Trend detection

[Handle123]

What are good methods for regime detection?

I'm looking for a method to detect the long term trend in S&P500 with good accuracy.

My Long Term Commodity System, right now I am averaging 12% winning trades for past 12 months, most would say I am losing my rump as all opening signals are against major trends, and add on signals are already where the herd are saying trend has changed. Now trend is allusive as you can only understand that the trend really is recent or distant past and most hopes it will continue, except me. I have learned to make long term money, you have to view trading as if you were a commercial entity, you don't see too many of them going belly up. I hedge all my trades and that is cause unlike the Commercials, I don't have or want the product in which they can hedge, so I have to use options.

There is no "perfect" moving average, indicator or charting patterns in which it will call tops/bottoms. If there were, no one would be trading as no one would take other side of trade.

Many have used 4/9/18 for years on monthly charts, when all are 4>9>18 trend is up and other way trend is down, when mixed up is often chop or trying to start new trend.

Nice thing about Monthly charts is seldom in uptrends you have three lower closes or three higher closes in downtrend. Matter of fact I get all giddy when I see two closes in a row counter to current trend, not going to bet the farm, but am looking for good opportunities in stocks, futures and options to trade back into the trend.

TaDaaaaaa

 

[Stockmarketmap]

A pure price cross heuristic above and below the 10 period monthly basis moving average applied to the S&P500 and long bond market has produced risk adjusted alpha over buy and hold over a 30 year sample. Using the Nasdaq 100 / QQQ * as the equity proxy is a decent way to increase and exploit the returns from the mid cap growth stock universe.

Using shorter basis such as the popular "daily" 200 period moving average produces whipsaws that can be counter productive in: 1) transaction costs 2) monitoring time and 3) behavioral affects. The monthly time frame alleviates these deficiencies as it contains the whipsaw signaling produced by the short basis = less transactions with cleaner signaling. It has also more or less accurately reflected a general economic cycle length in the markets tested over large test samples ( U.S. and Japan ** ).

It may not be as "sexy" as a trading regime requiring more frequent transactions, but it has produced decent returns for the effort expended.

* https://docs.google.com/document/d/1XwZjcWy7KlSwA7xi0rax7nevIBCtW0Uu4UZFH-Hc1ns/edit?usp=sharing
** https://docs.google.com/presentation/d/1Sn6BKRCKRU5tensBDFTkJXI3v2wRQ4M1bt8VoIM2Zmc/edit?usp=sharing

additional research http://www.advisorperspectives.com/...ook-back-at-the-performance-of-the-holy-grail

James
Director of Quantitative Research
XOXOX
Boulder, CO