[Cookie Monster]

I have been thinking about the best "basic strategy" for an action points contest. In contests with lots of players and not too many picks, taking a high variability approach is the best strategy. It increases the chance for a very high score (ie prize) at the cost of increasing the probability of a very low score.

So, looks like taking basketball totals is a no brainer. But I have 3 assumptions that would like to be confirmed:

1.- NBA totals have a higher variability (ie average action points) than NCAABB, as the totals are in the 200 zone, compared with 130s of hoops.

2.- The high total lines (220s) have a higher variability than the lower ones (180s). That also sounds logical.

3.- The over has better action points average than the under. There are two reasons: The overtime, and the fact that the line is a median, and the distribution curve has a slight skew in the form of a Poisson curve.

So, the best "basic strategy" should be taking the over in high totals NBA lines. But then, in high scoring games, the possibility of overtime decreases, and maybe the distribution curve start skewing to the other side on such high lines.

So, anyone can test my assumptions against a database? Is "over on totals 220+" the best basic strategy?

[LegitBet]

You can still fat hammered as in the case of the Knicks heat game thurs. I was playing that live and it was amazing to see the total plummet after the first quarter. Conversely low total games offer such a nice 'floor', and I can visualize the upside with less variance. What do you think?

[batigol]

Very,very difficult question !

[Cookie Monster]

I made a few queries at an online NBA database (1995-today), and the result were more or less what I expected.

First, the over/under have way more action points than the spread. The average spread action points is 8.9, the totals average action points is 13.5 when the under wins, 14.3 when over wins. The numbers for over and under should be slightly less, as I did not include the pushes. That also shows that the distribution is slightly skewed towards the over, mostly due to the overtime.

However, the effect of the totals line is far less than I expected. Certainly, the high lines games have a slightly larger average action points than the low lines, but the difference is not that much. For example, the games with line of 178 or less, have an average action points of 13.0 when under wins, 13.9 when over wins. The games with a total line of 216 or more have an average action points of 15.4 when under wins, 14.8 when over wins. So, my original thoughts were right. The high lines have a larger action points potential and the distribution skews towards the under.

In conclusion, most of the advantage comes from simply taking NBA totals. The high lines have slightly better average, but far less than expected. I have no access to NCAA basketball database, but I am almost sure that the average action points are lower. Then, answering the original question, the best basic strategy is to pick only NBA totals, with an slight preference for higher numbers. But any number deserves attention, especially if there is value in the over.

Thanks to dvsbmx for posting a link to the database.