This is Alpha, the first-born, when he was 2YO.
This is Beta, the second-born, when he was about 2YO.
This is Gamma, the third-born, when he was about 18MO.
Football: MPWHFA
Mar
6
2014
The last time I added a method of predicting the winners of NFL games, it wasn’t all that impressive. That was YPP, and it was not an improvement over MPW.
MPW has been the leading indicator for the last few years, better than any other automated, statistics-based predictors.
But now MPW is pushed to second place, because MPWHFA has taken the lead.
I added MPWHFA (and its results) to the list of NFL game-predicting methods.
It wasn’t just slightly better than the old leader – it averaged a full percentage point better.
Of course, that’s not entirely fair to MPW. HFA stands for “Home-Field Advantage”, and MPWHFA is really just an improved version of MPW. Take MPW, add a few points to the home team, and see which team has the larger number. That team should be the winner.
Next step: predicting not just the winner, but also the point spread. MPWHFA does that too (and at around a 60% success rate, I might add). MPW and MPWHFA are based on net points. That should be all you need to know to figure out point-spread predictions.
Enjoy!
As for every matter of wisdom and understanding about which the king consulted them, he found them ten times better than all the magicians and conjurers who were in all his realm.
Daniel 1:20
This little article thingy was written by Some Guy sometime around 6:25 am and has been carefully placed in the Sports category.
March 6th, 2014 at 10:00 am
When you make a new model, what are you doing this in (Excel, Perl, etc…)?
I’m wanting to experiment with my own models just to see what happens.
March 6th, 2014 at 12:40 pm
PHP.
That way, the website is updated as I work. There are no additional steps to convert from Excel to HTML.
Also, if you haven’t visited http://www.somefunsite.com/footballformula/mpw.php to see my research into stats vs. wins, you might want to do so.
March 7th, 2014 at 9:12 am
I was hoping you wouldn’t say some weird thing like Smalltalk :)
I’d be interested in trying to develop my own algorithms to see how well I’d do. It could be an awesome guest post!
Genetic algorithms to calculate “win momentum” weighted against “loss frustration”……it’d be awesome.
March 8th, 2014 at 12:40 am
I will not doubt you this fall.