9/14/2009 8:00:00 PM
One must be very careful when using raw numbers to handicap football, as it is common for teams to accumulate better stats vs. weak opponents. Thus, stats need to be adjusted.
Handicapping football using raw numbers is very dangerous, as these numbers do not take into consideration the opponents each team played.
The disparity is even more explicit in college sports, but it is colossal in professional sports handicapping as well. Rankings are just that, teams are ranked in specific categories first to last. Ratings have teams “rated” in various categories comparative to a mean number.
For example, let’s say Clemson is playing Maryland in football. Clemson runs for 252.4 (raw number) yards per game to rank (rankings of course) No. 7 in the country. Maryland as an illustration averages 239.8 to rank 18th. According to those “rankings” and “raw numbers” Clemson has a better rushing offense.
However rankings and raw numbers don’t scratch the surface. Conversely a rating would say the cumulative average of Clemson’s previous opponents’ defense allows 232.2 rushing yards per game. That would mean Clemson rushes for 20.2 more yards per game than their opponents normally allow (+20.2).
If Maryland’s cumulative foes allow only 197.8 that would put them at (+40.2). The inferior raw numbers make it look like Clemson is the better run offense by 12.6, but in the much more telling ratings, it’s actually Maryland by 20 yards per game.
So again, using the hypotheticals, here is a comparison (all illustrative rushing totals). The “advantage” numbers are ALL CAPS:
Raw numbers: CLEMSON 252.4; Maryland 239.8
Rankings: CLEMSON No. 7; Maryland No. 18
Ratings: Clemson +20.2; MARYLAND +40.2
Furthermore rating both offenses and defenses is most accurate using yards per play, yards per pass and yards per rush. These are much more telling as to whether teams outplay or underplay their stats. In short, games in which the proverbial “they have dominated them everywhere but on the scoreboard” are priceless to the handicapper. The more deceptive a won/loss record is, the more opportunity.
Team stats in those categories are a much greater precursor of future performance than points per game.
Maybe Maryland runs the ball 16 more times per game than Clemson. Yards per rush puts the raw numbers into better perspective, but yards per rush relative to the cumulative average of their opponents makes the stats rise to the level of truthful for handicapping excellence.
In basketball, shooting percentages offensively and defensively are more accurate than points per game. This is true in no small part do to the fact that half court teams will have lower scoring games than up-tempo. This is in no way to imply that ratings under this circumstance are flawless. Slow down teams will both get fewer easy baskets and give up fewer, but in also weighing points per game, the flaws of each statistic can cancel out some of the deficiencies of the other.
In short, ratings put raw numbers into perspective much more so than rankings. By no means whatsoever do we disregard rankings, but the square player is shockingly oblivious to the value of the more judicious numbers.
Yet again we must emphasize mere statistics are only part of the equation. But only a small percentage of handicappers are acute enough to use more precise ratings rather than the not to be trusted rankings.
Everything though we said about the strength and weaknesses of power ratings applies here. The learned player must make adjustments for injuries both for a team and that of their previous opponents.
Statistics can be used and statistics can be abused. But knowing the right valuations to use is just as important as knowing how to adapt them.