Here, we'll take a look at a slightly different way of handicapping NBA games. It is a bit similar to some of the other methods we have previously looked at, but does have several key differences, primarily the scoring averages that are used.
This method uses league average/league
median, which is something that's been discussed a number of times and will
give predictions for both sides and totals.
The
Method
The stats required are simply the
average league scoring for both home and away teams and the average home or
away scoring for the two teams involved in the game you wish to predict.
Using an NBA game with Milwaukee at
Brooklyn for an example, the first thing you will need is the average points
scored and allowed by home teams in the league. Teams that are at home average
98.5 points per game and allow 95.2, so naturally road teams score 95.2 and
allow 98.5.
Next, you will compare each team
against the league average in both offense and defense to come up with a
percentage of how that team does compared to the league average.
Milwaukee is averaging 96.6 points
on the road and allowing 97.7. You will divide Milwaukee's points for into the
average points scored by road teams, giving you 96.6/95.2=1.011, which means
Milwaukee's offense scores 1.1% more points than the average road team.
Next, divide Milwaukee's points
allowed of 97.7 by the average points allowed by road teams, which is 98.5 and
you will get 97.7/98.5=.992, meaning Milwaukee allows .8% fewer points than the
average road team.
The home team, Brooklyn, averages
97.4 points at home and allows 94.8, so you will divide 97.4 by 98.5 and get
.989, meaning Brooklyn scores 1.1% fewer points than the average home team, and
you will divide 94.8 by 95.2 and get .996, which means Brooklyn allows .4%
fewer points at home than the average team.
To calculate a team's predicted
score, you multiply its offensive percentage by its opponents' defensive
percentage by the league average. For Milwaukee, you will have
1.011*.996*95.2=95.862. Your predicted total for Milwaukee in the game is
95.862 points.
For Brooklyn, you will have
.989*.992*98.5=96.637, which is the predicted total for Brooklyn. Your line on
the game would have Brooklyn favored by .775 or 1 point and a total of 192.499
or 192.5.
The actual line on the game, which
was played Tuesday, Feb. 19, was Brooklyn -4.5 with a total of 194.5 and the
Nets escaped with a 97-94 victory, which was pretty darn close to our
prediction on the game. (If they were all so easy.)
Unlike the other methods that
require you to either add several points to the home team or subtract from the
road team and add to the home team, here there is no adjustment that needs to
be made, as you are using home and away averages, so the home court factor is
already built into your numbers.
Our sample
size is still close the lower limit, meaning there are enough games home and
away games that have been played to be able to use those stats separately, but
just barely. So I would probably look for differences of at least five points
between your prediction and the line for both point spreads
and totals.
