Updating Power Ratings
This article is a follow up to the previous article How to Create Your Own Power Ratings and will look at the updating process.To summarize, in the first article we showed you how to create power ratings by using several different sources and averaging them into one power rating number. Jeff Sagarin's ratings in USA Today are good to use, as the ones in the Gold Sheet and many of the other newsletters. There are also several other sources online to use.
Because it is less time consuming to update your own numbers rather than going through the process of averaging the others all over again, that's the best way to go. There is one additional reason for updating your own ratings, which we will discuss briefly and go into more detail.
The Updating Process
The adjustment to your power ratings is a bit of a subjective process, but it is something that must be done. If your numbers say one team will win by 10 points and they win by 35, you have underestimated one team and over-estimated the other. This is why there needs to be adjustments to your numbers.For the first three weeks of a football or basketball season, the maximum adjustment I will use is three points. In the example above of a team winning by 35, I would adjust the winner's number upwards by three points and reduce the loser's power rating by three points.
After three weeks are in the books, the maximum adjustment I will use falls to two points. This is so that one very good or very bad performance doesn't skew the power ratings, as all teams have good or bad games.
For the first three weeks, if my predicted margin, based on power ratings, is within seven points of the final score I will not make any adjustment. If the final score is between 8 and 14 points from the predicted difference, I will use a one point adjustment. Remember, that is adjusting the winner's power rating up one point and decreasing the loser's rating down one point.
If the final score is between 15 and 21 points from the predicted difference, I'll use a two point adjustment. Anything above 22 points is a three-point adjustment.
After three weeks are in the books, simply drop the criteria for a three-point adjustment and use anything above a 15-point difference is the maximum two-point adjustment.
There are always teams that play better or worse than expected, which is why we use a three-point adjustment early in the season. After three weeks our power ratings should reflect the team's performance level, so we'll drop to two points at that time.
Injuries
The other time to make adjustments to your power ratings is when injuries occur. Obviously, a team without its starting quarterback or point guard is not going to perform at the same level.This is especially true in college sports, where most teams don't have the luxury of having a competent replacement.
If a key college player at a crucial position is out, I'll make up to a four-point adjustment of a team's power rating. This is where some subjectivity comes into play.
If a college team's starting quarterback goes down, a four-point adjustment could be warranted depending on how much experience the backup has. But if a school rotates quarterbacks and one goes down, no adjustment many be necessary. There have been several instances where I've raised a team's power rating following an injury, feeling that the replacement player was better.
But as a rule, a four-point adjustment for a starting quarterback or one of a basketball team's leading two scorers is sufficient for colleges. A star running back or key basketball player might be worth a two-point adjustment, while a starting star wide receiver or starter on the basketball team is generally about a one-point difference.
The other factor to look for is what are referred to as cluster injuries. While a starting lineman might not warrant an adjustment in power ratings, two or three injured linemen is a different story and could be worth an adjustment of several points, as could injuries to the secondary, wide receiver position, etc.
In basketball, that could mean a non-key starter and one of the first players off the bench.
For the pros, the maximum adjustment I'll make is three points and that is pretty rare and only for the Peyton Mannings and Kevin Durrants on the game.
If you follow the above guidelines by the midway point of the season you'll have better power ratings than the majority of publications and web sites who publish their own, which will give you an advantage when it comes to spotting lines that are off a bit.
But there is one other thing we can do to tighten our power ratings a little bit and that's look for misleading final scores and will be the subject of next week's article.
