Game Predictions – 3/18/2006
We’ve been having a pretty good week for game predictions and it continued last night as my predictions were a perfect 4 for 4. Let’s hope it continues tonight. Before we get to the predictions I have 2 announcements, one related to a small bug I found in my program which incorrectly assigned strong/good/some confidence ratings to some games and the second a new feature added to the prediction table to aid in betting.
1. I found a small bug in my program which incorrectly assigned confidence ratings to some games though the predicted winner was correct. Games in which one team played the previous night and the other did not were far too often asigned the ‘Some’ rating when a significant number of them should have been given good or long. I have fixed this and I also tweaked hor the ratings get assigned to all games by increasing the value required to make a strong confidence rating. With these fixes/changes we will have many more Strong and Good predictions. If I retroactively apply these changes to my past predictions since mid-January my success rates would be:
Strong: 33 of 50 – 66.0%
Good: 45 of 70 – 64.3%
Some: 72 of 121 – 59.5%
Overall: 150 of 241 – 62.2%
2. The other major change I made was to add some betting advice to the game predictions. You will find a new column HTOdds with a bunch of numbers listed below them. The HTOdds can be used as an indicator of what odds you should look for when betting on the home team. My reccomendation is that you only bet on the home team if the odds are better than the numbers listed below, otherwise bet on the road team for the best risk/reward ratio. If the betting odds are pretty close to the HTOdds you might want to consider passing on the game altogether as you aren’t getting a great advantage either way. As an example, let’s look at Ottawa. Ottawa has HTOdds listed as -127 and Bowmans has the odds for Ottawa listed at -210. According to my calculations these odds have Ottawa as far too significant a favourite so I would take Buffalo and their +175 odds. Another example: The Boston/Carolina game has HTOdds for Boston listed as +106 while Bowmans has them listed as +160. In this case, take Boston. My calculations show only a minor risk but the Bowmans odds give a pretty godo reward. If you are unsure what these numbers mean, a positive odds, +150 for example, means you will win $150 profit on a $100 bet while negative odds, -150 means you have to bet $150 to win $100 profit. In Buffalo’s case a $100 bet will net you a win of $175 (above and beyond your initial lay out of $100).
Caution: I have not tested this betting strategy so bet at your own risk and use only as a guideline. Also, sine there isn’t a direct correlation between how I make calculate my picks and how I calculate my odds there will be the odd inconsistency between them where the prediction listed is for one team but the odds indicate the other team is the favourite. This should be pretty rare but take it as an indication that the game is about as close to a pick-em game as there could be.
Ok, so now on to the predictions.
|Home Team||Road Team||Predicted Winner||Confidence||HTOdds|
|San Jose||Dallas||San Jose||Some||-112|
|NY Rangers||Toronto||NY Rangers||Some||-115|
|Los Angeles||St. Louis||Los Angeles||Good||-187|