Using Pinnacle.com's Closing Line to Predict Profits
Posted 4th August 2016
In a previous article I talked about the importance of the closing line as a means of assessing whether a punter or tipster is capable of consistently beating the market. Pinnacle.com, furthermore, have published an article identifying what distinguishes winning bettors from losing ones. To reiterate, bettors can and do show profitability. The more important question is whether they are doing that because they have been lucky or because they are skilled. If the former, those profits will regress to the mean, that is to say, disappear. If the latter, they should be more consistent and long lasting. One way to set about analysing the distinction is to see by how much the bettor can anticipate the future movement of the market. And arguably the best way to measure that is to measure the magnitude by which he can beat the closing odds.
According to the efficient market hypothesis, the closing odds provide the most efficient or most accurate representation of the probabilities of actual results, since they reflect the most amount of information expressed in the form of wagers by the betting public. Writing for Pinnacle.com, I have shown just how efficient their closing odds prove to be. In that article, I formulated a hypothesis: the ratio of the odds bet by the bettor to the closing odds should provide a prediction of expected value he has managed to achieve. For example, if he bet a team at a price of 2.50 which subsequently closed at 2.00, his expected value would be 2.50/2.00 or 25%. This expected value would thus be equivalent to his expected yield over a larger sample of wagers.
Football-Data has now made available 4-seasons of closing match betting odds from Pinnacle.com dating back to 2012/13. Users of the historical betting odds and results data files can find these figures at the end of each spreadsheet. Including prices for all home, draw and away options gives a database of 87,960 betting odds pairs, including the original pre-closing market betting price originally collected and the final market closing price now added to the files. A ratio of pre-closing to closing price has been calculated for each pair, and these have been analysed for actual returns to level stakes. The following chart shows how this ratio, providing a measure for the expected yield, correlates with actual returns.
Given the strength of the correlation and its almost perfect 1:1 relationship between expected and observed yields (the graph slope is almost exactly 1.00), it is clear that not only are the closing odds highly efficient, but that the amount by which a pre-closing-market price beats the closing price provides an excellent means to predict the size of expected value. In this analysis, Pinnacle.com's betting margin has not been removed. For this reason, the expected yields are a few percent higher than observed yields, the difference being a measure of the bookmaker's margin.
Football-Data has chosen to make Pinnacle.com's closing odds available precisely because it will allow bettors to analyse whether their betting performance is beating that market consistently. By comparing the odds you bet to Pinnacle.com's closing odds, you will be able to determine whether you genuinely hold an edge over the market through skill, instead of wondering whether your profits and losses are the result of good and bad fortune.