FMC ranking system

[SkewbIsTheBestEvent]

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).

Yes, but that is the factor of luck that comes with sports. Those 99,975 other competitors will get other chances to get the 10-mover in other competitions. I wouldn't consider that "systematically discriminating".

Also under your current proposed system, which I believe is (# of moves in the optimal solution) / (# of moves in competitors' solution) you are still "systematically discriminating".

Let's say the optimal solution is 19 moves, and a competitor found a solution that is 20 moves. That would be 0.95.
However, if the optimal solution was 15 moves, and a competitor found a solution that is 16 moves. That would be 0.9375.

In that scenario, the competitor would be 1 move off in both cases, but the points are different, due to how percentages work.

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.

26.1 TPS is 0% of probability under the current regulations. A human cannot achieve that.

I sincerely hope WCA's opinion on that is different

The WCA's opinion will not be different. That competitor with the 15 move solution even though the optimal solution was 1 move. Under the regulations, the competitor got a solution that was 15 moves, which breaks the current WR of 16 moves, so it is WR.

 

[OtaMota]

The WCA's opinion will not be different. That competitor with the 15 move solution even though the optimal solution was 1 move. Under the regulations, the competitor got a solution that was 15 moves, which breaks the current WR of 16 moves, so it is WR.

To add a non-FMC example, if a competitor gets a 2 mover on 3x3, even if they only get a 3.12, it’s still WR.

 

[bcube]

Do you suggest that scrambles should be handpicked by delegates or another party?

No.

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.

I am only talking about the FMC event.

Correct me if I’m wrong, but in a round of FMC, don’t all competitors also get the exact same scrambles?

Make results in rounds mutually comparable under fair conditions.

In that scenario, the competitor would be 1 move off in both cases, but the points are different, due to how percentages work.

That is the point.

 

[Clock_Enthusiast]

Make results in rounds mutually comparable under fair conditions.

Isn’t this achieved by providing all competitors with the same scrambles in a round of FMC?

 

[bcube]

Isn’t this achieved by providing all competitors with the same scrambles in a round of FMC?

Yes, it is not. One round of FMC (say, optimal 16-mover) can be currently compared only to other rounds of FMC with an optimal 16-mover under fair conditions, not all rounds of FMC.

Edit: you probably thought of a round as consisting of 3 attemps (hence plural in "same scrambles in a round of FMC"), while I was thinking of a round as consisting of 1 attempt in my answer above. The principle is the same for any round format though.

 

[JDAW6]

Yes, it is not. One round of FMC (say, optimal 16-mover) can be currently compared only to other rounds of FMC with an optimal 16-mover under fair conditions, not all rounds of FMC.

Basically u want every FMC to have the same scrambles

 

[Jorian Meeuse]

Our definition of impossible differs then. Mine is 0% of probability.

I'm not going to argue about the FMC stuff because that's not going to change anyway, but I just want to say this definition of impossible of objectively wrong. Say you let a computer select a random natural number. What is the probability that it is going to be less than a million? 0. Is it impossible? No.

 

[cubenerd74]

I will say this once and for all. FMC is held under equal conditions because everyone has the same chance of getting a lucky scramble.

I'm not going to argue about the FMC stuff because that's not going to change anyway, but I just want to say this definition of impossible of objectively wrong. Say you let a computer select a random natural number. What is the probability that it is going to be less than a million? 0. Is it impossible? No.

You have angered every mathematician ever by assuming 1/infinity=0(It does work in most cases tho).

 

[Jorian Meeuse]

You have angered every mathematician ever by assuming 1/infinity=0(It does work in most cases tho).

I never assumed 1/infinity=0. I only assumed that the probability of said random natural number being less than 1 million exists and that it is not smaller than 0. If it exists, we can call it p. Since p is not smaller than 0, it is greater than or equal to 0. If p > 0, that results in a contradiction with the fact that infinitely many natural numbers exists, because since all numbers (implicitly) have an equal probability of being picked, it would mean there are (1 million - 1)/p natural numbers (or (1 million)/p if you consider 0 a natural number). Thus, p = 0.

 

[artless1der]

If each FMC competitor gets the same scramble(s), then:

They should just involve the amount of time a competitor takes to achieve their solve(s) when consodering ranking of solvers' attempts

Similar to how they do with multi-blind

 

[SkewbIsTheBestEvent]

That is the point.

What?

They are one point off but got different scores. That is unfair even though they were both one point off, which negates the point of you creating this new system in the first place

 

[bldcuber314]

What?

They are one point off but got different scores. That is unfair even though they were both one point off, which negates the point of you creating this new system in the first place

While I support the main argument here, I think this is incorrect as proportional difference matters more than linear difference. For example, which is more difference in solving ability- someone who averages 2 mins on 7x7 versus someone averaging 2:10, or someone averaging 2 seconds on 2x2 versus someone averaging 12? Both are 10 second differences.

 

[ProStar]

Yes, it is not.

I don’t have an interest in getting involved with this discussion since it’s clearly going nowhere, but I did wanna point out this gem

 

[bcube]

I did wanna point out this gem

I guess I analyzed the question too logically. Hopefully it is stil semantically correct though.

Thank you for pointing it out, I learned something new . We can switch to my mother language if you expect me discussing in gramatically perfect language ;-).

 

[JDAW6]

We can switch to my mother language if you expect me discussing in gramatically perfect language ;-).

Yes, I would rather that./srs

 

[kubesolver]

I never assumed 1/infinity=0. I only assumed that the probability of said random natural number being less than 1 million exists and that it is not smaller than 0. If it exists, we can call it p. Since p is not smaller than 0, it is greater than or equal to 0. If p > 0, that results in a contradiction with the fact that infinitely many natural numbers exists, because since all numbers (implicitly) have an equal probability of being picked, it would mean there are (1 million - 1)/p natural numbers (or (1 million)/p if you consider 0 a natural number). Thus, p = 0.

Just wanted to nitpick about math here. The argument is invalid because there exists no random, uniform distribution of natural numbers. I.e. a distribution where each natural number has the same probability to be picked.

There exists a uniform distribution of real numbers between zero and one where this argument doesn't hold that well because probability of getting a number from any range >0 is actually >0

 

[artless1der]

I don’t have an interest in getting involved with this discussion since it’s clearly going nowhere

This.

 

[Jorian Meeuse]

Just wanted to nitpick about math here. The argument is invalid because there exists no random, uniform distribution of natural numbers. I.e. a distribution where each natural number has the same probability to be picked.

There exists a uniform distribution of real numbers between zero and one where this argument doesn't hold that well because probability of getting a number from any range >0 is actually >0

Haha that was indeed the only issue I had trying to write out my argument, I was hoping to get away with saying it was implicit but I should've expected otherwise on a speedsolving forum

Thanks for the correction!

 

[Spencer131]

Do you suggest that scrambles should be handpicked by delegates or another party?

Here's a kind of related idea. It could be cool to have a 1v1 format where each person scrambles for the other person.

That way there is no reliance on randomness or other people. It's just your direct competitor trying to make it hard for you, and vice versa.

 

[bldcuber314]

Here's a kind of related idea. It could be cool to have a 1v1 format where each person scrambles for the other person.

That way there is no reliance on randomness or other people. It's just your direct competitor trying to make it hard for you, and vice versa

That would make ranking extremely hard.