The Predictor was designed to use the scores of games to determine the outcome of future games. This is purely a computer based model with no human factors included.
- Works well with a lot of data: This is great for something like the NBA, NHL, and college basketball. There are so many games played that eventually the data can see a lot of important trends that may give a slight advantage over the human eye. Late in the season, the Predictor ends up converging very closely with the Vegas Line.
- Zero Bias: There is absolutely no human massaging that goes into these predictions. Human Bias has a very interesting way of creeping its way into everyone’s thoughts and feelings about particular teams and games. The media, in particular, will drive up hype on players and teams to build ratings and this will sway public opinion. The computers, fortunately, don’t listen to Lee Corso.
- Early season predictions are based on previous years’ results: As one can imagine, this makes early season predictions quite unreliable. A great quarterback graduating, NCAA sanctions, a coaching change, etc. can have a huge impact on year-over-year performance.
- What happens with injuries: The quick answer is “nothing”. There is no way that the Predictor can account for injuries in a game. If a starting quarterback goes down with a season ending injury than that can change the line of a game by a touchdown or more. Again, since we can just see the scores of previous games, we cannot account for this.
- Predictions are not strong across leagues: Computer predictions work very well in leagues where there is a lot of games played and with a lot of inter-connectivity between teams. Where the methods sometime fail is when there isn’t that inter-connectivity (i.e. AL vs. NL in baseball, FBS vs. FCS in college football, an EPL team vs. a La Liga team). This will occasionally produce skewed results and should be viewed with caution.
In short, do not email me complaining that the predictions are horrible and that you lost all your money betting my picks. These predictions are a great base line to use when looking at games to bet, but should only be used as one piece of the puzzle. As stated above, there are many other factors that go into final outcome of a game, so research the games carefully.