Originally <a href='http://www.sportsbookreview.com/forum/showthread.php?p=26807401'>posted</a> on 02/22/2017:

Hello everyone, this is my first post on SBR! I've been posting on other websites about my NBA daily projections based on several statistical models I developed over the years. I'm a statistics university professor from Canada, and today I hope to give you some insight about the following question:

Is there evidence of pregame betting line movements having an effect on which team is going to win the game (in the NBA)?

I own a dataset which contains information about all games from the last 14 NBA seasons. I analyzed all games for which data was available on the opening & closing betting lines.

A fictitious bet was placed on the team in favor of which the line moved during the day. For example, assume the opening line has Team A favored to win the game by 5.5 points over Team B. If the closing line went higher, for instance 6.5, I would make a virtual bet on Team A at -6.5. On the other hand, if the closing line went lower, for instance 5.0, I would bet on Team B at +5.0. Notice that I am betting the CLOSING line (which is natural since we are studying whether we should place bets according to how the line moved during the day).

Let’s have a look at the results, where the first column indicates how much the line moved (absolute value of closing line – opening line):

Line movement Bets won Bets lost Ties Win %
0.5 1631 1673 38 49.4%
1 1131 1159 47 49.4%
1.5 664 669 12 49.8%
2 324 347 10 48.3%
2.5 229 210 6 52.2%
3 122 124 6 49.6%
3.5 49 65 2 43.0%
4 37 52 2 41.6%
4.5 29 24 1 54.7%
5 12 19 1 38.7%
5.5 9 6 0 60.0%
6 10 6 0 62.5%
6.5 7 7 0 50.0%
7 5 4 0 55.6%
7.5 5 2 0 71.4%
8+ 20 16 2 55.6%
TOTAL 4284 4383 127 49.4%


A person following the strategy described above would have won 4284 bets, lost 4383 and tied 127, which is good for a 49.4% win percentage. In other words, doing the opposite strategy would have generated a 50.6% win percentage. That is not enough to overcome the bookmaker’s vigorish (i.e. the percentage deducted from a gambler’s winnings as a form of commission).

I have highlighted in bold the win percentages that were below 50%. It is pretty interesting to note that they all occur in the upper portion of the table, which suggest following the plan below:


  • a) For smaller differences (4 or less) between opening and closing lines: bet AGAINST the team in favor of which the line moved;
  • b) For bigger differences (5.5 or more) between opening and closing lines: bet ON the team in favor of which the line moved.


Over the past 14 NBA seasons, portion a) of the plan above would have generated 4299 winnings bets, 4187 losing bets and 123 tied ones (win % = 50.7%). As for portion b): 56 wins, 41 losses and 2 ties (win % = 57.7%).

A word of wisdom: avoid putting this plan into action. Portion a)’s win percentage is not enough to overcome the vigorish (you need 51.2% at -105 and 52.4% at -110), while portion b)’s sample size is way too small to advocate putting any money on this strategy. But it’s still interesting to see the pattern that occurred here, and you could incorporate this knowledge into your betting strategy.

Now, what if we had had the opportunity to bet the OPENING line? In other words, if by some miracle God gave me a call to let me know in advance that the line was going to move by *x* points in favor of a certain team, how much of an advantage would I get by betting this team? Let's see the numbers:

Line movement Bets won Bets lost Ties Win %
0.5 1669 1625 48 50.7%
1 1215 1087 35 52.8%
1.5 736 574 35 56.2%
2 358 305 18 54.0%
2.5 265 172 8 60.6%
3 143 106 3 57.4%
3.5 68 45 3 60.2%
4 50 40 1 55.6%
4.5 34 20 0 63.0%
5 18 14 0 56.3%
5.5 11 4 0 73.3%
6 12 3 1 80.0%
6.5 10 4 0 71.4%
7 5 4 0 55.6%
7.5 6 1 0 85.7%
8+ 33 4 1 89.2%
TOTAL 4633 4008 153 53.6%


Overall, if you bet a specific team and the line eventually moves in their favor, you have a 53.6% chance of winning. You are in business! Only the case of a half-point movement is not enough to overcome the vigorish: even a 1-point movement yields a winning percentage of nearly 53%. Above 1 point, you are clearly in the driver's seat.

You may argue "that's great, but nobody knows in advance which way the line is going to move". That is true. But I believe those numbers should motivate you to try to anticipate line movements because we have clear evidence that it can be rewarding. When I saw those numbers, I figured maybe I should make my projections the day BEFORE the games occur so that I can possibly take advantage of early lines. Maybe I'll succeed, maybe I'll fail. But we may as well give it a try!

I hope I was able to bring you some value. I'd love to hear your comments about this little study!

Professor MJ