Stretching The Band

On the subject of Draw seeking, Scott writes:

Have you considered the merits of allowing for the backing of a higher priced draw above 5.0 in your XX selection process?

Maybe it's just luck, but Green Pullover has really boosted his ROI with two Barcelona draws at prices of 6.50 and 9.00.

Is it something worth considering when a short priced favourite offers no value?

The XX Draws are selected based purely on ratings and as you might suspect from a system with a 34.5% (implied price 2.9) strike rate where the average price is 3.48 (implied probability 28.7%), price is not typically an issue. It would be nice if the selections were all 4.1 (the maximum that I have yet seen) but so long as they are above 2.9, we have a value selection, and unless we're in a late season Serie A game, the draw price will always be well above 2.9.

The XX Draws are those games where the expected superiority of either team is 0.1 or 0. This is admittedly a very narrow and thus selective band, but that's why it is so profitable. If I widened the band, there would be more selections of course, but the quality would start to fall. If the band were widened to include the 0.2 to -0.2 selections, the draw price needed to be profitable would rise from 2.9 to 3.33. Interestingly, if we forget all about the crazy Bundesliga once we move beyond the 0.1 band, the results are better. The 0.5 band (0.5 to -0.5), needs a draw price of just 3.15 to be profitable, 249 winners from 785 selections. Excluding the Italian games would also boost the numbers, but that's another story.

The problem with games such as Barcelona's at Athletic Bilbao last weekend is the paucity of data. Of 2,355 matches recorded, only 7 have the same profile as Barcelona's game at the weekend. Three games have been draws, which is great, but from a sample so small, it would be reckless to get too carried away. If I include matches close to that rating, I have 8 draws from 24 selections, which is promising to say the least if the draw price is over 6.0. Analysis goes on, as always.