To reduce the overall number of transactions when you use dynamic weight, I would like to find out how volatile a dynamic weight factor is.
Lets say I weight after 1/RankPos, or 1/ MktCap or 1/AvgVol(91), but I wanted to find out, for instance, how volatile a RankPos is, like how often does a top rankpos fall near the sell limit (which is 50 in my system) and then go up again?
Is there any way of measuring this over a period of 3–6 months?
The question involves not only the volatility of the weight factor, but primarily the volatility of your ranking system. A ranking system based only on annual reports, for example, will have almost no turnover, while one based only on price data will have enormous turnover.
I have found the best way to measure the volatility of a ranking system is to simply create an equal-weight simulation with a very small number of highly ranked stocks (say five or ten stocks with a very tight rank-based sell rule) and then look at the turnover. Compare that turnover with various ranking systems and you'll get a good sense of how volatile each one is.
Yuval, what do you mean by "very right ranked-based sell rule"? What do you consider very tight?
Tony
Either the same number of stocks as you're holding or a few more. E.g. for holding ten stocks a sell rule like rankpos > 12.