10 Billion Threshold Prediction Challenge

[frozenchosen]

Okay, all you statmongers. Let's see how good you are at using your stats. I challenge you to predict when (day, hour, and minute), we will cross the 10 billion threshold on the current protein. The closest prediction will receive kudos and our undivided respect.:notworthy

How about it? Who's up for the challenge?

Post your best prediction right here. Of course, those who delay posting their prediction until we get closer are just demonstrating their lack of faith in their regression equations.

[Scoofy12]

hm... kinda like the anchor pool we had in the Free-DC side of the house for the last protein :)

[frozenchosen]

Oh.... I think we can do better than just an anchor pool. We've got all these stats and all these computers, let's see who can come up with a prediction based on a regression model. As I recall, the ancor pool let's people just pick time slots. I'm more curious as to how one might plot regressions to predict the crossing point.

[FoBoT]

25 Jan 03 01:20 EST

[TheOtherZaphod]

Just like our project here, I reserve the right to "refine" my estimate. I expect to take several snapshot looks at it, and then formulate a final guess from there.

The first swag point is being taken at about 11:30 EST, on Jan. 22nd. The calculation, at that time, is that we will pass the mark at 18:30 EST on Friday Jan. 24th.

I will shoot another reading later today...

Ok, My second revision calls for crossing the line at about 1:30 PM on Friday.

Quite a revision downwards, but I realized that the production totals I used in yesterday's computations were from statman, and I believe they only included the top 1000 contributors who are on a team. Today I was able to use the acutal delta in the project totals from yesterday to today. I got a much higher production rate that way.

[AMD_is_logical]

Fri, Jan24, 1:10 PM EST

[frozenchosen]

I think some folks are going to be dumping a bunch of crunches at the end. I'm taking the 1:00 pm slot. Any other takers?:eek:

[Kileran]

heh, i posted a guess, then noticed that we're not using Central time here :)

I'l take Jan 24th, 11:35 am

Sean

[Kileran]

heh, i was way off, i'l take a stab at 3:45 now :)

[AMD_is_logical]

Originally posted by AMD_is_logical
Fri, Jan24, 1:10 PM EST
Fri Jan 24 13:00:47 2003 EST - 9,997,282,228

Fri Jan 24 13:10:43 2003 EST - 10,001,677,730

:D

[frozenchosen]

It appears that the 1:10 slot was the best guess. AMD_is_logical is the winner with the closest prediction that didn't undershoot the exact threshold. Congratulations! I'm curious about how you did the prediction.



:notworthy

[AMD_is_logical]

Originally posted by frozenchosen
I'm curious about how you did the prediction.I looked at Dyyryath's stat page:
http://stats.zerothelement.com/cgi-bin/distributed-folding/render-teams.pl
which has the "Last Day Structs" for the entire project. I then assumed that this was the best guess for the future rate. I then took the latest stats from the distributed folding stat page, and figured out how long it would take to reach 10 Billion given that rate. :cool:

[bwkaz]

Heh heh, go Dyyryath! :D

Seriously, though, good work! I'll freely admit, I was too cowardly (... or at least too busy) to go try it. :p