More books equal better results

I’ve presented several studies in this blog that reach desirable conclusions. My latest study is not one of those. The question that I wanted to address is how significant is having more books on your potential results? Clearly having more books will improve your results. Sportsbookreview.com has a table of odds from various sports books that is updated regularly. You can glance at these to see different odds at different books. If you have accounts at two books, then you would take the one with the best odds. Add a third book, and there’s a chance you will get better odds. In the long run, will these will add up to a sizable difference? The obvious answer is yes, but I wanted to analyze.

Take Arizona -1.5 recently against the Mets. The odds range from +150 to +160. That is a non-trivial difference. I wanted to get a sense of how much this type of a difference this would make in the long run. I used the run line system that I’ve discussed previously. I have data from 8-10 books. I built a wager line for each book that met the run line system requirements and sorted by decreasing odds. I decided to look at the top 3 lines instead of just the top one. I was concerned with potentially bad data, hence used multiple lines. Following are the results.

Line	Number	Bet	Net	Ret/$	Ret/$ 100
3rd Best Line	961	$122,400	$6,956	$1.057	$1.059
2nd Best Line	988	$125,700	$8,738	$1.070	$1.073
1st Best Line	1031	$130,600	$11,088	$1.085	$1.089

As the lines got better, the number of wagers increased. That’s because more books met the minimum odds requirement (+140 and above). Also notice that the total bet increased. That’s because I bet more when the expected return increased. Finally, the Ret/$ increased significantly because the increasing odds. All of this simply confirms what was intuitively obvious. The more choices you have to pick from, the better results you will get.

But there was one bit of surprise when I was looking at the numbers. Instead of increasing the wagers as the expected returns increased, I looked at what would have happened with a flat $100 bet on each pick. The last column shows the returns per dollar on these wagers. These were slightly better than the returns when I wagered more on some of the picks. That may indicate that increasing the wagers with higher expected return isn’t particularly productive.

That result is disappointing to me. But unfortunately, it mirrors similar conclusions I’ve drawn in the past. I’ve used similar progressive betting systems in the previous years (like Kelly criteria). When the season was over, I’ve compared the results with those I would have achieved with flat bets, only to conclude flat bets would have been better. What does that tell me? I don’t know exactly, but it’s time to finally get to the bottom of this. My next project will be a more detailed analysis of these type of wagering systems. Follow me on Twitter, @ole44bill, to know when I post this stuff.