoy a satisfaction that they deserve as well.

There is no time in a competition to do three types of FMC events. One hour is already a big deal for the organizers, especially at one-day competitions (well, it's usually not even held at those!). You can't hold all 3 at once because then the best people will only be able to do one and thus winning STM or QTM will depend pretty much entirely on not having anyone better choose to do that event. In an unofficial setting you can use any metric, even axial turn metric if you wish, but for official solving it is absolutely necessary to standardize, and HTM is both traditional and logical (not that STM isn't logical, but I don't think QTM is to a non-theorist).

 

[TMOY]

It is perfectly possible to do 2 or even 3 FMC attempts in the same competition, it has already been done for example in several French competitions (Lyon 2009 and 2010, Clermont 2009, Montpellier 2010). Instead of two HTM attempts you could as well hold one HTM event and one STM event.

 

[Yes We Can!]

Scramble:
L D2 R F' B2 D' F' B D' U2 R2 F D' L D' R D2 U2 B2 F' U F2 U B' D'

Solution:
D' L' R' F' U L' (X-Cross)
x2 F' U' F U R U' R' y (Flip edge for following commutator + pairing 2nd pair)
x2 r U R' U' r' F R F' (commutator to solve BR pair)
x2 F' U F U' B' U B U2 L U L' (last 2 pairs)
y' L' U' L U B L F' L' F B' (forced PLL skip by using mirrored OLL).

= 42 moves.

Wow, I suck. I also found another PLL skip solution on this, also 42 moves >_>.
I think everything of this was bad, except for the X-Cross, which was very good I think.

 

[okayama]

A friend of mine got an easy scramble. His time was 7.92.
Scramble: B2 F2 L' R2 D' L2 F D U' B F' D2 U B2 F2 D R2 B U2 B D' B R2 D U'

I decided to solve this scramble for FMC, with a Fridrich-like method.

Solution: z2 F' D U R B' R2 U2 F U B U F' B' R B U' L' U R' L2 U' R' U L' U' R U2 (27 HTM)

Cross: z2 F' D R'
1st slot: R U R B' R2
2nd slot: U2 F U B
More pair: U F'
3rd slot: B' R B R'
4th slot: R U' L' U L R'
COLL: L U' R' U L' U' R
AUF: U2

I didn't try to search a corner 3-cycle insertion. Almost linear solve!

 

[sz35]

This is the solve from this week's weekly comp. I would like you guys help me improve the ending.
38 HTM
Sramble: B2 L' R F' U F' L' R2 U B2 R2 B' R2 U' L2 U2 L2
Solution: L' B L' F D R' F L' B L B' R' U R' U' R U R' U' R U R' U' R2 D R D' R' D B D B2 D L' F L2 F' D

Spoiler

2*2x2x2: L' B L' F D R' F L' B L B'
Switch to Inverse with premoves: B L' B' L F' R D' F' L B' L (11)
Leave 4 corners: D' F L2 F' L D' B2 D' B' D' R D R' D' (14)
Solve 4 corners: R2 U R U' R' U R U' R' U R U' R (13)

 

[Cride5]

I decided to try the 'hardest' scramble on cube20.org

The scramble is: F U' F2 D' B U R' F' L D' R' U' L U B' D2 R' F U2 D2

After seeing no 1-move blocks on both regular and inverse scramble I decided to go for a ZZ-based start, and managed this 35-mover (not great):

Spoiler

EO: F R L B' U' D' F (7)
Line+Pair: U' R' U' (3/10)
RH Side: D2 R2 L D2 * R (5/15)
F2L-1 + 1x2x2: D' L D L' D2 (5/20)
3-corners: L2 D L' D' L D L' D' L (9/29)
Insert at *: B' L' B R B' L B R' (8/37)
2 moves cancel:

I also found this mediocre blockbuilding start, but no good finish:

Spoiler

2 pairs: L' B2 (2)
2 1x2x2's + pair: F R L B U2 B2 (6/8)
F2L-1, no EO: D B' R B U D' F D R' U2 (10/18)

What do you guys think of the scramble? Hard for a human as well as a computer?

 

[Kryptonite]

Hey guys, I just put together the last four FMC results using the idea I had for comparing solve efficiency, which would hopefully let people work in any give metric.

As a quick disclaimer, I am not doing this to promote some random idea I had. I am simply pursuing an avenue of curiosity and figured it was worth sharing. I am by no means convinced that this kind of hybrid system is better than just sticking with HTM.

Here's the break down of optimal solve length done in each metric (sorted QTM, HTM, STM):

305: 19, 18, 17
304: 20, 18, 17
303: 20, 18, 18
302: 22, 17, 16

Here's the conversion. C is for Classic, O is for Official. I put the score, and which metric it came from. A big thing we are looking for is that there aren't any big upsets, since everyone was using the same metric. The "^" show anyone who went up a position. All the "^" boil down tie-breakers (same result in HTM, but one person had a more efficient solve in another metric) except for one. In the Official FMC 302, Всемирный jumps Tomoaki despite a 30 v. 29 HTM, because his QTM is 36. Comparing efficiencies, 36 QTM is more efficient than 29 HTM.

Could this satisfy STMers without making HTM obsolete? Not sure either way yet, but it's interesting to see it in action.

 

[iasimp1997]

I keep misreading this to be the "FMJ thread".

 

[Ranzha]

Hey guys, I just put together the last four FMC results using the idea I had for comparing solve efficiency, which would hopefully let people work in any give metric.

As a quick disclaimer, I am not doing this to promote some random idea I had. I am simply pursuing an avenue of curiosity and figured it was worth sharing. I am by no means convinced that this kind of hybrid system is better than just sticking with HTM.

Here's the break down of optimal solve length done in each metric (sorted QTM, HTM, STM):

305: 19, 18, 17
304: 20, 18, 17
303: 20, 18, 18
302: 22, 17, 16

Here's the conversion. C is for Classic, O is for Official. I put the score, and which metric it came from. A big thing we are looking for is that there aren't any big upsets, since everyone was using the same metric. The "^" show anyone who went up a position. All the "^" boil down tie-breakers (same result in HTM, but one person had a more efficient solve in another metric) except for one. In the Official FMC 302, Всемирный jumps Tomoaki despite a 30 v. 29 HTM, because his QTM is 36. Comparing efficiencies, 36 QTM is more efficient than 29 HTM.

Could this satisfy STMers without making HTM obsolete? Not sure either way yet, but it's interesting to see it in action.

This pleases me.
HTM isn't obsolete--I just find that cancelling those (R L')s and such are more desirable. Additionally, one can then go without the general ruling out of slice moves, which I always have to whenever I try using HTM.

 

[qqwref]

I see a problem with this approach: since a solver wouldn't know what the optimal solution is in HTM and QTM, they couldn't tell whether it was worthwhile to optimize for QTM or not. Some scrambles (like 305) would have only a 1 (or 0) move difference between HTM and QTM, and thus you can gain no benefit from optimizing specifically for QTM (your HTM score will almost certainly be better). However, some scrambles (like 302) would have a 5 (or higher!?) move difference, and thus optimizing for HTM is useless because a good QTM solve will rank much better. So, while your idea sounds good, in practice the optimal numbers are impossible to determine or even guess from the scramble, so it is basically luck that determines which one of two similar solutions will get ranked better.

 

[Kryptonite]

I see a problem with this approach: since a solver wouldn't know what the optimal solution is in HTM and QTM, they couldn't tell whether it was worthwhile to optimize for QTM or not. Some scrambles (like 305) would have only a 1 (or 0) move difference between HTM and QTM, and thus you can gain no benefit from optimizing specifically for QTM (your HTM score will almost certainly be better). However, some scrambles (like 302) would have a 5 (or higher!?) move difference, and thus optimizing for HTM is useless because a good QTM solve will rank much better. So, while your idea sounds good, in practice the optimal numbers are impossible to determine or even guess from the scramble, so it is basically luck that determines which one of two similar solutions will get ranked better.

Excellent point, and something I've given some thought to. I see two solutions, neither perfect. First, we let the solvers know the optimal move counts per metric ahead of time. The judges would have to know eventually, so why not let the solvers know? When it comes down to it, the solver must be good in the best metric for the scramble.

The second solution, I think, is to just never let the solvers know. The idea behind the system is that hopefully on average STMers and HTMers (and QTMers) win evenly. This however doesn't mean there will be an even playing field every time. So, if you don't know the optimal for scramble (and you're right, there'd be no way to figure it out), you might as well use what metric you do best. Some times you'll get lucky, other times unlucky, but hopefully it's the same for everyone in the long run.

Neither is great, but I think either could work. Any thoughts?

 

[DavidWoner]

Nationals solve 1

(premove) inverse scramble: (B') U D2 L B2 R2 U' L F' D2 L2 D F2 R' D2 R2 B2 U' D B2

2X2X2: U R2 U R

2X2X3: D' L D' L2

3xcross: B D B2 D' B.

leave 3 corners: D2 B D2 R D' R' B'

Insert F' U2 F D' F' U2 F D at the . to cancel one move, resulting in a 27 move solution. Pretty textbook FM solve, rather boring in my opinion. Wish I could have canceled more than one move :/


Nationals Solve 2

(premove) inverse scramble: (U') U R' F U2 B D B2 D' F2 D F D F2 U2 F' D' L2 U

2X2X2: B' L2 U' B

F2L: R U2 R F2 U' R2 F' R

OLL: R' D' F' D. F D' R U' R' D R U F' U'

insert D2 F R' L D2 R L F D2 at the . to cancel one move, resulting in a 32 move solution. I failed at transcribing it though The ending of the solve was an epic failure anyway.