And yet most trend following firms have either shut down or posted very poor returns.
Jim Simons Trashes Trend Following
- Thread starter marketsurfer
- Start date
And yet most trend following firms have either shut down or posted very poor returns.
If I know a certain coin is biased to come up tails more often than heads and I start betting on coinflips using that coin, betting tails every time, will I not be profitable in the long-run despite many losses along the way?
That's not prediction, that's just using deductive logic. I'm not trying to predict which side will come up on which trial, I just bet on the most likely outcome every time.
It's exactly the same with trend following: a market trend has a higher probability of continuation than of reversal. Nothing is certain, just play the odds. What kind of idiot can't do that?
Finding and trading trends is easy. The challenge is avoiding chop or minimizing the damage if you do get caught in chop.
Last edited:
Yet something going up or down can also be modeled after geometric brownian motion, which is no better than zero-expected-value gambling (before execution costs) -- in other words, completely worthless as a trading or investing tool.
Again, I am not saying that "trend following" is worthless, but the fact that hardly anyone (if anyone at all) on planet Earth was able to time (even in a very rough sense) the four trades implied by your examples demonstrates that if there is a "trend" to follow as an anticipatory signal, it is by no means easy to do.
"Trends" look seductively obvious in hindsight to the human brain. Whether they enable you to earn money trading is an almost entirely separate issue.
But there's a (potentially money-losing) problem with your analogy: financial instruments are nothing like a biased coin, whose odds are known and simple. If you're using "odds" in trading, you are implicitly predicting the odds.
If you want "odds" that are known, you're better off heading to your local casino; the market will give you no such gifts.
Yet the "chop" and the "trends" are not easily separable in securities price series, even in hindsight. Again, how does one know that they aren't just lunging at brownian motion here, being "fooled by randomness"? I think the answer here is on the individual trader level and comes down to whether a particular trader using the relevant technique is beating the indices with relative consistency and by a margin significant enough to justify the risk; if not, they're probably wasting their time (and money) and would be better served pursuing something else.
Who said anything about knowing the odds on the coin? That's irrelevant. If I know tails is more likely than heads, regardless if the edge is big or tiny, it still makes sense to bet on the most likely outcome. If I want to optimally size my bet, sure I need to know the correct probability but in terms of a straight-up bet, I just need to know the most likely outcome.
It comes down to knowing how to find trends and only trends, which admittedly most TF traders can't do. And that includes the professionals. You can find trends with a moving average crossover. But that crude tool will also find you chop, which you don't want. Finding trends while avoiding chop or minimizing damage from chop isn't easy but it's doable. Don't ask how, nobody is going to hand you such a wonderful gift. Remember: even most professionals don't know that stuff.
Clearly you are not aware of the shear brilliance of Jim Simons. His intellect goes far beyond the markets--- its best to learn from him rather than think u can chalkenge him in ANY topic. Peace, Surf
hmmm, try reducing size and widening stops everytime you get stopped out. I just got chopped to death on eur.usd, now I am sitting with a stop to enter a small position long way up there and another one to go short way down there, otherwise I am flat.
and conversely, adding size and tightening stops when you are not stopped out
you know it's going good when your size is maxed out and your stops are irrelevent