Modeling is a concept on the periphery for most sports gamblers. Plenty of people give out sports picks, but very few do so with the justification the picks are what their model kicked out.
For those gamblers who build a model and see success with it, sharing is not caring. It becomes a kind of trade secret, a secret algorithm of weighted values assigned to certain variables that can provide winning selections. A successful modeler may tell you he has a fast car, and even give you a taste of how fast the car goes by giving you a selection or two, but he probably won’t pop the hood and let you see the guts. The reason for this is simple. If you develop a system you believe is capable of beating the book and is “ahead of the curve,” why make it public? Sportsbooks prove year after year that the public loses.
If you want to try modeling, you’ll want to use data-driven resource sites that can provide you important statistical and historical information. From there collect any and all information you believe is noteworthy and assign value to it in order to compile a composite data point you can use to rank teams against one another. Once you’ve calculated a statistically based prediction, ignore any bias you might have and base wagers strictly off the numbers.
Sound easy enough? Not exactly. Where models go wrong is when they aren’t able to successfully determine stats that are most important and those that are relatively insignificant. A good modeler will constantly review their results to better sharpen the spear, but this takes mathematical interest and dedication. There is a substantial population that is involved in the market and enjoys this pursuit of statistical significance. They are the modelers, and they’re changing the game.