Point Spread Calculation

I have added a feature the the cumulative connections ranking system that predicts point spreads of future NCAA basketball games based on the difference in their ratings. The method uses a simple linear regression between all the score differences and associated rating differences observed during the season:


Using this method we can produce a point spread for the sweet 16 games being played tonight.

-Miami favored by 0.4 over Marquette (Vegas says Miami is a 5.5 point favorite)

-Ohio State favored by 5.3 over Arizona (Vegas says Ohio State is a 3.5 point favorite)

-Syracuse favored by 1.3 over Indiana (Vegas says Indiana is a 5.5 point favorite)

-La Salle favored by 3.7 over Wichita State (Vegas says that Wichita State is a 4.5 point favorite)

UPDATE 3/29:

Actual Margins:

Marquette over Miami by 10

OSU over Arizona by 3

Syracuse over Indiana by 11

Wichita State over La Salle by 14

So if you used this method’s numbers as a guide for how to bet against the Las Vegas spread, you would have won two bets (Marquette and Syracuse) and lost two bets (OSU and La Salle). Obviously for this method to be useful it would need to outperform the Las Vegas spread (have a success rate over 50%). More data is needed.

Here are the predicted spreads for tonight:

Louisville over Oregon by 22.3 (Vegas: Louisville by 10.5)

-Implicit recommendation: bet Louisville

-UPDATE: Louisville won by 8 so a bet on Louisville would have lost.


Michigan over Kansas by 1.1 (Vegas: Kansas by 2)

-Implicit recommendation: bet Michigan

-UPDATE: Michigan won by 2 so a bet on Michigan would have won.


Duke over Michigan State by 10.1 (Vegas: Duke by 2)

-Implicit recommendation: bet Duke

-UPDATE: Duke won by 10 so a bet on Duke would have won.


Florida over Florida GC by 27.1 (Florida by 13)

-Implicit recommendation: bet Florida

-UPDATE: Florida won by 12 so a bet on Florida would have lost.


After 8 games the method is 4-4 agains the Las Vegas spread.

This entry was posted in Sports. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s