The premise behind the Deviation Factor is that teams will exceed their average difference from the league norm. In this case, it relates to NBA teams and totals.
The first step is to determine the
league average of points scored and allowed. Because the NBA is essentially a
closed system, meaning that all games are played against other NBA teams, these
numbers will be the same. (That isn't the case in college sports, as Division I
teams often play Division I-AA or Division II teams in preseason games, etc.,
which is why the average points for is typically greater than points allowed
for college football or basketball.)
For ease of demonstration purposes,
let's assume the average points scored and points allowed in the NBA is 90. A
game played between two teams averaging and allowing 90 points per game should
have a total of 180. That should make sense.
How about a game played between two
teams that average and allow 100 points? While the simple answer would be that
the game should have a total of 200 that's not entirely correct because it's
not considering the fact that both teams exceed the league average. If teams
are scoring 100 points against a league that allows 90, they should score more
than 100 points playing against a team that allows 100, or 10 points more than
the league average.
We assign a points rating for each
point difference from the league average. A team that had a two-point
difference from the league average received a 3, a team with a three-point
difference received a 4.8, etc.
We end up with four numbers, one for
points scored for both teams and one for points allowed by both teams. We then
total them up and add or subtract from the league average to get his predicted
total. It’s a good method, but is time consuming and could be a bit confusing,
so what we'll do is use a modified version with a base deviation factor of 1.5.
Modified
Deviation Factor
Using our earlier example of teams
averaging 90 points per game, we have a base of 180 total points for an NBA
game. So a game involving two average teams would see a predicted total of 180
points.
Now, let's use our example of two
teams both scoring and allowing 100 points per game. The first step we will do
is to add all four numbers (Road team points scored, road team points allowed,
home team points scored and home team points allowed) together and divide by
two. In this case it's simple, as 100+100+100+100=400 and 400 divided by two is
200.
We now have a prediction total of 200, which is 20 points greater than our base league average of 180. Our next step is to take our 20 point difference and multiply that by our base deviation factor of 1.5, which gives us a total of 30. Our predicted total is now 30 points greater than the league average of 180 which is 210. So our predicted total for the game is 210 points.
Rather than try to explain the
process again, we'll give a few more examples.
If we have a game involving two
teams that average and allow 87 points, our first step is to add 87+87+87+87 to
get 348, which we divide by two to get 174. Since 174 is six less than our base
of 180, we have a figure of -6, which multiplied by 1.5 becomes -9. If you add
-9 to 180 you get a predicted total of 171 points in the game.
One last example, this time
involving a team that scores 98 and allows 102 playing a team that scores 85
and allows 81. Adding together all four numbers gives us 366, which divided by
two gives a figure of 183. Subtracting 180 from 183 gives a total of 3, which
multiplied by 1.5 becomes 4.5. Our predicted total for this game becomes 184.5
points.
