thanks, nbates. i am familiar with those papers.
best,
surfer
trend following delusion shattered
- Thread starter hank rollins
- Start date
- Status
thanks, nbates. i am familiar with those papers.
best,
surfer
that paper on trend following stocks is a difficult thing in my opinion. i did not get the feeling that survivorship bias was a true concern ... my opinion is that standard CTA technology does not work on stocks.
in addition something to the thread as such. i think one must distinguish between trendiness and fat tails. the point of this distinction is that on a single market the trendiness is to weak to bet on, but the once-in-a-while-happening-fat-tail-concept is good for a bet ... if you increase the likelihood of a fat tail by trading many markets.
well, maybe it's just semantics ...
peace
in addition something to the thread as such. i think one must distinguish between trendiness and fat tails. the point of this distinction is that on a single market the trendiness is to weak to bet on, but the once-in-a-while-happening-fat-tail-concept is good for a bet ... if you increase the likelihood of a fat tail by trading many markets.
well, maybe it's just semantics ...
peace
Someone said that they have NEVER seen any prove that the markets are trending, but I think they really didn't look hard enough or simply refuse to look at the subject objectively.
I suggest doing googling on Hurst exponent, variance ratio, or other statistical tools for studying brownian motion and random walk.
A hurst expoent of 0.5 suggest a random walk, lower than 0.5 suggest mean reversion. In the book( forgot the title I read), the S&P has a hurst expnent of 0.67, suggesting it's persistent, meaning that the variance of the market expands at a rate faster than suggested by random walk, now, it doesn't say that the markets are "trending" to be exact, but trends or trend-following is just trader 's terminology.
Look at LME copper or HG, ANY statsical test will tell you that it's not a random walk, let alone mean-reverting.
you don't need to be a stats guy to understand that. just look at the chart. I don't know what else to call it but ..trending...
I suggest doing googling on Hurst exponent, variance ratio, or other statistical tools for studying brownian motion and random walk.
A hurst expoent of 0.5 suggest a random walk, lower than 0.5 suggest mean reversion. In the book( forgot the title I read), the S&P has a hurst expnent of 0.67, suggesting it's persistent, meaning that the variance of the market expands at a rate faster than suggested by random walk, now, it doesn't say that the markets are "trending" to be exact, but trends or trend-following is just trader 's terminology.
Look at LME copper or HG, ANY statsical test will tell you that it's not a random walk, let alone mean-reverting.
you don't need to be a stats guy to understand that. just look at the chart. I don't know what else to call it but ..trending...
Of course markets can be trending.
The problem is with what is meant by trending.
Rephrasing your above paragraph one could say that nobody has EVER been able to define 'trend' in a way acceptable or useful - i.e. making money - to another guy.
ecritt is the author of the trend following on stocks paper, he may have forgotten to mention that.
I tested it myself and found using basic rules like he did, that the drawdown was higher than suggested. I contacted him and then
determined that certain proprietary details were missing , and in my mind key to pursuing such a strategy- the position sizing on entry is dynamic and not static. That changes the results significantly. I have not found improvement changing the position sizing via scaling in, but apparently he has. They are smart and Tom Basso is an investor, of Market Wizards fame.
The work is legitimate and easy to code if you have a suitable portfolio level backtester. Doing the scale in modifications is beyond my current capabilities however, but I did develop my own trend following system that worked well without scale in.
Eric Paradis
I tested it myself and found using basic rules like he did, that the drawdown was higher than suggested. I contacted him and then
determined that certain proprietary details were missing , and in my mind key to pursuing such a strategy- the position sizing on entry is dynamic and not static. That changes the results significantly. I have not found improvement changing the position sizing via scaling in, but apparently he has. They are smart and Tom Basso is an investor, of Market Wizards fame.
The work is legitimate and easy to code if you have a suitable portfolio level backtester. Doing the scale in modifications is beyond my current capabilities however, but I did develop my own trend following system that worked well without scale in.
Eric Paradis
for people who believe you can't make money with trends, even more asinine, that you can't identify a trend............put your money in someone else's hands.........that is as far off the reality base as it gets, seriously............wow, unbelievable.......stop trading now.........your money is ducksoup..........
Counter-trend trading is OFTEN, by definition, simply getting in on a trend before it manifests itself - ie, fading into a position in anticipation of a retracement or reversal.
I love it when people refer to it as trying to catch a falling knife - not at all the case. The reason countertrend trading works is that entries are safe - ie, momentum of the current trend is attenuating. The counter-trend trader often gets out at breakeven when he is wrong, whereas the trend trader gets caught in reversals.
When I follow the trend, I define a siginficant level (often a pivot) and get long above it and short below it and I don't allow myself to get shaken off.
I love it when people refer to it as trying to catch a falling knife - not at all the case. The reason countertrend trading works is that entries are safe - ie, momentum of the current trend is attenuating. The counter-trend trader often gets out at breakeven when he is wrong, whereas the trend trader gets caught in reversals.
When I follow the trend, I define a siginficant level (often a pivot) and get long above it and short below it and I don't allow myself to get shaken off.
- Status