Why HTM for FMC?

[blah]

As the thread title says. I've looked around the forum and the WCA regulations but couldn't find an explanation to this.

That's question one. Here's question two: I'd also like to know why not QTM, STM or SQTM (or any other metric that I don't know of)?

How would other metrics have affected the development of FMC? Domino reduction/Human Thistlethwaite (if used at all) would be terrible for QTM. STM/SQTM would be sweet for Roux/CF. So, why not?

:confused:

(Apologies if this is in the wrong subforum, will a mod please move this to the appropriate subforum if necessary, thanks )

Edit: In the case of STM, is it because we choose to see corners and edges as permutable pieces but not centers? If so, why not? Why shouldn't centers be considered as permutable pieces?

 

[byu]

I don't know why HTM is used, buy I would prefer STM. ThY would make edge commutators a lot less moves

 

[stiwi griffin]

i want you to answer this:why not HTM for FMC?first,all the scrambles of the WCA follow that so if they changed that they will have to change everything.why? because people will get confused if for FMC was STM amd for 3x3x3 was HTM amd because is logical to try to have everything in the same way

about the centers,they are always in the same relative positions because even if you rotate them the yellow will be always(in my cube)opposite to white

 

[trying-to-speedcube...]

Try M' E2 M E2. It "permutes" the centers.

Though if you see the centers as fixed, it has permuted 12 edges and 8 corners, and has done an x2, but that's somewhat inconvenient

 

[Swordsman Kirby]

Try M' E2 M E2. It "permutes" the centers.

Though if you see the centers as fixed, it has permuted 12 edges and 8 corners, and has done an x2, but that's somewhat inconvenient

You can still write slice moves in your solution, though.

 

[blah]

I don't know why HTM is used, buy I would prefer STM. ThY would make edge commutators a lot less moves

Exactly.

i want you to answer this:why not HTM for FMC?first,all the scrambles of the WCA follow that so if they changed that they will have to change everything.why? because people will get confused if for FMC was STM amd for 3x3x3 was HTM amd because is logical to try to have everything in the same way

about the centers,they are always in the same relative positions because even if you rotate them the yellow will be always(in my cube)opposite to white

I don't mean to be rude, but your answer really doesn't make much sense to me, mind clarifying?

Try M' E2 M E2. It "permutes" the centers.

Though if you see the centers as fixed, it has permuted 12 edges and 8 corners, and has done an x2, but that's somewhat inconvenient

Sorry, but I don't get what you're trying to say. At least, I don't see which part of your post answers my question(s). I don't see why you need quotes when you say the centers are permuted. What's wrong with seeing centers as permutable pieces? Roux and CF work that way. I do 5x5x5 BLD that way.

 

[qqwref]

That's a very interesting question. I'm not sure why HTM is used (possibly because some people think of the 6 possible moves as being BDFLRU, since that's all that is necessary to solve the cube) and I guess it would make more sense for slice turns to be allowed (that is, use STM) because slices can be executed as a single move and thus should be counted as one.

I don't like the idea of using QTM/SQTM for FMC, though. My reasoning is that when you do a 180-degree turn on a cube it doesn't feel natural to consider it two moves, since in real life it's easy to execute as one move, even if mathematically it might be better in some ways to consider it as two. FMC ought to mirror real life - you should be able to execute a FMC solution in front of someone and have them confirm that it is (say) 28 moves. STM feels the most natural for counting moves in real life.

 

[blah]

I don't have any idea why, but I've always thought it was more "mathematical" to consider a 180-degree turn as two moves, since, well, you're "turning 90 degrees twice", so how on earth can that be considered equivalent to a 90-degree turn? Am I the only one here with this inexplicable gut feeling?

As for STM, to any layman who doesn't know what a cube's internal structure looks like, there's no reason for him to believe that the centers are fixed, because they aren't! They're just as permutable as the edges and corners. Take a look at this:

Any face turn = simultaneously moving 4 edges and 4 corners = 1 turn
Any slice turn = simultaneously moving 4 edges and 4 centers = 2 turns?!

Why? It's like centers are inferior to edges and corners or something.

 

[stiwi griffin]

i want you to answer this:why not HTM for FMC?first,all the scrambles of the WCA follow that so if they changed that they will have to change everything.why? because people will get confused if for FMC was STM amd for 3x3x3 was HTM amd because is logical to try to have everything in the same way

about the centers,they are always in the same relative positions because even if you rotate them the yellow will be always(in my cube)opposite to white

QUOTE]I don't mean to be rude, but your answer really doesn't make much sense to me, mind clarifying?[/QUOTE]

i mean,if FMC was changed into STM,for example,we have two posibilities

1º: all the scrambles will change into that metric,leading to a big change on the rules.

2º: just FMC will change,confusing people on how to count the moves etc.

centers:the centers are fixed so even if you rotate the cube they will be(in my cube):white-yellow,blue-green and red-orange

you weren't rude

EDIT:about what you said in the post above:
let's take this example,
SLICE MOVE:M'=2 HTM
M'=R' L X'=2 HTM
that is the explanation

 

[fanwuq]

I don't like the idea of using QTM/SQTM for FMC, though. My reasoning is that when you do a 180-degree turn on a cube it doesn't feel natural to consider it two moves, since in real life it's easy to execute as one move, even if mathematically it might be better in some ways to consider it as two. FMC ought to mirror real life - you should be able to execute a FMC solution in front of someone and have them confirm that it is (say) 28 moves. STM feels the most natural for counting moves in real life.

So should we also allow anti-slices as 1 moves the way Tony Snyder counts? (L R = 1 move)
For FMC, my style suits HTM very well and I can see a lot of people using Roux if the rules are changed to STM. Perhaps the HTM rule is the reason that I'm not very familiar with Roux.
Personally, I think SQTM is the best way to reflect execution speed. (RU 2gen H perm is very slow; at least that alg cannot be counted with HTM.) When ever I try to find algs with ACUBE, I set it to UFLR QTM optimal.

 

[Stefan]

If I remember correctly...

There were discussions about this in the past. After HTM had already been used as main metric in the unofficial competition, some of us felt STM would make more sense. In the end HTM persisted I think because of tradition and because at some point the main guy arguing for STM changed his mind. Not towards HTM but towards PTM, plane turn metric. Same thing (at least for 3x3x3), but different name/perspective The explanation being something like we have two hands and are twisting two things against each other on a plane.

Something like that.

 

[qqwref]

I don't have any idea why, but I've always thought it was more "mathematical" to consider a 180-degree turn as two moves, since, well, you're "turning 90 degrees twice", so how on earth can that be considered equivalent to a 90-degree turn? Am I the only one here with this inexplicable gut feeling?

No, it's true that mathematically it makes more sense to consider a 180-degree turn as two 90-degree turns, but that's not how people would see it if you performed your solution on a physical cube. Do an R2 and then ask random people on the street how many moves you did, and I'll bet you money that most people will say it was one move. It's probably easier to do math-related stuff in QTM, but when it comes to fewest moves solving a 180-degree turn intuitively feels like one move and I think it would be confusing to insist that it's two. Besides, it would ruin the beauty of having all standard corner commutators be 8 moves

So should we also allow anti-slices as 1 moves the way Tony Snyder counts? (L R = 1 move)

Not really... the point is that a slice is just pushing one slice in one direction, not that anything you can do all at once should count as one move. For instance, on bigcube patterns I use block turn metric, where any contiguous group of slices on one axis moved in the same direction counts as one move, and the reasoning behind that is that you can basically do that with one finger push, so it feels like a single move. Using axial metric on bigger cubes leads to silliness such as Rw R Lw' L2 being counted as one move.

Personally, I think SQTM is the best way to reflect execution speed. (RU 2gen H perm is very slow; at least that alg cannot be counted with HTM.) When ever I try to find algs with ACUBE, I set it to UFLR QTM optimal.

I agree. I also usually use sqtm when looking for good algorithms. As it turns out, for the hi-games applet, the metric of choice is aqtm (axial quarter turn metric), as you can do something like l R all at once but not something like R2 (even though they're equivalent ).

 

[rjohnson_8ball]

If I remember correctly...

There were discussions about this in the past. After HTM had already been used as main metric in the unofficial competition, some of us felt STM would make more sense. In the end HTM persisted I think because of tradition and because at some point the main guy arguing for STM changed his mind. Not towards HTM but towards PTM, plane turn metric. Same thing (at least for 3x3x3), but different name/perspective The explanation being something like we have two hands and are twisting two things against each other on a plane.

Something like that.

"Plane turn metric" is the way I think of HTM. I found it appropriate for computer programs, keeping centers fixed in place and rotating only the 6 faces. Also, most cubes were stiff years ago, so people found it easier to perform a slice move by rotating 2 opposing faces.

 

[Stefan]

"Plane turn metric" is the way I think of HTM. I found it appropriate for computer programs, keeping centers fixed in place and rotating only the 6 faces.

To clarify: Plane turn metric would also let you use double layer turns like r, in that case *not* keeping the centers fixed in plane. Or triple layer turns like 3R on the 7x7x7.

 

[rjohnson_8ball]

David Singmaster's Notes on Rubik's Magic Cube discussed fewest moves before Rubik's Revenge hit the market (I think), before an importance to distinguish double layer turns. He wrote about different metrics, but he seemed to push HTM. So maybe HTM became the precedent from way back then.

I see now that for cubes larger than 3x3 it would be useful to treat multiple layers (like a triple layer turn on 7x7) as one or even an interior slice (like in a 7x7) as one. This could be important soon when FMC involves cubes larger than 3x3.

 

[AvGalen]

...This could be important soon when FMC involves cubes larger than 3x3.

soon????? I don't really think that would happen. For a REAL bigcubes FMC (not just optimised reduction) I would need about 6 hours. FMC has changed from 90 minutes to 60 minutes and MBF now has a 60 minute time limit as well. We already have tournaments becoming 3 days because the amount of competitors and events require that. Having more events doesn't seem likely, especially not such a massively long event.

HTM makes sense.
* Give a random person a cube and they scramble it like this: R L R L2 (rotate cube) R L2 L' etc. They don't use slice turns.
* For speedsolving I perform M' as 1 move, but M as 2 moves (r' R)
* The cube has 6 faces so having 6 possible base-moves seems logical

And once a system is in place, it becomes very hard to change it.

With STM notations my FMC solutions wouldn't change much. They would just be shorter sometimes

 

[blah]

* The cube has 6 faces so having 6 possible base-moves seems logical

The cube has 9 slices so having 9 possible base-moves seems logical too

 

[TMOY]

And once a system is in place, it becomes very hard to change it.

That's the main problem. Because some people in the past years have decided that HTM should be used, now we're stuck with it. That's just stupid.
Personnally I tend to consider STM counting as the actual number of moves and HTM counting as the score it is worth under the WCA rules. For example; my winning solution at UK Masters was 27 moves, for a score of 33.

 

[blah]

And once a system is in place, it becomes very hard to change it.

That's the main problem. Because some people in the past years have decided that HTM should be used, now we're stuck with it. That's just stupid.
Personnally I tend to consider STM counting as the actual number of moves and HTM counting as the score it is worth under the WCA rules. For example; my winning solution at UK Masters was 27 moves, for a score of 33.

CF FTW.

 

[qqwref]

With STM notations my FMC solutions wouldn't change much. They would just be shorter sometimes

Using STM does remove a lot of the bias towards corners-first and Roux solutions. While those methods can undoubtedly be efficient, if slice turns are counted as two, they lose a lot of their power.