Page 1 of 1

Data mining the VF scores

Posted: Wed Dec 05, 2012 9:52 am
by Zippy8
Yeah, yeah, yeah. I know :roll:

So... taking the numbers available on the most excellent Virtual Freestyle page I have dredged up some "results".

  • I only used the normalised scores, not absolute and not placing. Because.
  • Two tables:- one is total, so biased towards participation. Another, decayed against time. 100% for the last round, 80%, 60%, 40%, 20% then 10% for all other rounds.
  • I am considering a "top X" table with just someone's best few numbers.

And...
for the decayed scores:-
  • Bryan Beasley
  • Adrian Atherton
  • Alex Belov
  • Tom Patterson
  • Krijn Hemminga

For all available results:-
  • Bryan Beasley
  • Tom Patterson
  • Alex Belov
  • Dave Hearn
  • Gerardo Fisanotti

I find this interesting but then again I am distinctly odd that way. However, I don't think it proves much :wink: Presenting all the data for 96 different people in a comprehensible way seems unlikely. I may try to dig out the missing 10 rounds as the evenings draw in.

Mike.

Re: Data mining the VF scores

Posted: Wed Dec 05, 2012 1:14 pm
by Zippy8
And if anyone expected me to have an idea about playing with the data and not immediately do it.... :doubt:

All of the available scoring data is in (missing 14,9,7~1) and taking the top 5 rounds for each person:-

  • Alex Belov
  • Bryan Beasley
  • David RĂ¼ger
  • Davide Equizzi
  • Will Sturdy

(I am going to need those missing numbers to get 6-time winner Piero Serra to the top of this list)

Looking at the squiggly lines that the scores produces I can see some things of note. But mostly it's just squiggly lines.

Mike.

Re: Data mining the VF scores

Posted: Fri Dec 07, 2012 2:25 pm
by ObijuanKenobe
I am not sure how you can extract a rank based on entirely subjective scoring values from a never-the-same set of judges. I would say the only thing that can be averaged across rounds is rank.

But for sure, if Piero is not #1, the system is broken.

obi

Re: Data mining the VF scores

Posted: Fri Dec 07, 2012 11:21 pm
by Coogee
To Quote Disreali " There are 3 types of lies, white lies, damn lies and statistics"
Pretty much sums up how statistics works to me.
Mike