Welcome Subscribers. I have moved to sixshootin.com

John -- "Conquistador"

Things are different today. In most ways, the information age has made it more difficult to bet sports. The lines are tighter because we have much more information at our fingertips. We have to look for ways to improve. We can use this information age to our advantage. An extremely difficult thing is to sift through the data, and use what is meaningful. Many people use back fitting (find a trend after the fact and think that it will hold up going forward). It may work going forward, but a theory has a stronger foundation if it is put to a hypothesis first. A good prognosticator learns from others. Even when somebody says something is an original idea, it is likely a modification of a similar idea. Isaac Newton once said, "I have seen further than others because I have stood on the shoulders of giants.

I use a normal distribution model to create probabilities. Many people hear statements about the normal distribution, but don’t know exactly what it is. In a nutshell, many distributions (including football scores) have a ‘bell shaped’ curve. We find the mean and standard deviation and use a chart to find the corresponding probability. It is basic math. What is difficult is to find MEANINGFUL data to arrive at a projected point spread. The key is to find statistics that CORRELATE to scoring. Yards per play is a very good statistic, but taken out of context can be misleading. For example, an 8 yard play on third and nine is different than an eight yard gain on third and seven. (thanks to the Hidden Game of Football for largely pioneering context with regard to yardage)

I have been in the Carib contest for five years, and although I have not placed in the top 20, twice I have placed in the top 50 (once at 25 and last year at 39).

My five year record has been:
50-35
46-36
41-42
45-35
92-75

for a record of 55%. This has been picking five games a week (ten last year, the first year they included select over/unders). I will likely have an average of 3 to five 5 in the pros and from 5 to 10 in college. The number of games during the first few weeks of pro games may be sparse, but that will likely not be the case with college football since there is less parity (and therefore more predictability from year to year with most teams). I am hoping for somewhere between 55 and 60%.

During the last five years, picks projected at 55-57.5% have been 55.3%. And ones projected at 57.5% and above has been 57.2% (note: projections may slightly exceed actual results because models are predicated on past data, and even though a trend may hold up over time, it is more likely than not to be a little idealistic).

I will attach a probability to each of my picks. These are determined by respective ratings in a normal distribution model. You can decide what you would like to play, but 55% and above is a good cutoff. Keep in mind, there are no guarantees. A person who can pick 55% has about a 10% probability to hit 46% over a span of 50 games. You must be disciplined, and grind it out. I suggest that you should never bet more than 2% of your bankroll. Remember what the Great Gazoo once said. "If you can't afford to lose, then you can't afford to win"

Here is a sample format, which I will use for pro football. In college, I will give a probability for every game the model shows a favorite at 55% or higher. I have just one college game for this week. The first few weeks are a little more sparse.