The John De Lange Sophisticated Prediction Contest - Standings after Round of 16
Futbolmetrix
assuming that everything from here onwards goes according to rank:
Brazil WF
Spain LF
Argentina, Uruguay SF
Ghana, Germany, Netherlands, Paraguay QF
Rank Name Total Score
1 AQB -32.9604
2 Voros McCracken -33.653
3 Collective_Simulations -33.8126
4 Castrol Football -34.5961
5 Chance de Gol -35.0522
6 ELO -35.234
7 Werner Pichler -35.6579
8 Futbolmetrix -36.0331
9 Lléo -36.0754
10 Collective_Humans -36.4458
11 CHRIS HORYMSKI -37.2276
12 MH -38.0779
13 Jesus Petry -38.2283
14 Miguel -38.8641
15 Joachim Parsch -39.1777
16 milivella -39.4617
17 Philipp Lucas -41.2321
18 Alessandro Riolo -42.4216
19 Agnostic -42.9751
20 juanvazquez -47.4937
21 Olcay Kara -69.5579
22 JK -106.167
23 Luiz Mello -106.802
24 Myk -108.037
25 Anders T -112.684
26 7h@ch -174.041
27 Jim Goloboy -179.221
28 Jussi Uosukainen -249.926
29 Nate Silver SPi -975.708
30 Binder Dundat -1187.68
31 Mark V. -1187.8
32 El Kot -1202.47
It looks as if AQB has cemented his position at the top of the standings and
is unlikely to lose it. Among the humans, though, everything is still wide
wide open.
D
Binder Dundat
> Not exactly the standings. Rather, the standings for the final competition
> assuming that everything from here onwards goes according to rank:
>
> Brazil WF
> Spain LF
> Argentina, Uruguay SF
> Ghana, Germany, Netherlands, Paraguay QF
>
> Rank Name Total Score
> 1 AQB -32.9604
> 2 Voros McCracken -33.653
> 3 Collective_Simulations -33.8126
> 4 Castrol Football -34.5961
> 5 Chance de Gol -35.0522
> 6 ELO -35.234
> 7 Werner Pichler -35.6579
> 8 Futbolmetrix -36.0331
I really do not remember doing the sophisticated predictions for the
World cup? It was way too sophisticated for me, I must of done it
after a night at the local bar?
Futbolmetrix
news:4f0769f5-31f6-4dad...@n8g2000prh.googlegroups.com...
> I really do not remember doing the sophisticated predictions for the
> World cup? It was way too sophisticated for me, I must of done it
> after a night at the local bar?
There was an unofficial thread in which a bunch of people posted
sophisticated predictions for the first round. I used those. The reason that
you are close to the bottom of the standings is that you did not enter a
full matrix of predictions. In fact, it should really count as a "no entry".
I'm going to erase those next time I post the standings.
D
MH
Why do I get error thingies (# NUM) in the total column for all of those
teams for which I predicted a zero somewhere? How can I adjust those ?
>
> D
>
>
Jesus Petry
The solution I use is to replace those zeros with a very small value
(1E-99, for instance).
You might have a problem with 1s too, in partial scores. The
workaround is more complicated then.
Tchau!
Jesus Petry
Futbolmetrix
"MH" <MHno...@ucalgary.ca> wrote in message
news:i0ftln$2qh$2...@news.ucalgary.ca...
If your columns named d_r1-d_wf are in columns M-R, you can use the formula:
M3*LN(MAX(C3,1E-32))+N3*LN(MAX(D3,1E-32))
+O3*LN(MAX(E3,1E-32))+P3*LN(MAX(F3,1E-32))+Q3*LN(MAX(G3,1E-32))+R3*LN(MAX(H3,1E-32))
D
milivella
Now you _have to_ tell what the complicated workaround is, since I
feel silly having thought "well, just put .9999 instead"! :)
--
Cheers
milivella
Binder Dundat
OK good, I was beginning to wonder if I somehow managed that
complicated, err I mean sophisticated matrix in a drunken stupor one
night?
Jesus Petry
Actually, IIRC, it's only a problem for partial scores, and the
workaround was changing the formulae in the cells to avoid that the
sum of the fields inside a ln calculation was zero. So instead of (1 -
1) it was (1E-99 + 1E-99 + 1E-99 + ...).
Tchau!
Jesus Petry
milivella
Ah, now I got (well, almost...) it. Thanks a lot! :)
--
Cheers
milivella
Voros
2nd is so bad. I'd like to become a multiple winner if that's
possible, but I'd certainly live with second. Certainly highlights
that you can indeed predict football matches with at least decent
accuracy using a statistical system.
--
Voros