I feel bad for Kian?

[ImmolatedMarmoset]

ok so going back to the topic that this thread was meant for, Kian's kind of regressing at this point (pretty sad, but we have to give him credit for trying; it is very hard to keep moving at that level). Sean's 6.42 avg doesnt undermine Kian; it's just a pretty nice step contributing to the Rouxvolution.

I do hope that Kian keeps improving, however, specifically in OH because I see him as being the best in the world at this moment and I really want him to get OH WR average again.

 

[Krisrejas]

No es un AsR ni siquiera un NR, pero esta mañana, Sean Patrick Villanueva obtuvo un promedio de 6.42 3x3 en el campeonato de Luzón. Esto lo hace oficialmente mejor que Kian Mansour, quien es considerado el solucionador de roux más rápido, pero solo tiene un promedio de 6.52 en la actualidad.: eek:View attachment 10482

GJ a Sean Patrick!

parece que sean tiene un gran futuro, batió a leo borromeo y ahora va ala mundial en Australia, va aser una competencia increible y espero ver entre los primeros a sean roux vs cfop, vamos roux

 

[xyzzy]

Actually, if you take all of the times in the top 100 averages and find the mean time for each method, Roux is lower. The problem is, the sample size is low enough that I don't think it's a very good comparison. There's also a significant chance that I messed up. I don't have the numbers anymore, and the website is down, so I'll recheck it sometime later. Also, in case anyone was wondering, when I did the T-test, the p-value for whether Roux is faster was about 0.3, if I remember correctly. If I'm interpreting it correctly, that means that it's a pretty credible statistic. Then again, the sample size is by no means large, and that could bias the results a bunch.

Do you have a list of all the fast (sub-10, say) Rouxers so I can try to replicate this finding? A list of people who use neither CFOP nor Roux (ZZ, Petrus, etc.?) would also be very welcome, but this is probably harder to get.

(No offence, but (i) I find this somewhat suspect and (ii) a p-value of 0.3 is very weak evidence.)

 

[Sue Doenim]

Do you have a list of all the fast (sub-10, say) Rouxers so I can try to replicate this finding? A list of people who use neither CFOP nor Roux (ZZ, Petrus, etc.?) would also be very welcome, but this is probably harder to get.

(No offence, but (i) I find this somewhat suspect and (ii) a p-value of 0.3 is very weak evidence.)

Yeah, I don't put a whole lot of stock into it either, the main reason being that the only Roux solvers I knew of were Kian, Sean, Alex Lau, and Kaijun Lin. That's a tiny and very possibly inaccurate sample size. What I did was go through the list of top 100 averages on the WCA website and, using my trusty TI-84, plug the four Rouxers into L1 and everyone else into L2. The results after doing "2-SampTTest" are as follows:
\[ \begin{array}{l}2-\mathrm{SampTTest}\\\mu_1<\mu_2\\\\\mathrm t=-.3758836974\\\mathrm p=.3655357854\\\mathrm{df}=3.12593854\\{\overline{\mathrm x}}_1=6.9625\\{\overline{\mathrm x}}_2=7.0715625\\{\mathrm{Sx}}_1=.574361965\\{\mathrm{Sx}}_2=.405624478\\{\mathrm n}_1=4\\{\mathrm n}_2=96\end{array} \]
I don't know stats very well; I know what the means and standard deviation are and I have a general grasp of what the p-value is, but I don't really know the rest of the stuff. You probably understand it a lot better than I do. At any rate, given the lack of surety on who uses what method and the small sample size, it's not an incredibly convincing bit of evidence.

 

[PapaSmurf]

Reading this thread was quite funny. It started off with someone saying that the fastest rouxer at home isn't the fastest officially, then it turned into a Roux v CFOP debate with the CFOPers/anti-Rouxers giving no good argument, and the Rouxers not really giving any proper argument, just pointing out that the CFOP argument was flawed. I just think that now's the time for everyone to realise that ZZ-A is the future. But seriously, CFOP and Roux are both really good methods, as is ZZ, it's just that people haven't listened to the arguments for and against each method. Trust me, I've had a lot of discussions on ZZvsCFOPvsRoux(vsPetrus a little bit) and the conclusion that I've come to is ZZ≈Roux>CFOP but other people have come to many different conclusions. The hard part of the whole debate is that it's impossible to prove with evidence of 'this person is the fastest, therfore their method is the best.' The easiest counter example is corners first in the 80s. Instead, the best way, imo, is to look at the concrete evidence in movecount and to look at the not-too-super-hypothetical ergonomics. For example, you could point out to different U perms. At the end of it all, you weigh out the evidence and choose the method that you think is the best. I have come to the conclusion of ZZ or Roux, and as I already mained ZZ, I chose ZZ. Some people would rather Roux, some would rather another, less well known method such as LEOR. It all depends on the person. In conclusion, use good evidence.

 

[Etotheipi]

As it has already been pointed out, CFOP had been around for 22 years before Roux came. And within those 16 years Roux has existed it has gotten within fractions of a second from CFOP. Considering that many many more people use CFOP, and has existed longer, while Roux has many less users and is younger, proves that the method itself has more potential. You cant just look at the solve time. There are so many factors that so many people fail to notice, which gives them a very distorted picture. Now go tell me go wrong I am =D

 

[ImmolatedMarmoset]

Do you have a list of all the fast (sub-10, say) Rouxers so I can try to replicate this finding? A list of people who use neither CFOP nor Roux (ZZ, Petrus, etc.?) would also be very welcome, but this is probably harder to get.

(No offence, but (i) I find this somewhat suspect and (ii) a p-value of 0.3 is very weak evidence.)

Anthony Brooks has a

Reading this thread was quite funny. It started off with someone saying that the fastest rouxer at home isn't the fastest officially, then it turned into a Roux v CFOP debate with the CFOPers/anti-Rouxers giving no good argument, and the Rouxers not really giving any proper argument, just pointing out that the CFOP argument was flawed. I just think that now's the time for everyone to realise that ZZ-A is the future. But seriously, CFOP and Roux are both really good methods, as is ZZ, it's just that people haven't listened to the arguments for and against each method. Trust me, I've had a lot of discussions on ZZvsCFOPvsRoux(vsPetrus a little bit) and the conclusion that I've come to is ZZ≈Roux>CFOP but other people have come to many different conclusions. The hard part of the whole debate is that it's impossible to prove with evidence of 'this person is the fastest, therfore their method is the best.' The easiest counter example is corners first in the 80s. Instead, the best way, imo, is to look at the concrete evidence in movecount and to look at the not-too-super-hypothetical ergonomics. For example, you could point out to different U perms. At the end of it all, you weigh out the evidence and choose the method that you think is the best. I have come to the conclusion of ZZ or Roux, and as I already mained ZZ, I chose ZZ. Some people would rather Roux, some would rather another, less well known method such as LEOR. It all depends on the person. In conclusion, use good evidence.

I really liked reading that post, but I did not mean that the fastest router at home is not the fastest officially necessarily. I meant that others are catching up to Kian and this is a good thing. Sorry for any confusion.

 

[Sue Doenim]

Anthony Brooks has a

Does he really?
[It seems to me that you forgot to finish your thought.]

 

[Mike Hughey]

Anthony Brooks has a

Does he really?
[It seems to me that you forgot to finish your thought.]

I accidentally a 5x5

 

[ImmolatedMarmoset]

I accidentally a 5x5

I accidentally the whole solve over
Beat that.

Does he really?
[It seems to me that you forgot to finish your thought.]

Oops! Anthony Brooks has a 7.60 average that was top 100 at the time he got it with full ZB, so that’s cool

 

[waltermcy0110]

Sean got a 5.98 avg today. Still waiting for reconstructions. I hope his improvements are from more efficient solutions so that I will not have an excuse for insisting that Kian is the best Rouxer.

 

[PapaSmurf]

Efficient!=good at Roux. Although Kian's the better OHer, and it's definitely a place where Sean can improve.