average point spread of actual NBA games played

[widebody2]

Does anyone know the average point spread of NBA games that were played in the past. Not looking for the odds...looking for the actual numbers from the games that were actually played. Whats the average and if there is a website that lists these types of numbers I would love to know what it is.


At first I thought this number would just be the average of all team pt differentials which I've figured to be 3.96 for the 2010-11 season but this number is actually not a real life example of what really happens in NBA games. 3.96 is an average of all the averages....these numbers include both wins and loses.

Here's an easy example of why this method does not work:
Team A wins 5 games by 10 pts each
Team A loses 5 games by 8 pts each

Team A now has a point difference of +2, meaning they average a win by 2 points.....but the actual pt spread average of these games was 9.

Off the top of my head I would guess an average NBA point spread is somewhere around 10.5. Does anyone know the actual number?

[ManBearPig]

You have to know SDQL but you should be able to get this info from the Killer Sports database. I ran a quick query from mine from 2003 the average line is about -3.3 for all games played. Without looking I would expect the average total to be around 199. Hope this helps.

[widebody2]

No it doesn't help....not looking for the average line. Not looking for an average of all team's averages as I explained. In simple terms what I am looking for is:

1. take every game played in the NBA during 2010-11 and subtract the losers score from the winners score.
2. Add up all those numbers and then divide by 2460

I think it should be somewhere around 10.5

[ManBearPig]

Using your method.

It's looks to be around 3.16...if broken down by H vs. A...or 1.57 overall (+/-3886/2460)

[widebody2]

check out this link:http://espn.go.com/nba/standings/_/type/expanded

Look at the team records for games decided by 3 pts or less and look at the records for games decided by 10 points or more. All teams have significantly more games decided by 10 pts or more than 3 pts or less.

There is no way the numbers you are coming up with are correct.

Maybe you are not understanding what I am really looking for or maybe I am losing my mind. I understand that 3.5 or so may be a typical point spread predicted by Vegas but that's not what the games are typically won or lost by....Vegas point spreads are the tipping point where they get 50/50. 50% favorites score both above and below that tipping point.

The lines are lower than the actual pt spreads that really occur in the games because the oddsmakers need to consider the underdog also has a chance to win.

[suicidekings]

check out this link:http://espn.go.com/nba/standings/_/type/expanded

Look at the team records for games decided by 3 pts or less and look at the records for games decided by 10 points or more. All teams have significantly more games decided by 10 pts or more than 3 pts or less.

There is no way the numbers you are coming up with are correct.

Maybe you are not understanding what I am really looking for or maybe I am losing my mind. I understand that 3.5 or so may be a typical point spread predicted by Vegas but that's not what the games are typically won or lost by....Vegas point spreads are the tipping point where they get 50/50. 50% favorites score both above and below that tipping point.

The lines are lower than the actual pt spreads that really occur in the games because the oddsmakers need to consider the underdog also has a chance to win.

ManBearPig's numbers are accurate. There is some variation from season to season, but the average margin of victory (across all games) hovers around 3.1ppg. Not vegas lines, but actual game results.

[ManBearPig]

Hmm...your right maybe I don't understand so I'll do you one better. Here's all the games from this season and the scores...knock yourself out. Based on the example you gave above it sounded like an avg margin of victory you were looking for, which is what I'm giving you not the average point spread.

I know you have a number in your mind that your expecting to see but this is the way the numbers average out...the math doesn't lie. If I took the ABS value of the point diff it would probably give a larger number due to the fact we're eliminating who scored more and looking at the raw differential data, but in your example 40-50 = -10/5 = -2, so I left the actual numbers. The numbers will reflect the same whether it's H-A or A-H but I don't know if you care about it being a +/- for your purpose.

Like I said KillerSports will let you query their DB but you need to know SDQL (Sport Data Query Lang) or check out sites like Basketball Reference or NBA Stuffer...there are lots of places to get data. Either way you have all the raw scored for each game this season and you can manipulate however you wish.

[widebody2]

manbearpig thanks for the help. Actually I figured out at least one of the problems. I told you to devide the total point difference by 2460 because that's the 82 games x 30 teams but that actually should have been 1230 because it takes two teams to play one game. Sorry about that.

So that brings our average point spread in an NBA game up to 6.32. To me that seems a little more realistic but I was still expecting a bit higher. I did a 35 game sample towards the end of the season and came up with like 11....but maybe that sample had a few large blowouts to skew the numbers.

Unfortunately I don't know how to use excel. Heres a real quick sample from the first 13 games of the season:
Absolute value pt diff:
8,2,14,15,5,8,4,10,3,10,4,11,15

That average margin of victory is 8.38

The first 35 game sample I did was 10.7 but it did include one 32 pt blowout.

In order for the average margin of victory in the NBA to be 3.1, for every 20 pt blowout we would need 8 games that were only decided by 1.

Suicidekings that 3.1 you are talking about is the average pt diff. Which is actually an average of all teams averages. This is not the average margin that an NBA game is decided by. I initially thought the same thing. Here's another quick example of why its different.

The knicks win 10 games by 20 pts each. They lose 10 games by 20 pts each. Their pt diff is actually 0, but those games were decided by an average of 20. See the difference?

[widebody2]

To answer my own question, for anyone who is interested, the average NBA game is won or lost by 6.32 points and the average vegas line for this season was 6. Over the course of the past 3 season's the average vegas line was 6.1

I am fairly certain of these numbers and anyone coming up with an average spread somewhere in the 3.xx is reading into the queries incorrectly, as I did when I first tried to figure out the answer to my question.

Average point spread involving a road favorite is 4.4. Average point spread involving a home favorite is 6.7. There are many more home favorites than road favorites and that is why we have the 6 point average spread.

[HoulihansTX]

7, and 5

[benschwab]

The actual number is 10.57 from the data ManBearPig has given though the data seams to be missing 3 games. Notably the median is 9 while the mode is 5. These are the measurements of central tendency for the data of the margin of victory for each individual game. The histogram for the season given is bellow:



If one averages the point differential for each team on the season one should get 0 as for each game a team wins by x points another team looses by x points.

It should be noted that these statistics are only for the 2010-2011 season.