Using the Closing Betting Odds to test for Skill Part 2
Posted 3rd July 2017
In December 2016 I wrote an article looking at 4 different tipsters to see whether their tips were beating the closing price. The closing price provides a useful measure of whether a punter has any real long term profit expectation. Indeed, I have previously published an article showing that the ratio of the odds you bet to Pinnacle's odds at closing is a good predictor of your expectation.
At the time, I made some predictions about these 4 tipsters: Jeremy Price would show a long term yield of 6%, Eredivisie2000 and Aris Papadopoulos would regress to their mean of no profitability at all and Pinnacle Tipster was cheating.
Before turning to the main point of this article, let's briefly see how those predictions stacked up. Dealing with the last one first, it's just a simple thing to observe that Pinnacle Tipster has fewer recorded tips than he did in December 2016, so I don't think anything else needs to be said about this fraud. Eredivisie2000 decided to stop verification with Betrush soon after a bad February 2017. His overall record there finished with an ROI of 98.24% from 82 picks. He continues to post on Blogabet but over the 6-month period has lost money. Similarly Aris Papadopoulos has fallen backwards. Finally Jeremy Price is now showing an ROI of 106.86% from a total of 166 picks. It would seem my initial predictions about how these tipsters would perform proved to be quite accurate.
Let's now take a look at another betting record. I wanted to show it because it provides a nice example of how the closing price really is a reliable indicator of the true odds. Consistently beat it by 10% and you can expect to see a 110% ROI over the long term. This is a private record of bets on Australian horse racing. All of the 3,354 bets are win-only placed with Australian bookmakers. Together with the price he bet, the punter also chose to record Betfair's starting price to allow us to make this analysis. He managed to beat the starting price 56.5% of the time. [Read my latest Pinnacle article to find out how often you should be beating a bookmaker's closing line to have a chance of being a winner.]
The first chart shows the betting history time series (blue) and compares it against what would be expected (red) assuming this model. Whilst the actual returns do underperform somewhat compared to expectation during the latter half of the history, this is not beyond the margins of chance. A 10,000-run Monte Carlo simulation confirms that the difference from expectation is less than one standard deviation.
The second chart re-orders the bets in descending order of expectation. That is to say bets with the highest expected value (as defined by Betfair starting price to odds bet) appear at the top. This effectively shows how closely actual performance mirrors expected, for all but the worst value bets where the starting price ended up much longer than the odds the punter managed to bet. Again, this is arguably due to chance. For these bets, the strike rate is so low (just 4 winners in 443 bets where expected return on investment was less than 60%) that variance is a significant factor. Just 4 more lucky winners in this sample would have seen actual returns more or less mirror expectation.
If you're wondering whether your betting record is showing evidence of skill, don't look at the size of the bank balance; good or bad luck have too much of a short term influence. Instead, find out whether you're beating the closing price, and more specifically the closing price of an efficient bookmaker like Pinnacle or an exchange like Betfair. These brands are the ones most likely to have odds which reflect the true result probabilities. Other brands may not be so accurate. To learn more about why, have a read of another of my articles on market efficiency at different bookmakers.