odds and money traded

[Joe]

Why is it that a disproportionately large amount of money goes on a favourite, compared to that on a longshot? For example, suppose we had odds in a fair match-up of 1.5 and 3. For the fair bookie to break even whatever the result, he would be hoping that if £3000 went on the 1.5, £1500 would go on the 3. From a a little bit of data collection from Betfair, it seems that the 1.5 attracts considerably more than £3000 and the 3 considerably less than £1500. The simplest answer is that fewer punters back the longshot. However, if everyone was backing at level stakes, there should be no disproportion, provided that there were 3000 punters at 1.5 for every 1500 punters at 3. The next simplest answer is longshot backers stake less money than favourites backers. In the extreme, if each stakes money to win the same amount, then a high priced longshot would attract disproportionately a lot less cash. What strikes me as rather odd about this is that, whilst a longshot punter is reducing his risk level by reducing his outlay, he is also reducing his long term profits (because he won't win as often). Which then begs the question, why bother with the longshots, if you can keep you long term profits higher without the added risk by dumping the longshots in favour of the favourites? Anyway, without access to bookies' data sets on punters betting habits a lot of this will remain speculation. For now I'll show below what I've collected this morning on tennis match bets for Hamburg 2nd Round. I've only bothered to show the data for the favourites (the books are all virtually full term and about fair). First column are the odds, second column the expectancy based on the odds, third column the expectancy based on the money traded if we assume that punters knew what the real true odds were and were trading proportionally. 1.16 86% 94% 1.26 79% 95% 1.70 59% 82% 1.74 57% 85% 1.75 57% 91% 1.78 56% 73% 1.79 56% 66% In all cases, there's a disproportionate volume on the favourite, and hence if we were to assume that the only reason for this was the punters were more savey about true odds, in all cases the favourites should have been priced considerably lower. Clearly this can't be the case, but for reasons already mentioned, I won't carrying on speculating as as to why this deviation exists. [As an aside at this point, it migth be worth wondering whether this sort of data could be used in anyway for a betting system. One of the rows above deviates much more markedly than the rest. This was for Ferrer to beat Hrbaty. It's interesting to note here that whilst the average price (based on volumes traded) was 1.75, this had fallen to 1.67 by the time of the start of the match. Obviously hindsight is a wonderful thing, but with so little going on Hrbaty, this might indicate that Ferrer is a good bet.] Anyway, back to the main point. If punters are putting disproportionately more money on the favourites, why aren't the bookies dropping the prices to manage their liabilities? Of course, Betfair couldn't give a monkeys how much is traded on either player, and bookies' prices will also have the overround built in to protect them somewhat, but in general I would guess that the volume proportions traded at Betfair will be reflected in those with the bookies. And it simply is not the case that a bookie will drop his price for Ferrer to 1.2. It might drop a little but not that much. Doing the maths, then, we can see that in all cases in the table above, should the favourite win, the bookie will be down on the deal somewhat, up on the deal more significantly if the longshot wins. Over the long term of course the bookie will not lose (and the overround will generate him the profit) However, the longshots win less frequently, so the bookie has to suffer a series of smaller losses, and recoup those with fewer bigger wins. Now to me this it the complete opposite of what a risk averse punter backing favourites will be doing. Yet I thought that bookies were risk averse. And it's all the more puzzling when one considers the presence of the favourite-longshot bias, where the prices on the favourites are disproportionately higher than they should be on the basis of the overround alone. Of course, I've attempted to explain the bias in terms of the influence that slight errors a bookie makes in estimating the true chances of a result will manifest themselves more damagingly (in terms of the bookie's experience) in the longshot. It has also been explained away by the presense of insider traders, who may come in with a disproportionately large stake on the outsider, and similarly by less savy punters who are lured by large payouts, despite the lower chances. But regardless of all this, if the data from Betfair is replicated more widely over all betting books and bookies at large, this makes for a puzzling conundrum. Ideas and comments appreciated.

[Joe]

Re: odds and money traded For the record, Ferrer won 6-0, 6-3 :D