Niederhoffer thinks "follow the trend" sucks

Anyone remember VN's posts on his dailyspeculations blog around 2006 and 2007? He posted weekly articles making fun of the perma bears, how bear markets only happen rarely, how the optimists always win, how history teaches us to buy after 3 consecutive down days.

He's gloating, he loves to polarize, he's an arrogant gambler at heart. And then he blew up again in the summer of 2007.
 
Drift is obviously a second order trend. But I'm not sure why we're even discussing it as trend or counter trend given that Vic or others are not selling 10y options on anything. It would be like saying one should not short the market because its a counter trend trade due to drift...
That's an important point. I'm reminded of what I read about Keynes's early days as a currency speculator. When he first started, he based his leveraged currency trades on his long term views of the respective economies of the currencies in question. Initially his trades made good money, but then he quickly went into the red. That was what inspired him to remark that the markets can remain irrational longer than you can stay solvent.

And so Keynes smartened up and began employing shorter term criteria for his shorter term plays. He went on to become a successful shorter term speculator, as well as a long term investor ala Buffett in terms of approach. In fact, Buffett has acknowledged Keynes’s influence on his thinking. In 1991 he said Keynes was a man, “whose brilliance as a practicing investor matched his brilliance in thought.” (You'll have to forgive my plug for my favorite economist.)

EDIT: Time horizons are an important consideration not only in trading. If memory serves, the S&L crisis of the '80s was attributed in large part to mismatched funding. They funded long-term assets (loans) with short-term money (deposits). It worked great until it didn't. When short-term rates went through the roof, the S&Ls' bottom lines went through the floor.

And so, to the extent that he heavily relied on drift, I think you correctly assessed VN's mistake of placing too much emphasis on the expected long term while he was highly leveraged and vulnerable to a potentially uncooperative short term -- the potholes along the way.
 
Last edited:
...He's gloating, he loves to polarize, he's an arrogant gambler at heart. And then he blew up again in the summer of 2007.
Arrogance sets a high bar. If he were more humble, I think we'd all be at least a bit more forgiving. Maybe. But he set his own stage.
 
"In general it is not time per se but nonlinearity to volatility that counts, so a six month out of the money (measuring out of the moneyness in low delta) is preferable to a 1 year OTM with higher delta, for squeezes. Niederhoffer went bust with short term options... they went from .2 to $38 ! The advantage of shorter term is that vol explodes the most in them. Short but of course not too short." -Taleb

http://www.reddit.com/r/options/com...sim_nicholas_taleb_ask_me_anything_on/crwlrog
 
"In general it is not time per se but nonlinearity to volatility that counts, so a six month out of the money (measuring out of the moneyness in low delta) is preferable to a 1 year OTM with higher delta, for squeezes. Niederhoffer went bust with short term options... they went from .2 to $38 ! The advantage of shorter term is that vol explodes the most in them. Short but of course not too short." -Taleb

http://www.reddit.com/r/options/comments/38onec/i_am_nassim_nicholas_taleb_ask_me_anything_on/crwlrog
Okay. I confess that I know next to nothing about options. So if my comments did not quite fit the scenario you'll have to forgive me. It was the idea of a doggedly-held leveraged short-term trade made on the basis of a long-term premise (drift) that got me going on my diatribe.
 
Last edited:
Wow, that's a good source when it comes to definining statistical terms. I'll have to remember it.

This may help you and the other math illiterate among us---


A series with drift can be modeled as yt=c+ϕyt−1+ϵt where c is the drift(constant), and ϕ=1

A series with trend can be modeled as yt=c+δt+ϕyt−1+ϵt where c is the drift(constant), δt is the deteministic time trend and ϕ=1




http://stats.stackexchange.com/ques...tween-series-with-drift-and-series-with-trend

Any questions about why a drift is not a trend

Peace

Surf

I don't need to discuss definitions like you do. I know for myself what drift and trend is. My bankaccount confirms me every day again that I know what I talk about. So I don't need to discuss every . or , in every word.

I am glad I could offer you a good link. I will keep you busy in symantics. As a writer it is important for you and for the other nicknames you use here.
 
Back
Top