FMC ranking system

[bcube]

Hi,

Our mission is to have more competitions in more countries with more people and more fun, under fair and equal conditions.
- WCA Mission Statement

As I understand it, FMC can be well quantified and measured. We have already a relation between solved and unsolved cubes in MBLD. Why don´t we have a relation between optimal and used number of moves in FMC?

I mean, if an optimal solution requires 18 moves, how is fairness and equal conditions guaranteed for all competitors (when it is mathematically impossible to tie/break the world record of 16 moves in such a case)?

Assuming all FMC scrambles are backward available (so the optimal number of moves is known for a given scramble), what are the arguments against the following ranking system: (optimal number of moves) / (used number of moves)? If you want points, multiply by 100.

Edit: I am more and more sure that option number 3 in this post is the way to go as it is solving all my objections and fairness concerns. For the mean of 3 format, all 3 solutions should have the same optimal move count.

 

[artless1der]

You seem to be confusing the WCA's aspirations for the enviroment(s) is which cubing competitions are held with regulations about the individual event FMC ???
¯⁠\⁠(⁠°⁠_⁠o⁠)⁠/⁠¯

It seems pretty straightforward...

Each FMC competitor gets the same scramble(s)...
and they use the time given to try to find solution(s) using as few moves as they can conceive

Elsewise, you might consider watching several of the videos on COLORFUL POCKETS YouTube channel for more information and to get some help with your difficulties

 

[Jorian Meeuse]

I don't think it would be very practical if every optimal solution ever found by any competitor is a guaranteed eternal WR.

Edit: Also, the event is 'Fewest Moves Challenge', not 'Proportionally Closest to Optimal Challenge'.

 

[bcube]

Each FMC competitor gets the same scramble(s)...

I am afraid that is not the case. That´s why I am proposing a system in which different scrambles are mutually comparable in a more fair way (in my opinion).

I don't think it would be very practical if every optimal solution ever found by any competitor is a guaranteed eternal WR.

That was my concern too. But I am looking at it as if the best possible outcome doesn´t guarantee a WR, I don´t know what should.

Edit: if there were fair and equal conditions for all FMC competitors from the beginning (by that I mean scrambles requiring the same number of optimal solving moves), that is exactly what would happen.

Edit: Also, the event is 'Fewest Moves Challenge', not 'Proportionally Closest to Optimal Challenge'.

Point taken (although event name is not my point)

 

[artless1der]

I don't think it would be very practical if every optimal solution ever found by any competitor is a guaranteed eternal WR.

Edit: Also, the event is 'Fewest Moves Challenge', not 'Proportionally Closest to Optimal Challenge'.

[Edit:[€d!t:]]
As well, the cubing event nomenclature is "Fewest Moves Count"
not
"Proportionally Closest To And Simultaneously Not Diametrically Opposite To Optimal Quantification Yield"

 

[Billabob]

It would be an interesting metric to compute and use to compare existing results (I might take a look in a few days). But I wouldn't want to see it used by the WCA officially.

 

[bcube]

I made a quick check for all 46 single results better than 20 moves, if anyone is interested:

Spoiler

          as of 13. 02. 2024
          ------------------

Cale Schoon            100 WR    18/18
Chad Batten            100 WR    18/18
Harry Savage           100 WR    17/17
Sebastiano Tronto      100 WR    16/16

???                    100 WR    20/20

???                    95        19/20

Wong Chong Wen        94,736    18/19
AJ Kamal            94,736    18/19
Ang Chin Zhen        94,736    18/19
Anonymous            94,736    18/19
Baiqiang Dong        94,736    18/19
Bartosz Karpiński    94,736    18/19
Marcel Peters        94,736    18/19
Mark Boyanowski        94,736    18/19
Szymon Jeziorski    94,736    18/19
Firstian Fushada    94,736    18/19
Louis Sarthou        94,736    18/19
Shuto Ueno            94,736    18/19
Simon Kellum        94,736    18/19

Evan Brown            94,444    17/18
Guido Dipietro        94,444    17/18
M. M. Ortega López    94,444    17/18
Theodor Nordstrand    94,444    17/18

Yiwei Liu            94,117    16/17

???                    90        18/20

Fabio Schwandt        89,473    17/19
Jaehyun Jeong        89,473    17/19
Jayden McNeill        89,473    17/19
Krish Shah-Nathwani    89,473    17/19
Levi Gibson            89,473    17/19
Owen Widdis            89,473    17/19
Radomił Baran        89,473    17/19
Robert Yau            89,473    17/19
Tim Wong            89,473    17/19
Uladzislau Ushakov    89,473    17/19
Wojciech Rogoziński    89,473    17/19

 

[bcube]

I mean, if an optimal solution requires 18 moves, how is fairness and equal conditions guaranteed for all competitors (when it is mathematically impossible to tie/break the world record of 16 moves in such a case)?

Aside from already proposed solutions to this problem, namely:

1. Using fixed number of optimal moves to solve the cube for all scrambles (which is destructive to the current ranking system)
2. Modifying the current ranking system to reflect non-fixed number of optimal moves to solve the cube for all scrambles (which is non-destructive to the current ranking system)

this is an alternative to point 2: compare apples with apples, not oranges. By that I mean make mutually comparable results separated into categories sorted by optimal number of solving moves, such as FMC16, FMC17, FMC18 etc., similarly to categories boys, juniors, seniors in sport events. However, ranking system for the mean of 3 format should be modified in this case, too - there are several ways how to do that, depending on the scrambles used for this format.

 

[bldcuber314]

Interesting idea, but would that mean each category would have separate records? I think we can agree that that wouldn't be realistic.

 

[SkewbIsTheBestEvent]

Hi,



As I understand it, FMC can be well quantified and measured. We have already a relation between solved and unsolved cubes in MBLD. Why don´t we have a relation between optimal and used number of moves in FMC?

I mean, if an optimal solution requires 18 moves, how is fairness and equal conditions guaranteed for all competitors (when it is mathematically impossible to tie/break the world record of 16 moves in such a case)?

Assuming all FMC scrambles are backward available (so the optimal number of moves is known for a given scramble), what are the arguments against the following ranking system: (optimal number of moves) / (used number of moves)? If you want points, multiply by 100.

I understand that having fair and equal conditions, but virtually all sports require luck in order to get a good result.
The point of World Records is that it is broken to extents we have never seen before in a sport. If the World Record was fixed, then we know it will never be broken again, which isn't very fun ("more competitions in more countries with more people and more fun, under fair and equal conditions.)

Also, the factor of luck is in every event. Is it fair that we were all not in Grudziądz Poland on March 19th, 2019, when Maciej Czapiewski got the 2x2 WR with a 4 mover? (It was just R U R' U', arguably easiest 4-mover). ‎I'm sure a lot of us would be able to Sub-0.49 that scramble, but we weren't at that comp.

Having luck in a sport is what makes it fun. And if you have more skill, you can reduce the necessity for luck.

 

[bcube]

would that mean each category would have separate records?

Yes.

I think we can agree that that wouldn't be realistic.

You might be surprised to know this is already real in physical sports like athletics, cyclocross or boxing, as well as mental sports such as chess. And for a good reason: fair and equal conditions.

virtually all sports require luck in order to get a good result.

I don't care how lucky the scramble is as long as it is fair for all competitors. By a fair scramble I consider a scramble every competitor could theoretically break/tie a world record with, just like in any other WCA event.

The point of World Records is that it is broken to extents we have never seen before in a sport. If the World Record was fixed, then we know it will never be broken again, which isn't very fun ("more competitions in more countries with more people and more fun, under fair and equal conditions.)

What you are describing is fitting well for the current ranking system. Imagine someone gets a scramble requiring 1 move to solve (if this is covered by the regulations, leave them behind for a moment). Please, stop systematically discriminate the vast majority of competitors. Another example: imagine someone gets a scramble requiring 1 move to solve and he will solve it in 15 moves - if WCA really wants to recognize that as a WR then I quit trying to change anything immediately.

Is it fair that we were all not in Grudziądz Poland on March 19th, 2019, when Maciej Czapiewski got the 2x2 WR with a 4 mover?

Yes. You have plenty of attemps to beat/tie that, and more importantly, in each attempt you can actually do it (so you are not being systematically discriminated).

 

[bldcuber314]

Say that a countries NR for FMC 20 is 74 because 20 move scrambles are exceedingly rare. If it is not at the top level, the optimal solution has virtually no difference on a competitor's score. There will still be 'lucky' scrambles, i.e scrambles with a rarer optimal move count as they have less competition.

Do we really need to plan for 2 move scrambles? I think if a 2 move scramble came up in any 3x3 event the FMC ranking system is the least of our worries.

I really wish I had plenty of official attempts to beat a 0.49. Zayn Khanani hasn't even tied that. There are only 6 scrambles that lucky out of 3.6 million. Are we really going to have plenty of those scrambles?

 

[OtaMota]

The problem here is that singles don’t determine who is the best and they are overrated. This system only makes sense if we determined best cubers by single. If you compare between cubers the average of their results in the past let’s say 12 months, you can find who the best cuber is without having much luck involved, and the system proposed would have much benefit.

 

[bldcuber314]

Also, only 4-5(I'm not sure, correct me if I'm wrong) people have had optimal solutions officially. Even top FMC solvers are barely limited by the optimal solution for the scramble.

 

[bcube]

The problem here is that singles don’t determine who is the best and they are overrated.

Even though I don't see what is the problem here, I agree single solves are less representative of competitor's skills in comparison with a mean.

This system only makes sense if we determined best cubers by single.

Please elaborate/explain, I am not getting your point.

If you compare between cubers the average of their results in the past let’s say 12 months, you can find who the best cuber is without having much luck involved

Good luck trying to change mean of 3 to average of 12 months with WCA

 

[bldcuber314]

Even though I don't see what is the problem here, I agree single solves are less representative of competitor's skills in comparison with a mean.



Please elaborate/explain, I am not getting your point.



Good luck trying to change mean of 3 to average of 12 months with WCA

Instead of having this whole debate and trying to make singles completely fair to every FMC solver on the planet with zero luck involved, why not just rank people by mean. The current WR mean is 20, which clearly wasn't optimal. Nobody has ever had an optimal mean.

 

[SkewbIsTheBestEvent]

I don't care how lucky the scramble is as long as it is fair for all competitors. By a fair scramble I consider a scramble every competitor could theoretically break/tie a world record with, just like in any other WCA event.

I see your point.

What you are describing is fitting well for the current ranking system. Imagine someone gets a scramble requiring 1 move to solve (if this is covered by the regulations, leave them behind for a moment). Please, stop systematically discriminate the vast majority of competitors. Another example: imagine someone gets a scramble requiring 1 move to solve and he will solve it in 15 moves - if WCA really wants to recognize that as a WR then I quit trying to change anything immediately.

???

How am I systematically discriminating the majority of competitors?

15 moves is WR, so it should be recognized as such. Optimal solution would be 1, but since their solution was 15 moves, it is still WR.

Yes. You have plenty of attemps to beat/tie that, and more importantly, in each attempt you can actually do it (so you are not being systematically discriminated).

The highest optimal solution is 11 moves. You would need 26.1 TPS in order to beat the current WR. That is impossible to achieve under the current regulations.

 

[bcube]

How am I systematically discriminating the majority of competitors?

If 25 persons get a 10-mover and 99 975 persons get a 11 and above mover, you are discriminating 99 975 persons for not giving them a chance to be best (side note: it is not clear to me how this can be named a competition). It applies to the current ranking system, as should be obvious from the first post in this thread. Several solutions to this problem have been proposed (that is the point of this thread).

The highest optimal solution is 11 moves. You would need 26.1 TPS in order to beat the current WR. That is impossible to achieve under the current regulations.

Our definition of impossible differs then. Mine is 0% of probability. Therefore, if you achieve at least 26.1 TPS, I consider it as possible.

Compare the same logic with 10 and 11 mover above to see my point.

15 moves is WR, so it should be recognized as such. Optimal solution would be 1, but since their solution was 15 moves, it is still WR.

I sincerely hope WCA's opinion on that is different

 

[bldcuber314]

If 25 persons get a 10-mover and 99 975 persons get a 11 and above mover, you are discriminating 99 975 persons for not giving them a chance to be best (side note: it is not clear to me how this can be named a competition). It applies to the current ranking system, as should be obvious from the first post in this thread. Several solutions to this problem have been proposed (that is the point of this thread).



Our definition of impossible differs then. Mine is 0% of probability. Therefore, if you achieve at least 26.1 TPS, I consider it as possible.

Compare the same logic with 10 and 11 mover above to see my point.



I sincerely hope WCA's opinion on that is different

"Discriminating' implies that people, before getting lucky or unlucky scrambles, have different chances of this. Everyone is treated equally, with the same chance of a lucky scramble. This is the way it is in every WCA event. Also, it is not like most experienced FMC competitors don't get a chance at WR. 20 move means are achievable every time, and 16 move singles are achievable roughly every 40 scrambles(?)

 

[Clock_Enthusiast]

If 25 persons get a 10-mover and 99 975 persons get a 11 and above mover, you are discriminating 99 975 persons for not giving them a chance to be best (side note: it is not clear to me how this can be named a competition). It applies to the current ranking system, as should be obvious from the first post in this thread. Several solutions to this problem have been proposed (that is the point of this thread).



Our definition of impossible differs then. Mine is 0% of probability. Therefore, if you achieve at least 26.1 TPS, I consider it as possible.

Compare the same logic with 10 and 11 mover above to see my point.



I sincerely hope WCA's opinion on that is different

I think systematic discrimination is an enormous claim in the context of this discussion. The WCA isn't selecting groups of people and arbitrarily developing a caste system for scramble difficulty. From my understanding, you’re essentially calling for the removal of complete randomization in the WCA (outside of its current set scramble minimums), which is completely counterproductive to distancing oneself from bias in the generation of scrambles. Do you suggest that scrambles should be handpicked by delegates or another party?

As you did mention previously, switching up the means of generating scrambles would cause all prior rankings to become incomparable to new results. This is completely true, and I fully agree with you that it would destroy the current ranking system. For instance, if 2x2 minimum move count is increased to 5 or 6, it would not only present challenges in breaking current records, but also invalidate results achieved under a different system.

It is unfortunate that not every person is presented with scrambles of equal “difficulty” within and across competitions. It isn’t ideal, but it also isn’t unfair. Scrambles are random, and this randomization is equally applied to all. We must also consider the optimization of scrambles, which inherently requires skill to fully recognize and effectively execute. The best FMC singles are not typically random--they are solutions developed by those of the highest skill in move optimization, which simply is the event. When Sebastian Tronto got his 16, did everyone else in the round also get a 16? The answer is no, as the next lowest was tied at 21 by several competitors.

Correct me if I’m wrong, but in a round of FMC, don’t all competitors also get the exact same scrambles? Not every competitor is guaranteed the optimal solution.

In the context of other events, couldn’t one also have a less optimal scramble yet still execute it faster than a competitor with a more optimal scramble? For example, everyone in one round of 2x2 gets a 4-mover, but someone in the next round who (extremely luckily) receives a 5-move optimal solution scramble completes the solve faster. Should we make separate categories for both optimal solution AND ability to execute an optimal solution?