When I’ve spoken to traders about the use of correlation in the past, I’ve got quite a few shrugs of the shoulders, and the occasional ‘that’s something to do with statistics isn’t it?’ So it’s nice to know I raised some interest. True, it is used, as you mention, ‘to get a sense of what stocks are correlated with other stocks’ but that’s only the beginning, there are in fact several strategies that follow from correlational analysis and that are used by a small number of adherents. For example, pairs trading (though this seems to be used much more commonly in the Forex market, though not exclusively). Here’s one: take two stocks that are highly correlated, and if news or something moves one, you buy the other. That’s a very simple strategy, of course, but under the right conditions it can be effective. So as far as whether correlation should be part of the trader’s toolkit, my feeling is that it should be.

However, we still need to know some of these conditions – as a scientist, it bugs me that so much of technical analysis has not been scrutinized methodically: we need to know, for example, things like how the correlation changes for different time frames or periods, how is the corr coeficient related to the holding period, is negative corr just as useful as positive corr, and so on…

To begin, it really depends on how much is possible on your site and that can lead into further analysis. Time frames of, say, 365 days, and 250 days, are useful but more importantly perhaps, the shorter time periods of 100 days, 30 days, and 7 days really need to be looked at.

Not having index data is a problem, as sometimes it would be useful to look at these correlations in the context of the market. Is it not possible to use another source? What about Yahoo! they have a lot of available historic and recent data (I know that it can be downloaded into Excel, so maybe it’s downloadable into something else!). We don’t have to have up- to-the-minute data for this – end of day data or data a day or two old would work, I think, just as well.