Monday, October 04, 2010

2010 Simulation Results

The 2010 regular baseball season is over and below are the statistics of how the simulator did for the 2010 season. Method of measurement is Return On Investment (ROI), which is calculated by taking the total amount won betting, divided by the total amount of money bet.

 Measurement ROI Overall 3.42% April 8.95% May -1.37% June 1.20% July 5.52% August 1.28% Sep/Oct 5.66% National League 4.98% American League 1.53% Interleague 5.03% Favorites -4.00% Underdogs 8.04% Visitors -0.75% Home 6.06% Home Underdogs 15.55% Home Favorites 0.19%

To put this in perspective, let's say that you average bet was \$10 on each game, (this is a simplified example) - but just to get an idea on what you'd make, it would be around \$730 and \$7300 on average bet sizes of \$100. From the table above, the simulator cleaned up on "home underdogs".

Josh Engleman said...

I have a couple questions for you. Have you thought about looking at how the sim would have done using Kelly? Do you have an idea how often you beat the Pinnacle no-vig closing lines? How often did the lines move with your predictions? I just found your site after seeing your post on Tango's site. Good work.

Xeifrank said...

Yes, my betting system used the Kelly system for bet sizes. I didn't keep track in detail of how often the lines moved with my predictions as I only have a limited amount of time to work on this. Thanks for the great question, feel free to visit any time.

Pat said...

Very nice results. I've followed (very lightly) your simulator for a couple of years now. I was wondering if you had any totals not on ROI but on the overall accuracy of the simulator. You'd probably have to do ranges but in a format for example like this:

There were 400 games where a team was given 60-65% odds of winning, the favored team won 63% of those games.
There were 900 games where a team was given 55-60% odds of winning and the favored team won 54% of those.

Just as an example but I would love to see a grouping with percentages to compare to the simulator results.

Thanks and great work!

Xeifrank said...

Great question Pat. That is something I should be able to do, just not sure when. When I do do it, I will post the results in a blog entry (not this thread), since it is probably worthwhile of its own post. Hopefully there are enough 2010 results to make this work.