Predicting the Winning 3x3x3 Average for World Championships 2009

[KConny]

11.11

Are you able to read this? I only wrote it because my message was too short. I hope no one quotes me. If so everyone in the world would see this secret message.

 

[philkt731]

I see your secret message

 

[shelley]

I don't see why sub10 is so out-of-range. One solve sub9 (which is happening now), 2 solves sub10, 1 bad solve (12+), and one solve sub10.5ish, and you could have a sub10 average.

We did all, however, assume that a WR average would be set, which is sub-10 at the moment, and according to my calculations, a WR average of sub-10 seconds is probably going to happen at the end of next year, which could be at the WC09, but, keep in mind that last year, WR average was not set at the WC

Nobody assumed a WR average would be set at WC 09. I only mentioned the WR average because of all the people predicting that the winning average at WC 09 would be at 8 or 9 seconds. If that's the winning average, the WR average at that time would obviously have to be at least as fast.

If we look at the progress of the WR average so far (for which we have more data points to work with), most of us would probably conclude that an 8 second WR average in a year is highly unlikely. I would of course love to be proven wrong, but I'm just trying to take a realistic view on things.

 

[pjk]

You guys can't look linearly, since cubing has progress greatly over the last several years. 20 second avg to sub-15 avg to sub-12 avg. It will be leveling out soon.

 

[blah]

Too difficult to say exact number. Best prediction is sub-12. I think that is probably 99.999999999% likely. Perhaps it will be somewhere around 10.5? sub-10 is possible, but not too likely.

Ditto to everything you said. The 0.000000001% is when everyone's having a very very bad day, like this: http://www.worldcubeassociation.org/results/c.php?top3=Top+3&competitionId=ItalianOpen2008.

...so assuming a constant rate of change...

Err... it's called the Law of Diminishing Marginal Returns, at least I think it is. Stupid Economics jargon. I hate Econs.

I don't see why sub10 is so out-of-range.

Everyone pray that Jason Baum finishes ZBLL in time! (or are there any others who (intend to) practice full ZB out there?)

 

[EmersonHerrmann]

Maybe I'm being unrealistic, but there are so many variables to determine (e.g. new methods), why not just guess? 8.78 is what I say. This is a really fun thread

 

[cubejunkies]

You guys can't look linearly, since cubing has progress greatly over the last several years. 20 second avg to sub-15 avg to sub-12 avg. It will be leveling out soon.

I don't know, if you do an exponential regression of years vs. averages and then look at the graph, it levels out in times that are really incomprehensible, and in a really long time, according to my equation made by 8 made wr averages, (20, 16.53, 15.38, 14.52, 13.22, 11.76, 11.48, and 11.28) giving us an equation of 20.095933921477*.92299933741117^x

 

[hawkmp4]

That function has an asymptote of y=0. Not very realistic for this application.

 

[cubejunkies]

...so assuming a constant rate of change...

Err... it's called the Law of Diminishing Marginal Returns, at least I think it is. Stupid Economics jargon. I hate Econs.

Yeah, I already corrected myself in after that posting the exponential regression I got

That function has an asymptote of y=0. Not very realistic for this application.

okay, change it, I'm a bit rusty on my algebra, but i believe that it still levels off in a number of years

 

[Hadley4000]

Let's just look at Erik, Edouard and Harris. The only three who have gotten the equvilant of a sub-10. Let's take their 5 fastest times and make them into an average.

[7.08]
9.77
[9.96]
8.90
9.93

9.53 average.



Edouard.
9.18
9.66
[8.66]
9.90
[10.36]

9.58 average



Harris.
[9.44]
9.80
9.96
10.03
[10.15]


9.93 average.



So there are 3 cubers out there who have hit enough times to make them reach a sub-10 average. I think that this means that sub-10 average in competition can happen.

 

[Sa967St]

10.94, just a guess

 

[blah]

@Hadley, doesn't your data mean it can't happen? Even with their 5 fastest solves, of which most are undeniably lucky, they're barely sub-10 (actually 9.5 is quite a convincing sub-10 to me, but whatever), and you expect them to get such results under pressure in the WC? By the way, did you forget an extremely fast Japanese cuber who has very good shampoo/hair conditioner?

 

[cubejunkies]

@Hadley, doesn't your data mean it can't happen? Even with their 5 fastest solves, of which most are undeniably lucky, they're barely sub-10 (actually 9.5 is quite a convincing sub-10 to me, but whatever), and you expect them to get such results under pressure in the WC?

actually, some aren't lucky. like harris' 9.44 and edouard's 9.66, both were full step.

anyway, I was reading the beginning of this thread, and the original poster said that they couldn't calculate the winning average because they couldn't do exponential regression, and then vault312(i think) went on into doing a logarithmic equation, more accurate, but I don't think we've even gotten an exponential regression equation from the original data set of the WC winning averages. So, i took the liberty of doing it and, if you're curious, the equation is:

24.96626725*.7893034904^x​

Just thought that might help a bit...

 

[blah]

I might try that instead of sleeping tonight. I can probably make it out of K'nex pieces.

Good luck. Send me a prototype when you're done, I'll PM you my mailing address as soon as you PM me a picture of it

My inspiration for that idea:

Quoted from a Wikipedia article on Claude Shannon:

One of his more humorous devices was a box kept on his desk called the "Ultimate Machine", based on an idea by Marvin Minsky. Otherwise featureless, the box possessed a single switch on its side. When the switch was flipped, the lid of the box opened and a mechanical hand reached out, flipped off the switch, then retracted back inside the box. In addition he built a device that could solve the Rubik's cube puzzle.

The last sentence is just a coincidence This guy is the coolest Mathematician ever.

 

[brunson]

We should make a pool. Everyone who wants to participate picks a time, accurate to 2 decimals and puts in a dollar. Closest to the actual result without going under takes the pot.

 

[blah]

We should make a pool. Everyone who wants to participate picks a time, accurate to 2 decimals and puts in a dollar. Closest to the actual result without going under takes the pot.

I think the forum management team has had quite an unenjoyable experience with the V7.

 

[pcharles93]

Or PJK could make a contest out of this. He sets up a range of times from 8-11.5 by .25 increments and to participate, you have to submit at least 2 dollars. Whoever has the most money on the closest time interval wins half the total money to spend. The other half goes to me, the original creator of this contest

 

[Harris Chan]

of which most are undeniably lucky

actually, some aren't lucky. like harris' 9.44 and edouard's 9.66, both were full step.

I think I need to create a machine to wear on my head that sticks a hand out to smack my forehead everytime cubejunkies posts.

By the way, 9.44 = 2-move cross, that's not technically lucky - it's literally lucky.

Well I do have 9.80 and 9.96 non lucky official solves too, without the 2 move crosses

 

[cubejunkies]

We should make a pool. Everyone who wants to participate picks a time, accurate to 2 decimals and puts in a dollar. Closest to the actual result without going under takes the pot.

I think the forum management team has had quite an unenjoyable experience with the V7.

I agree. But if anything, the betting should be done outside of here. Here's my idea of how it could happen:
Money is collected by people who know they can go to the WC, like WCA delegates, so that whenever you go to a comp, you can sign up, giving your name, info, money, and guess of the average, to the delegate or whatever, and at the WC, everyone gets together and sees who guessed correctly, splitting the money between same guess winners, etc.

 

[blah]

Well I do have 9.80 and 9.96 non lucky official solves too, without the 2 move crosses

I had a feeling you'd be offended, but I didn't know how to phrase that better No offense, really, just pointing out a fact.