Predictability
Talk to anyone who bet on sports 10-plus years ago and they’ll reminisce about the time when realizing profits was far more attainable with the right information. Back in the day, rogue numbers and middle opportunities were ever-present for savvy bettors to take advantage of and those with access to actionable knowledge enjoyed a pronounced edge over the competition. Today—in a world dominated by technology where information becomes instantly ubiquitous—those with valuable insight have only minutes, if not seconds, after lines are posted to get their money down before the marketplace corrects itself.
The NFL betting market has undoubtedly become more efficient over the years so this should cause lines to also improve in accuracy, right?
This was my thought process which led me to the question: Are point spreads today better at predicting NFL contests than those 10 or 20 years ago?
In short, the answer is no. Here are the numbers…
| Seasons |
% of games within 3 pts. of line |
within 7 pts. |
within 10 pts. |
|
1989-92 |
22.0% |
45.2% |
59.2% |
|
1993-97 |
21.7% |
45.1% |
59.8% |
|
1998-2002 |
20.8% |
45.2% |
60.2% |
|
2003-07 |
21.6% |
43.0% |
57.5% |
|
2008-12 |
20.3% |
44.9% |
57.8% |
*Note the point spreads used in this analysis are those that were bet into by the greatest number of people, not necessarily the opening or closing line.
Obviously everyone wants to know what the takeaway is from this analysis. Although the NFL betting community comes to a consensus line far more quickly in 2013, this agreement has not translated into an improved ability over the years to forecast the outcome of football games. But enough of the college kid, let’s get an expert’s take on the topic.
Todd’s Take: Justin is completely right in his analysis that the market in 2013 corrects itself faster than ever before in sports betting. Information is widely available in real time and those “insider trading” tips of yesteryear are increasingly difficult to keep under wraps. However, I don’t agree with the thesis that lines haven’t grown more accurate over the years. Pointspreads are not designed to be predictive measures of the future but rather to force bettors into making price sensitive decisions.
If I know my true line as an oddsmaker on a game is Pk yet the public will largely lay -3 with reckless abandon, there’s no benefit to opening the price too short and giving away the store. Forget arbitrage in these scenarios, it’s as straightforward as supply and demand. If I take a sharp bet at +3 then I know I can move off the key number and let the public lay through -2.5 with limited reservations about accruing liability. No matter the sport, there will always be variance when comparing final scores to closing lines and I don’t feel this is an indictment of a line’s accuracy. If you ask me the best indicator for how precise spreads are during the course of a full season is a one-word answer: profit. The only thing that matters is a sportsbook’s balance sheets because if they’re generating more revenue now than they did 10 years ago, I’ll contend oddsmakers are still doing exactly what they need to in the grand scheme of things by adjusting to a dynamic marketplace.
