The beauty of drivel such as that of Nostradamus the Prognosticator is that is is expressed so vaguely and metaphorically that anyone with inclination and half a wit can scry from it whatever they wish, to almost any time and place. His words are clouds and mist, and so for that reason he is nigh well unfalsifiable. Hence, for a certain crew, nigh well infallible. His track record is free of all proven error, the enviable lad.
But I am an Anti-Nostradamus, which means I prognosticate as clearly and sharply as I can. And thus can be judged easily, once the evidence is in, by that proverbial half wit with the inclination to do so. Which means, sadly enough, that my track record is resplendent in proven error.
It is time to extend this track record.
Today, Germany plays Costa Rica, and Poland plays Ecuador. I am on edge about the Ger-CR game, and the absence of the great Ballack, but will go with a 2-1 German victory. Poland has hurt itself by not going with Dudek, who is the better keeper in dangerous moments, but then Ecuador's keeper and defense are truly mediocre. Poland 2-0.
Let the embarassment begin!
Subscribe to:
Post Comments (Atom)
Posts
6 comments:
Going out on a limb, AA? OK, we'll take note, and compare afterwards. Just so we can do this rigorously, how do you measure deviation from prediction to actual value? Is there a good way to assign a numerical value, say between -1 and 1? Eg, if Germany beats Costa Rica by 2-1, you obviously get a score of 1. If Costa Rica beats Germany by 10-0, well I guess you get a -.99 or something. What if it's 0-0? or 5-0? We need some math here, if we are to truly move away from the Nostradamus flummoxery!
Germany-Costa Rica 4-2 vs predicted 2-1: quite good, I'd say a score of 94%.
Poland-Ecuador 0-2 vs predicted 2-0: very bad, I'd say a score of -98%.
Overall, thus, a score of -4% for our budding Anti-Nostradamus. But my mathematical model is still under-developed, with some fiddling one could surely transmute my -4% into a +2% or something...
yeah, have been trying to fiddle with points to get above 0, and unless I employ "creative" accounting it cannot be honestly done. How about this? For G-CR had the winner [70%] and ratio of goals correct, but lose 10% for error in magnitude. So 90% (+). For Pol-Ecu had correct score [so correct magnitude], so 30% (+), but wrong winner and so -70%. Giving me a grand total of 50% [+]. ...Hey, I'm good. so, think the pollster folks could appreciate this skill?
We'll count magnitudes by units of 1? So if Germany had beaten CR by 6-3 I'd lose 20%?
agh, This needs refinement.
Correct magnitude (+30%) but wrong winner (-70%) for a score of only -40%? Hmmm... that sounds rather fishy. Try selling this to Gallup when they'r polling for elections or something...
Post a Comment