trend following delusion shattered

[hanseng1]

Quote from NickelScalper:

Cue the theory of market efficiency which says, "If there's a method in the public domain, it doesn't work."

The efficient market theory I am familiar with states any new information will be fully and instantly priced into the market, where the market becomes stable until the next market-moving event.

The efficient market hypothesis is actually pure bullshit and is just one of many assumptions economists use when defining models (along with rationality, etc.). The best acamedic work I have seen is "zero-intelligence" modelling by J.D. Farmer (I was initially put off by the name and the pathetic understanding of markets by many academic economists, but it is actually very good), which does not make the assumption of rationality and efficiency.

Making a system public does not necessarily ruin it. Look at Richard Dennis and the Turtles. They all used the same trend following system for years with very good results. Many of them still manage large sums of money with trend following systems.

Also, don't be so quick to simply dismiss trend following success as random occurrance. The exact same could be said about any other form of trading. The generally accepted odds of trading success (usually quoted as 5%) could also be dismissed as simply luck. I would attribute it to long periods of hard work and learning (areas where the other 95% usually fail).

 

[trade-ya1]

Hank, if you are short Oil, you would agree that you are going against the longer-term trend, no? If I am a short in Oil, I am certainly going to cover quicker on a downtick than I ordinarly would in a sideways or falling overall environment, hence perpetuating the trend. Trend must include some timeframe of reference, I agree on that point. Some of our quants don't think about this issue, others utilize trends as an edge. How does one explain the fortunes of trend-followers such as John Henry, Stan Druckenmiller, etc.? Luck? I respect everyone's opinion and I really haven't tested trend-following scientifically. However, I see the market as an independent living organism made up of many independent living cells (similar to the human body made up of many cells, DNA, RNA, etc. each acting independently and with their own agenda) and as such, I believe the market has a distinct personality and the laws of nature and physics can be applied successfully to such an object.

 

[NickelScalper]

Quote from jsp326:

Cue Alexander Elder quoting (paraphrasing?) Warren Buffet on market efficiency:
"Investing in a market in which people believe in efficiency is like playing poker against those who believe it does not pay to look at cards."

What significance do you attribute to that saying in the present discussion?

 

[NickelScalper]

Quote from hanseng1:

The efficient market theory I am familiar with states any new information will be fully and instantly priced into the market, where the market becomes stable until the next market-moving event.

The efficient market hypothesis is actually pure bullshit and is just one of many assumptions economists use when defining models (along with rationality, etc.). The best acamedic work I have seen is "zero-intelligence" modelling by J.D. Farmer (I was initially put off by the name and the pathetic understanding of markets by many academic economists, but it is actually very good), which does not make the assumption of rationality and efficiency.

Making a system public does not necessarily ruin it. Look at Richard Dennis and the Turtles. They all used the same trend following system for years with very good results. Many of them still manage large sums of money with trend following systems.

Also, don't be so quick to simply dismiss trend following success as random occurrance. The exact same could be said about any other form of trading. The generally accepted odds of trading success (usually quoted as 5%) could also be dismissed as simply luck. I would attribute it to long periods of hard work and learning (areas where the other 95% usually fail).

The variety of the principle of market efficiency I have been citing says that all publicly available information and techniques are already in the market price, because there's no way for the individual trader to reliably compete against the resources and access of the collective competition including the bigger players.

Second, the citing of a few past examples which appear to contradict the hypothesis is statistically irrelevant and experimentally unreproducible.

Third, you haven't answered the previously posted challenge, which is reproduced here:

Any false method, including "trend" following, will yield success to some degree, otherwise it would be negatively correlated to the market, which is an edge. Trends exist only in the past, so traders on trend have made money in the past, but only as luck has provided for them.

Cue the theory of market efficiency which says, "If there's a method in the public domain, it doesn't work."

Anyone can prove both of these wrong simply by explaining, in terms that can be used to trade, how any combination of publicly available tools can determine a specific minimum difference between present and near-future price based on past price action.

 

[hanseng1]

Quote from NickelScalper:

Third, you haven't answered the previously posted challenge, which is reproduced here:

Any false method, including "trend" following, will yield success to some degree, otherwise it would be negatively correlated to the market, which is an edge. Trends exist only in the past, so traders on trend have made money in the past, but only as luck has provided for them.

Cue the theory of market efficiency which says, "If there's a method in the public domain, it doesn't work."

Anyone can prove both of these wrong simply by explaining, in terms that can be used to trade, how any combination of publicly available tools can determine a specific minimum difference between present and near-future price based on past price action.

I will produce some information that will confirm the usefulness of trends when I get a chance (maybe in a day or two).

In the mean time, why don't you produce some proof that markets are really efficient? I doubt you will because you can't.

As I said before, the concept of an efficient market is an assumption used in academics. Most of these core principals don't hold perfectly (or at all for some). Take the PPP and IS/LM models; both are valuable concepts but are useable in the real world by themselves.

 

[NickelScalper]

Quote from hanseng1:

I will produce some information that will confirm the usefulness of trends when I get a chance (maybe in a day or two).

In the mean time, why don't you produce some proof that markets are really efficient? I doubt you will because you can't.

As I said before, the concept of an efficient market is an assumption used in academics. Most of these core principals don't hold perfectly (or at all for some). Take the PPP and IS/LM models; both are valuable concepts but are useable in the real world by themselves.

Did you read my post before you replied?

I'm not asking for "information that will confirm the usefulness of trends." There is virtually unlimited apparent or partial confirmation and contradiction on this topic.

That's why I'm cutting to the chase with a precise proposition that any method of trend must satisfy, which is repeated as a reminder:

Any false method, including "trend" following, will yield success to some degree, otherwise it would be negatively correlated to the market, which is an edge. Trends exist only in the past, so traders on trend have made money in the past, but only as luck has provided for them.

Cue the theory of market efficiency which says, "If there's a method in the public domain, it doesn't work."

Anyone can prove both of these wrong simply by explaining, in terms that can be used to trade, how any combination of publicly available tools can determine a specific minimum difference between present and near-future price based on past price action.

 

[hanseng1]

Quote from NickelScalper:

Did you read my post before you replied?

I'm not asking for "information that will confirm the usefulness of trends." There is virtually unlimited apparent or partial confirmation and contradiction on this topic.

That's why I'm cutting to the chase with a precise proposition which any method of trend must satisfy, which is repeated below so we don't drift away from it:

Any false method, including "trend" following, will yield success to some degree, otherwise it would be negatively correlated to the market, which is an edge. Trends exist only in the past, so traders on trend have made money in the past, but only as luck has provided for them.

Cue the theory of market efficiency which says, "If there's a method in the public domain, it doesn't work."

Anyone can prove both of these wrong simply by explaining, in terms that can be used to trade, how any combination of publicly available tools can determine a specific minimum difference between present and near-future price based on past price action.

You are starting to become circular in your responses.

I am going to do a small study on trend "extrapolation" that I am sure will show trends persist into the future. This is in direct relevance to this thread. Don't knock it before you see it.....

On the other hand, I am quite confident you cannot prove markets are efficient.

I have a feeling no matter what I produce as proof, it will be shot down with all kinds of "what-ifs".

As far as public domain systems, Abberation and several systems sold by Mind-Fire Systems (Catscan and others) work well into the future and are publicly available. Many other systems (such as the Turtle 20 day breakout) still work to a degree without any fancy optimization. All of these are publically available and work. I'm sure they are going to be shot down because of drawdowns, sharp ratios, etc. But these are daily trend following systems that do work.

Also, what is a "specific minimum difference between present and near-future price"?

You try to prove trends have no "influence" on future price and I'll try to prove the opposite. If you make claims of this sort but cannot back them up, then why should I?

Edit: Are you asking my to prove the existence of a trend-following "holy grail"? You say trend following has to work to a degree. I call that an edge (perhaps not a good one, but one none the less). You give two things to prove wrong, but they are not clear demands.

 

[hank rollins]

Quote from hanseng1:



I am going to do a small study on trend "extrapolation" that I am sure will show trends persist into the future. This is in direct relevance to this thread. Don't knock it before you see it.....

looking forward to your findings.

 

[NickelScalper]

Quote from hanseng1:

You are starting to become circular in your responses.

I am going to do a small study on trend "extrapolation" that I am sure will show trends persist into the future. This is in direct relevance to this thread. Don't knock it before you see it.....

On the other hand, I am quite confident you cannot prove markets are efficient.

I have a feeling no matter what I produce as proof, it will be shot down with all kinds of "what-ifs".

As far as public domain systems, Abberation and several systems sold by Mind-Fire Systems (Catscan and others) work well into the future and are publicly available. Many other systems (such as the Turtle 20 day breakout) still work to a degree without any fancy optimization. All of these are publically available and work. I'm sure they are going to be shot down because of drawdowns, sharp ratios, etc. But these are daily trend following systems that do work.

Also, what is a "specific minimum difference between present and near-future price"?

You try to prove trends have no "influence" on future price and I'll try to prove the opposite. If you make claims of this sort but cannot back them up, then why should I?

Edit: Are you asking my to prove the existence of a trend-following "holy grail"? You say trend following has to work to a degree. I call that an edge (perhaps not a good one, but one none the less). You give two things to prove wrong, but they are not clear demands.

Of course there are "trends" that persist into the future. Again, that's not the issue.

The point is that there is no combination of publicly available tools that can be used to determine a specific minimum difference between present and near-future price based on past price action. Your avoidance of the issue reveals your lack of an adequate response. Can you or can't you describe a suitable method, and if so, what is it?

As far as the meaning of "specific minimum difference between present and near-future price" goes, if you can't achieve such a difference, then you don't have a tradable edge.

 

[Spydertrader]

Quote from NickelScalper:

Anyone can prove both of these wrong simply by explaining, in terms that can be used to trade, how any combination of publicly available tools can determine a specific minimum difference between present and near-future price based on past price action.

I'm not sure if what follows meets the criteria you set forth above. However, I use the methods everyday.

My Tools: Price and Volume.

For a specific set of stocks, the relationship between price and volume allows me to anticipate a "specific difference between present and future price."

Using the following two assumptions:

If Volume is increasing, then Price will continue its trend over time.
If Volume is decreasing, then Price will change its trend over time.

I arrive at the following:

If Volume is increasing, with an increasing price, then price will continue to increase.

If Volume is increasing, with decreasing Price, then price will continue to decrease.

If Volume is decreasing, with increasing price, then price will begin to decrease.

If Volume is decreasing, with decreasing price, then price will begin to increase.

I begin by creating a Universe of High Quality Stocks (top 90% of EPS and RS Rank). I further refine the list by focusing on those stocks that cycle a minimum five times in six months for gains of 20% or more occurring over 6 to 8 days.

Finally, by comparing price, volume and accumulation / distribution, one can anticipate trend continuation or trend reversal with a certain degree of accuracy when a specific set of criteria have been met previously. Without delineating the exact criteria used (Interested individuals can locate the methods used in my Journal for additional detail), I have enjoyed profitable results with these methods over the last three years.

Now, I cannot anticipate correctly everyday (at least not yet) because the specific set of criteria I require do not appear each trading day (or I do not yet poses the experience required to recognize the set ups in real-time or hindsight on days I determine "no signals exist"). I don't claim to know the future everyday, and I don't claim 100% accuracy when the specific signals required for correctly anticipating trend reversal or continuation do appear. However, the signals appear with enough frequency and prove correct with enough regularity to permit me to make a living.

On the surface, it appears to me, that my methods contradict your theorem outlined above. However, I'll leave it to those with far more experience than I to determine if I met the criteria you specified.

- Spydertrader