Saturday, March 17, 2007

Spread Betting

Paton and Williams from Nottingham have investigated financial spread betting in ‘Quarbs’ and Efficiency in Spread Betting: can you beat the book?. They

examine a relatively novel form of gambling, index (or spread) betting, that mirrors (and indeed overlaps with) practices in conventional financial markets. In this form of betting, a number of bookmakers quote a bid-offer spread about the result of some future event, and bettors are invited to buy (sell) at the top (bottom) end of the quoted spreads.

This example explains betting in detail.

For example, in a cricket game between England and Australia, the bookmaker might set spread for runs in England’s first innings of 240-250. A bettor who believes England’s batting is weak may sell total runs at the price of 240 on a stake of, say, £5 per run. If England score 215 runs, this is the termination value of the asset and the bettor will win £125, calculated as the difference between the value and the price (240 - 215) times the stake (£5). On the other hand, if England score 290 runs in the game, the bettor would lose £250, calculated as the difference between 290 and 240 times the £5 stake.
They test two strategies:

We consider two alternative predictors of the actual value of the asset: (i) the average midpoint of the market bid-offer spreads (MID); (ii) the mid-point of the outlying bid-offer spread (OUTLIER).
And find:

Taken together, the evidence appears to be conclusive in suggesting that the market mid-point is systematically superior to the outlier in predicting actual assetvalues.
They also find

Using the notion of quasi-arbitrages or Quarbs, we find that it is possible to devise a trading strategy on the basis of the outlying spread that yields returns, both within and out of sample, that are consistently positive and superior to those that might be expected from noise trading.

No comments: