by Mike Hayes - 09/23/2005
If there was ever a doubt that information in the wrong hands could be dangerous just take a look at some of the so-called NFL trend analysis that litter internet sites, newspapers and television broadcasts these days.
While some trends can be useful in helping make selections, many, if not most, can be deemed useless data that does little more than fill space in a newspaper or add further graphics clutter to a network broadcast.
For example, one of the papers in my area offered this insight last week, "The Lions are 0-6 ATS as a favorite of 3 points or less since 1992."
So let me make sure I've got this straight. Every other year, once a year, the Lions are favored by 3 or less and have never covered. WOW, with information like that the bookies are bound to go broke.
Ironically the Lions did get buried by the home Bears which must have left them pounding their chests at another newspaper that offered this nugget prior to kickoff, "The Lions haven't beaten a current divisional foe back-to-back on road since 1991."
The only analysis needed here is the undisputable fact that the Lions have basically stunk since 1992 and so, for the most part, have the Bears, meaning the game was pretty much a toss-up in the eyes of the bookies.
The fundamental problem with analyzing trends as far as the NFL is concerned is that the sample used is often too small - "the Saints are 7-3 ATS on road in last 10"- to have significant meaning or too large - "Bengals are 15-33 ATS in September since 1992" - for the same reason.
While the Saints 70 percent cover rate on the road sounds impressive it is just two games over 500 with a sample of just 10 games - just two games better than the likely outcome if you wagered on these games based on the flip of a coin. The 70 percent cover record could have more to do with the caliber of team the Saints faced over that stretch or simply happenstance.
The sample concerning the Bengals is too large because it covers 12 years. Over this time the Bengals have had four different head coaches, seven different starting QBs and too many different players for this statistic to have any real meaning. In fact, it took me about five minutes of additional research to learn that the Bengals overall record since 1992 is 68-140, a winning percentage of .326 barely better than the "15-33 record ATS (.312 Percent) in September since 1992." Again, the only analysis needed here is that the Bengals have been a bad football team since 1992, not only in September but in October, November and December as well.
When looking at a trend it is important that it be analyzed further to determine whether it makes logical sense. I came across this "trend" this week. "The Giants are 9-15 ATS in their last 24 with fewer than 6-days rest" While a short practice week can certainly influence a team's performance on Sunday this needs further review as well. Without doing any research it is safe to conclude that this "trend" covers quite a few years as most teams are likely to play just a few games a season on short rest. Again, this "trend"covers too many different variables to be meaningful. If Tom Coughlin were the Giants coach the entire time and the roster was for the most part the same, it might have more meaning.
It is also interesting and perhaps most important to note that the 9-15 record results in a .375 percent success ATS over the last 24 games in which the Giants played on short rest. Over the last two seasons the Giants have posted an overall record of 12-20 ATS - the exact same .375.
Some of the best trends can only be determined after it is too late to take full advantage of them. The Chargers for example were clearly better than handicappers thought last year as evidenced by their 13-1 cover record ATS.
I also find the fact that the Patriots have not lost two games in a row in more than two seasons to be of value especially considering that the Patriots have covered the spread at a remarkable 24-6 rate over the past two seasons.
Obviously trend analysis is not and can never be a precise science or bookmakers would go broke. No useful trend is going to select winners at a rate much greater than the 60 percent or so that a solid handicapper can expect in a winning season. Also, any useful trend is likely to become public knowledge and influence the point-spread and betting accordingly making the trend less useful over time.
Oh, by the way………..after starting 0-4 in September home games under coach Herman Edwards the Jets have won two in a row.