The Limper went 3-11 ATS!!! Oh my!!! The model might have had a better week chewing razor blades (hence the appended warning). How could this happen? For the benefit of some, a few words of explanation:
If we can assume that more-or-less 2, plus more-or-less 2, equals more-or-less 4, this can be useful, but only if our assumptions are safely assumptive; and our assumptions are “safe” only if they are “historically true”. Not, by any means, guaranteed, merely safely assumed. The algorithm which drives the model relies solely on the historical truth of the variables it uses in its calculations. (True, the weights given these variables are mostly subjective, but even these are objectively driven (ie. not based on mere “eye test”); so, I can honestly say the model’s projections are as objective as they can be.)
However, when the model fails – when the sum of 2+2=5, or 3 – when what is historically true fails to project a current outcome – the problem is not that mathematical abstractions cannot project current or future realities, it’s simply that those abstractions are somehow insufficient (ie. What was “historically true” was, in fact, 3+2, or 2+1; that the data mined and used was either wrong or misapplied). Of course, there are limits to the extent which accurate and relevant data-mining is possible, which would sound like an excuse, but my model is – on the season – beating out Number Fire and Team Rankings, and their extensive resources, on their own ATS projections; so The Limper (with a staff of one doddering, old fart) is doing as well as should be expected.
Anyway, the point is, I’m not “quitting”, “going away”, or finding “another hobby”. Data projections are not crystal ball predictions, so I’m constantly reviewing data as it is gleaned and revising how the algorithm uses it - and enjoying every minute of this strange NFL season.
GLTA
If we can assume that more-or-less 2, plus more-or-less 2, equals more-or-less 4, this can be useful, but only if our assumptions are safely assumptive; and our assumptions are “safe” only if they are “historically true”. Not, by any means, guaranteed, merely safely assumed. The algorithm which drives the model relies solely on the historical truth of the variables it uses in its calculations. (True, the weights given these variables are mostly subjective, but even these are objectively driven (ie. not based on mere “eye test”); so, I can honestly say the model’s projections are as objective as they can be.)
However, when the model fails – when the sum of 2+2=5, or 3 – when what is historically true fails to project a current outcome – the problem is not that mathematical abstractions cannot project current or future realities, it’s simply that those abstractions are somehow insufficient (ie. What was “historically true” was, in fact, 3+2, or 2+1; that the data mined and used was either wrong or misapplied). Of course, there are limits to the extent which accurate and relevant data-mining is possible, which would sound like an excuse, but my model is – on the season – beating out Number Fire and Team Rankings, and their extensive resources, on their own ATS projections; so The Limper (with a staff of one doddering, old fart) is doing as well as should be expected.
Anyway, the point is, I’m not “quitting”, “going away”, or finding “another hobby”. Data projections are not crystal ball predictions, so I’m constantly reviewing data as it is gleaned and revising how the algorithm uses it - and enjoying every minute of this strange NFL season.
GLTA
