Going Under In The NBA

I've spent a little more time looking at the free football data and while I was focused on Pinnacle's numbers, it seems rude to ignore those of Bet365. 

Having done so, I can see why I pretty much ignore other books. Based on 3306 matches where we have data from Bet365 and Pinnacle for both the Money Line and the Under / Over 2.5 Goal markets, the average over-round (Money Line) for Pinnacle is 102.4% compared with Bet365's average of 106.7% while for the Over / Under market the numbers are 102.8% and 105.8% respectively.

If you were stupid enough to back the Home, Draw, Away, Under and Over in every game, I hope you'd have at least saved yourself 622.43 units by betting with Pinnacle!

Backing the Draw at the better of the two prices in every match would have resulted in just about breaking even, though not quite, with a loss of 29.49 units, or slightly less than 1%. 

As with financial investing, the impact of commission or fees on your transactions should never be an after-thought. The impact of higher fees on actively managed funds over time has been covered in this blog previously, and one big reason why investing in low fee index tracking funds is my preference, but the same logic applies to betting, with the added complication that the "best" prices might not always be an option for you, in which case you have to go with the best prices available to you.

Polish NBA fan Wojciech Malinowski tweeted:

I'm not sure if Pinnacle operate in Poland or whether Betfair allow bets from there, and if so at what rate of commission, but for sure, overcoming a 12% tax would be quite a challenge to say the least.

Ben sent me this suggestion a few days ago, and rather rudely I don't think I ever said thank you as I was away at the time, but hopefully he will accept a late thank you here and not mind me sharing this personal recommendation more widely:

Have you looked into using a betting brokerage platform like Sportmarket to place your wagers (avoiding the Betfair premium charges)?

I’ve been using the company for a bit over a year now and have withdrawn with ease each time.

I regularly beat the Betfair starting price (inclusive of commission), simply due to the availability of multiple asian books.

Personally I have no experience of this platform, but Ben is a quality individual and I have no reason not to trust his opinion here.

Regarding the second part of Wojciech's comment on the NBA Unders, and the challenge right now is that finding Unders is a little like trying to find a stock to short in a soaring bull market.

They are out there, but a rising tide lifts all boats as the expression goes, so caution is needed. As a refresher for any new readers out there, point totals in the NBA have increased quite dramatically since the 2012-13 season. It's been covered in this blog for some time, but here is an interesting article from this season with tables to show the changes:

The 2012-13 season, or seven years ago, is when the league truly started exploiting the math and realizing they could reallocate some of the hardest field goals to convert into jump-shots with a greater reward (three-pointers).

Sports and consequently sports markets do evolve, even if it is impossible to prove to the satisfaction of a statistician, but then:

Arguing with a statistician is a lot like wrestling with a pig. After a few hours you begin to realise the pig likes it.

To find early signs of changes, it is important to keep tracking key metrics in games. 

If you see the average number of possessions changing, the likely effect is a corresponding change in point totals. 

If you see the number of three-point shots being taken is increasing, likewise. 

Are all teams changing their tactics the same way though? 

Fortunately (from an investing perspective) not. Three point attempts this season range from 44.3 per game (Houston Rockets) to 27.4 (San Antonio Spurs), while possessions per game range 104.4 (Milwaukee Bucks) to 97.1 (Charlotte Hornets).

Not coincidentally, the Houston Rockets lead the league in points followed by the Milwaukee Bucks, while the Charlotte Hornets languish in 28th place (out of 30 teams).

I'm getting slightly away from the original question, but the point I'm trying to make is that any evolution isn't likely to be taking place at the same pace across the league, and with teams playing in conferences, changes may be somewhat localised. 

So, is there value to be found backing Unders when the total is set low? The bias that is in play here is that if the total is set low, instinctively less-informed punters will see value on the Over.

What is a 'low' total? Because this sport is changing so fast at the moment, there is no mean total, so I start with looking back three seasons and the 211.2 number. 

Unders when the total is set below 211 over the two and and a bit seasons since hit at 56.3%. This increases to 57.2% in games between Western Conference teams, I suspect because the West is known for being higher scoring.

Probably the most active bias in sports betting is that of recency and in games between teams where both are coming off an Under, the win percentage goes up to 59.5%.

I'm not sure I like the name 'Anti-Beast' for a system though. 'Teddy Bear' perhaps, but the answer to the question is yes, value is there on lower point totals. At least for now.