Thursday, 9 November 2017

The Boys Are Back

The College Basketball (NCAAB) season begins this Friday, and having watched the rising totals in the NBA, here are some observations on the College version.

It's important to note that the NBA and College games are quite different. College games are shorter, with two 20 minute halves rather than four 12 minute quarters, the shot clock is 30 seconds rather than 24 seconds, but with an extra two seconds to cross the half-court line, and the college game has a possession arrow.

The three-point line is closer to, and always the same distance from, the basket (unlike in the NBA), and the college lane (paint) is 12' wide rather than 16' wide. 

Possibly the biggest difference in how the game is played is because of the rules around fouls and defences. If you're interested, there are plenty of places you can read up on these, but I don't have all day. 

The other confusing aspect for the College game is the number of teams - 351 spread over 32 different Conferences. There are also a lot of mismatches, so you might want to wait for the Conference season which starts in January. Not that there aren't mismatches on Conference games, but the average line over the last three seasons in Conference games is 3.4 while in non-Conference games it is 8.7.

Despite their many differences, backing Over when the Total is set high is also currently a successful strategy in the College game.

In the past two seasons 2015-17, Overs has a 120-101-2 record which is a 5.9% ROI.


The increase in points over the past two seasons is clear, at least in part because the the shot clock was reduced to 30 seconds in 2015, so older data isn't relevant.

The same two seasons also show an edge on road teams favoured by between 6 and 10 points in Conference games:
If that's too many selections for you, try the 7 to 8.5 point range:
     

1 comment:

Habotron said...

Hi,

I think there is a mistake in the first image. Total P/L and ROI%. In the text it is ok 5,9ROI%, is it?

Anyway I really love you blog.

Thanks!