 |
TSGannGalt
Registered: Jun 2004
Posts: 1980 |
12-09-09 06:58 AM
Quote from Trend Following:
Are you going to post the list of systematic trend followers who are taking massive discretionary bets who will soon be going the route of Amaranth?
OK... This is beyond my comprehension (and it's getting scary...) It's either, I'm very tired or Mike's on something... Wait... Is this another one of those nit-picking phrases and taking things out of context?
I don't get it... So...
Can someone else tell me how my series of posts in this thread leads to me posting a list of funds?
|
| |
|
Edit/Delete Quote Complain |
AFJ Garner
Registered: Mar 2008
Posts: 98 |
12-09-09 12:04 PM
Quote from TSGannGalt:
Can someone else tell me how my series of posts in this thread leads to me posting a list of funds?
Intelligent readers of this thread (incredibly, there seem to be few) will have taken your points on board. On occasional recent visits to his multiple sales venues, I note that Trend Following has taken to naming any view that does not coincide with his own 100% bollocks.
I tactfully suggest that there is no point trying to argue with such a person. There is no common ground.
|
| |
|
Edit/Delete Quote Complain |
ssb11
Registered: Nov 2009
Posts: 14 |
12-09-09 02:15 PM
we trade a mechanical simple trend following system and have done so for the past 10+ years. by mechanical i mean that each entry, exit, and size decision is dictated by rules that have been created before the trade is entered. we do not use any mean reversion. we do not view markets as changing. they have and always have had an infinite variance. past history only shows us a tiny sliver of the possible market outcomes.
any winning model must have some predictive value. random entry and exit have an expectancy of 0 less costs. that model may be predictive in the distribution of moves rather than the path of a given market. however, you must have a w% > 1/1+(avW/avL). if my after cost av W is 3 units and av L is 1 unit, then i must win >25% of my trades. random trading would win 25%. if av W is 1 and av l is 3 i must win >75%. the importance of costs and time frame, which are part of the same continuum, can not be overstated in my opinion.
i think of profitable trading as being either frequency or magnitude. i am either going to win more or larger. i would suggest that outside of market making the gross majority of long term profitable traders have done so through magnitude.
|
| |
|
Edit/Delete Quote Complain |
AFJ Garner
Registered: Mar 2008
Posts: 98 |
12-09-09 08:20 PM
Quote from ssb11:
we trade a mechanical simple trend following system and have done so for the past 10+ years. by mechanical i mean that each entry, exit, and size decision is dictated by rules that have been created before the trade is entered. we do not use any mean reversion. we do not view markets as changing. they have and always have had an infinite variance. past history only shows us a tiny sliver of the possible market outcomes.
Yes, each entry ,exit and size decision is dictated by rules that have been created before that trade was entered. A good definition of mechanical. But this does not preclude, of course, the changing of rules in between trades. Which certainly seems to be what most CTAs have done periodically over the years. I can understand the reason you do not view markets as changing within the infinite time frame you are considering you may well see different market conditions come and go around and around over the millennia. Good point. Nevertheless, in a more human timeframe (and more particularly within the timeframe of a human working life) it may well appear that markets change (on the assumption that one agrees with your viewpoint on the very long term and no reason not to we can hardly foresee what an infinite future might bring). In which case, there is a very good argument to be made that trend following systems need adjustment periodically to keep performance positive. Back testing shows that the Original Turtle Rules went flat to negative in and from the 1990s for instance. And that the addition of a longer term filter would have restored some profitability. Yes, I can create a set of rules which would have traded profitably over the past 50 or 100 years with hindsight but I am not at all sure that such rules would remain profitable over the next 50 or 100 years. You may take a different viewpoint?
Quote from ssb11:
i think of profitable trading as being either frequency or magnitude. i am either going to win more or larger. i would suggest that outside of market making the gross majority of long term profitable traders have done so through magnitude.
I do not have a wide enough view of other trading methods to either comment or to categorise profitable trading in this way. I can of agree that the two extremes you point out are indeed sources of profit. But I have my doubts that the matter is either this simple or this black and white. Of course, I admit my ignorance on the point the joy of discussion is to learn from others and to listen to their views.
|
| |
|
Edit/Delete Quote Complain |
TSGannGalt
Registered: Jun 2004
Posts: 1980 |
12-10-09 02:28 AM
Quote from AFJ Garner:
Yes, each entry ,exit and size decision is dictated by rules that have been created before that trade was entered. A good definition of mechanical. But this does not preclude, of course, the changing of rules in between trades. Which certainly seems to be what most CTAs have done periodically over the years. I can understand the reason you do not view markets as changing Ewithin the infinite time frame you are considering you may well see different market conditions come and go around and around over the millennia. Good point. Nevertheless, in a more human timeframe (and more particularly within the timeframe of a human working life) it may well appearEthat markets change (on the assumption that one agrees with your viewpoint on the very long term Eand no reason not to Ewe can hardly foresee what an infinite future might bring). In which case, there is a very good argument to be made that trend following systems need adjustment periodically to keep performance positive. Back testing shows that the Original Turtle Rules went flat to negative in and from the 1990s for instance. And that the addition of a longer term filter would have restored some profitability. Yes, I can create a set of rules which would have traded profitably over the past 50 or 100 years with hindsight but I am not at all sure that such rules would remain profitable over the next 50 or 100 years. You may take a different viewpoint?
I do not have a wide enough view of other trading methods to either comment or to categorise profitable trading in this way. I can of agree that the two extremesEyou point out are indeed sources of profit. But I have my doubts that the matter is either this simple or this black and white. Of course, I admit my ignorance on the point Ethe joy of discussion is to learn from others and to listen to their views.
What ssb is saying is the same things as "The fact that markets change(has infinite variance), does not change". So we're talking about 2 different things. He uses the phrase "market variance" as to what we refer to "market change". This is just semantical (again), and there's really no point of starting a discussion about this. I don't see how we can...
Eventually, if you continue to accept this debate, it will end up with "Markets are Random" debate. Which is also a different topic from whether the markets change. Let's say it does, get into this debate... then how does something random change, if there is nothing consistent to measure between 1 state to another (change).
Back to square 1.
Anyways...
All this is typical between Discretionary and Mechanical traders... Subjective semantics causing misunderstanding...
|
| |
|
Edit/Delete Quote Complain |
TSGannGalt
Registered: Jun 2004
Posts: 1980 |
12-10-09 02:41 AM
Quote from ssb11:
we trade a mechanical simple trend following system and have done so for the past 10+ years. by mechanical i mean that each entry, exit, and size decision is dictated by rules that have been created before the trade is entered. we do not use any mean reversion. we do not view markets as changing. they have and always have had an infinite variance. past history only shows us a tiny sliver of the possible market outcomes.
any winning model must have some predictive value. random entry and exit have an expectancy of 0 less costs. that model may be predictive in the distribution of moves rather than the path of a given market. however, you must have a w% > 1/1+(avW/avL). if my after cost av W is 3 units and av L is 1 unit, then i must win >25% of my trades. random trading would win 25%. if av W is 1 and av l is 3 i must win >75%. the importance of costs and time frame, which are part of the same continuum, can not be overstated in my opinion.
i think of profitable trading as being either frequency or magnitude. i am either going to win more or larger. i would suggest that outside of market making the gross majority of long term profitable traders have done so through magnitude.
As much I hope I understand your post... I'll be a bit stingy...
If past history only shows us a tiny sliver of the possible market outcomes... then, what was the basis of your decision to trade your trend following system? If what you mention is taken literally... trend-following can just be equally a noise/outlier that is completely against how the market is. If I sound like a flamer... isn't what you're saying close to saying that you make trades blind-sightedly. If what we observe is only a tiny pinch of how the market is. And if the markets have an infinite variance... then how can anyone conclude anything about the market? There's going to be a conflict of philosophy between how you trade and how you view the market.
I know... I'm being very picky... But you provided 2 ironies, in one post...
No offense or harm intended. Just curious.
|
| |
|
Edit/Delete Quote Complain |
| Receive
an email whenever a new post is added to this thread by subscribing
to it. |
|
|
|
|
