Spread records are often misleading

Nov 2, 2004 1:31 AM

One of the most commonly referred to statistics that is an all-out gambling number (as opposed to a conventional/official stat used for prediction purposes) is the "spread record."

You’ll often hear people talk about how a certain team is 5-0 against the spread at home or 8-1 as a favorite, as if this somehow signified a quality pick to beat the line in the next game. Similarly, many handicappers see a poor spread history as a sign of doom for a team when placed in a similar circumstance.

We are here to answer the question of whether NFL team specific spread records have any predictive value. If they do, great! Then we will know they should be something we look at when pondering a matchup. If not, even better. Not only can we discard the information from consideration, but can buck the tide from those who do follow the numbers.

For a first pass analysis, we’ll simply use in-season spread records to try and understand whether backing the "hot team" and fading the cold ones is a smart way to go.

The numbers show the cover rates in Week 4 and beyond. Basically they tell us that there’s next to no value in the current season spread record, at least on a surface level.

In looking for matchups where one team has a good spread record and the opponent a bad one, there is no reason to think spread records will be any help in predicting the future games. Playing the team with the better spread performance on the season coming into the matchup yields a 50 percent win rate.

This seems logical, given that the lines are adjusted based on a team’s performance. A bad team so far will be getting more points as a dog or laying less on those rare occasions it’s a favorite, and vice-versa.

Will teams that win games straight-up but lose against the line, or lose and cover change the outlook some? No.

All right, so far we’ve been putting the kibosh on spread records as something a handicapper should think about. The adherents to this factor though seldom emphasize current year records (after all, the NFL season is short), but instead discuss with enthusiasm things like a team’s record as a "home favorite."

For the next test then we’ll grab a team’s last 10 games in a similar circumstance (home favorite, home underdog, away favorite, away underdog) and see if the cover percentage during that span influences the chance of a cover in the next one.

Backing a team doing well in the past in similar circumstances would have cost you money over the years, since the results are slightly the opposite. There just aren’t many cases of teams with extreme records, so we believe historical spread records don’t matter.

There is one ray of hope for spread stat lovers. Teams which have been good plays as home dogs in the past have done well in their next game as a home dog. This, however, is most likely due to the overall success of home underdogs through the years.

Okay, but maybe we’re not giving the spread records their due. After all, if both teams in a matchup had good spread records in that situation, why would you presume it would

tell us anything? So for the final look, we compare the spread performance in the category for one team versus the other and look at what happens if you play the better spread history team. (This again uses the last 10 games for a team as a home fav./home dog/away fav./away dog.)

Our verdict: Past spread performance is not a useful indicator for forecasting future efforts. Now there are still some diehards who will staunchly stand by their favorite handicapping numbers. They will cry out that if only we had tested a team’s past spread won-lost in "divisional games" or on "artificial turf" or "when traveling cross country" or "when playing the second straight home game," then we would have witnessed the true power of spread records.

All we can do then is leave the further research up to you and restate our opinion that your time is being wasted.