The problem is you only know the market was trendy after the fact.......
How long does a system last?
- Thread starter pwells
- Start date
The GRAB trend following system featured in the book (buys pullbacks on uptrends and sells them on downtrends) has poor results in a long-term backtest.
Last edited:
Your point is only valid for those who do not know how to find a trend.
For those who know a trend, there is no problem.
If it would be a money making system it would not be sold in a book. Published systems are only good to have some general idea which direction you should look. That's the most favourable outcome you can expect.
Pullbacks on trends have one very weak point: when the pullback finally happens, at lot of the potential move can already be gone. You can then in best case try to catch the remainings that are left on the table.
I'm at one year and two months, who knows?
The description of the method is too philosophical, the parameters are prob. set wrong, if it this obviously wrong, just reverse it.
Of course it won't work, this is the opposite of a trend-following system, it buys when prices are "cheap" and sell when they are "expensive".
While it’s true that all data analysis is historical, fighting markets is riskier than being on the same side as liquidity takers.
I mean, those who bought and held have done really well the past 11 years.
Paul Tudor Jones - “The very best money is made at the market turns. Everyone says you get killed trying to pick tops and bottoms and you make all your money by playing the trend in the middle. Well, for 12 years, I have been missing the meat in the middle, but I have made a lot of money at tops and bottoms.”
This is a very good question and one that quants are always pondering.
Most systems go through periods of drawdown - so when a system which was profitable starts to lose is it from a temporary drawdown period or has the system stops working?
For this reason it is important
- even before starting to trade a system which has shown profitability in out-of-sample back-testing-
to set out (market,trade) conditions which would cause it to stop working.
Unfortunately, one cannot always determine why it works and therefore when it might stop - but at least outlining the conditions is a start.
One approach that is used when running several systems in parallel, is to reduce the weight assigned to a system when it is going through drawdown.
E.P. Chan has some good talks on quant system development.