Nov

2

50 200 MAI don't want to touch off a quant vs tech battle, but was interested in thoughts about the use of moving averages. The claims are seen so often in the media for example as 50m/a has crossed the 200m/a as bullish or bearish. As quant this is easily testable, but is there some fallacy built into the assumption, checking once premise. As an average of levels over some period is this overly sensitive to a large event that may have happened at the beginning of the period, or as new data are added and taken away from 200 days ago. This seems a drawback. So on an unchanged day the 50m/a could cross the 200m/a, so it is a lagging measure. I have not used them much but interested in testing and what the merits might be if any. I am prepared to buy a round if venturing too far from our charter.

Craig Mee comments:

Normally people use exponential moving averages which put more weight on the most recent numbers, however it must be noted when looking at technicals from any of the major houses, one day they may be looking at the 50 day exp m/a, the next day the 20 day exp m/a, next day 16 and so on. There seems to be little uniformity and who's to say the market's in a 50 day cycle or 20 or whatever, no doubt this will change at any rate, it must be proven, as to the reason it's used , and this is never done. 200 day just constitutes a longer term trend number. Could be 167 or 198 for the purpose. They may have more of a chance to perform some sort of return in a slower moving stock or short-end rate market, but for futures markets, there is way too much given away on the turn (i.e. consolidating flat markets forever paying away the spread to get in and out.) It's much the same in running a moving average over your profit and loss on a trading system as a filter , normally what you'll find is that the major gains are made on the swing, for that's when markets are extending, and if you're not in then you're not in!

Victor Niederhoffer comments :

Not to mention that it is extremely spurious and dangerous to your wallet to see cycles in moving averages.This is due to the Slutsky-Yule effect, a consequence of regression to the mean, creating false cycles.

Anatoly Veltman suggests:

The best M/A use I know of has nothing to do with crossovers, or even with price charts' current levels vis-a-vis averages. It has to do with current slope of all averages in an optimized pitch-fork of three averages. The method was originally described by Stan Weinstein, as applied to longer-term stock strategies. I’ve had positive results applying the general overriding idea; and over the years, I’ve worked with people who did optimization and chose a pitch-fork of 14-, 30- and 50- day pivot simple M/As. Furthermore, a few effective signals were developed: dubbed Moving-Average Fake-Out Trade and Moving-Average-Divergence Trade, at the time…


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search