Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : Buy and Sell Signals, and Other Market Perspectives -- Ignore unavailable to you. Want to Upgrade?

To: SGJ who wrote (29882)	3/15/2012 12:37:56 AM
From: FCom777	1 Recommendation Respond to of 224215

Sorry for not answering sooner but I've been pretty busy with the day job today ... You ask a good question about the correlation percentage. I actually started developing an Inter-market Explorer module to my financial analyses package and have been wrestling with the best and/or different ways to calculate scalar values that represent how different markets correlate to each other. It would be useful to have a matrix of values for a particular market/stock/future that indicate how it correlates to the key major markets (i.e. bonds, equities, oil, sector, etc.). Its not as straightforward as one might think. An obvious way would be to specify the time interval (i.e. hourly, daily, weekly) and statistically calculate percentage changes and do a summation type index over whatever target time interval one is considering. However, I see certain pitfalls with this approach. For example, what if one market moves up a lot one day and another moves down a little that day - but then the next day the first market moves down a little and the second moves up a lot. The two markets could be very correlated but statistically that behavior wouldn't show good correlation. Another approach would be to use moving averages of x intervals to dampen the effect of a particular interval and calculate statistics on that basis. Another approach would be to place a greater emphasis on time. Determine dates/times associated with ringed highs/lows and compare different markets in terms of when key turning points occur. I could easily implement different algorithms since I already have data for most all the markets but haven't really settled on a particular approach. I'm real open to any ideas you (or anyone else) might have in this regard. If anyone is aware of a good technical discussion of how this might be accomplished, that would be real useful and I'd be happy to share the results ... For now, I just plot it visually and evaluate the data. So, no I really don't have a specific correlation percentage per se ....