NCAAF WEEK 7 Tico Joe's Statistical ATS Week #2

but are they really 75% or do people just believe they are?  and even if they were, the problem is finding them.  if a bettor really KNEW certain games were best bets, why would he make any other wagers?

yet people constantly do this--even 'jdn money' who wins big on his 'monster plays' yet still wagers on other games.  the only exception i know of was a poster named 'mr. perfect' who only showed up a few times a year with just one pick and was almost never wrong

but that's different from what you said--you said there are 5-10 games each week that have a 75% likelihood of covering

there are no mathematics to prove this

In a historical data sense, it is possible to prove that a matchup will have a given % chance of covering or not covering.

The question of how is it possible to assign a % success rate to a given matchup could be  answered as follows:

Each game picked would need to fit a given scenario, or have several basic tenants in common, or satisfy a given number of parameters as all other matchups that are also used in the comparison.  That is one way to get into a situation where you are truly comparing oranges to oranges, at least in the statistical sense.

So, it is possible to look for common denominators using  different criteria as a way to sort out each week's matchups. Point spread situation filters, game stat parameters and characteristics of the teams involved in each individual matchup could be used  to "sort" through the matchups and put them in different piles......and then run the success rates of each defined type of matchup vs the point spread over a given number of historical occurences (past results of games vs the points), in order to establish a % chance of success in a future event.  This is basic data backtesting, and its validity in accurately predicting future events vary depending on whom you talk to.

Personally, I am not real big on using historical data sets as the end all be all, but they can certainly be a great place to start.

The major pitfall in using historical data sets in an attempt to develop a system or betting approach, is that it is very easy to find yourself back fitting data in order to fit your system.  If a high winning percentage is desired then the tendency is to add more filters in order to weed out the losses, even though you may be losing a disproportionate amount of winners to losers with each filter, and killing profitability.

In any case, in the historical sense it is possible to mathematically assign projected winning percentages to a future event.  The question of how valid they are is another thing altogether.

BUT if Ticojoe has found something then more power to him....it certainly is possible to develop a successful approach.....and if he can run it up at an 8-4 clip every week, then go get 'em.....my curiousity is certainly peaked, and I hope you continue to post your system picks. 

GL with your picks  

You might have a point regarding simultaneous wagers, but you need to keep your bankroll in play if you want to make money, so you need to have something going in all the time slots (assuming you can get something +ev).