adnormal spread ratios trading idea

[cwest@xxxxxxxxxxxx]

Here's another trading idea for anyone to critique, exploiting abnormal
spread ratios :-) .

The idea is to scan for spreads that are very much out of norm. Using IBM
and MSFT as an example, calculate a spread ratio by using the relative beta
of their respective volatilities: 

		relbeta=IFF(VolatilityStdDev(Length)of data1<>0,
			VolatilityStdDev(Length)of
data1/VolatilityStdDev(Length)of data2,1);

		spreadratio=IFF(close of data2<>0,
			relbeta*close of data1/close of data2,1);
			
		spread=close of data1-(spreadratio*close of data2);

Looking at a gif at www.traders2traders.com/myfiles/cwest/spreadratio.gif
<http://www.traders2traders.com/myfiles/cwest/spreadratio.gif>  you can see
that the adjusted or ratio'd spread diverges, sometimes extremely. When an
extreme spread occurs there are a few trading choices. My preference in this
example would be to use an inter-security option spread. When the spread
narrows I'd sell IBM calls and buy MSFT calls, choosing ATM strikes.
Sometimes naked calls are an issue, but they're very coverable with stocks
like IBM and MSFT.

TS doesn't back test this kind of strategy efficiently. It's necessary to
connect 2 instances of the strategy using a DLL, and then consolidate the
results. 

Colin West