You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Systemized profit-taking?
- Thread starter illiquid
- Start date
In my experience attempting to develop mechanical trading systems for intraday, most of them benefitted significantly from the addition of a "take profit" rule. Admittedly, most of them are mediocre at best (not worth the trouble to trade). However, from contemplation of those few which are a little better (they make minimum wage), I have a working hypothesis that the very best systems are optimized with a "time in trade" exit rule and are degraded by a "take profit" rule. (All this is independent of consideration of stops, the optimization of which typically must be determined iteratively with the other exit rules.)
Since I began, I have let the market dictate what to do by the signals it provides to me. Your question is one that could be asked of the market.
I gather data periodically (frequency is dictated by market conditions) as a prescribed set. I analyze the complete prescribed set in a standard way. After that I make decisions based on what I believe the market is telling me. I take actions that fit with what is in the space.
The third part, making decisions, is a belief system template that I have created by trying to understand the workings of the market. The market seems to work consistently in the same way in each of its operating points. By understanding the market in this way, I, in effect, abide by what the market tells me to do.
So I do not get targets or % limits for profits as most ET posts suggest. What I get periodically are go/no go decisions based on complete sets of data and the corresponding analysis . A go is often a "continue". No go is more complicated; it means to act to take profits and usually, and more important, reverse to continue in the market where the market has a new operating point and more profits in sight for me.
The decision is made by looking at three consistent distinct complete facets of the market. A programmable tree so to speak.
The idea that there is an answer to your question in terms of "targets" and/or % amounts of profit would be like looking at a poker game and expecting "targets" and/or a % win. It is not possible to gin up a set of answers in playing poker at all usually. Comparing this to poker is not a great analogy, of course. But it is possible to gain a little of the necessary insight required. Set ups can be compared to hands beginning to be dealt. I a trader can keep an open mind to some extent, then it is possible to continue to play the hand instead of do the deadly deed of "targeting the hand or % profitting the hand. you may be able to get to a place to "see" every poker hand has many consistent distinct complete facets to consider. Often by the play of the hand they are presented in a series of time frames as in the market's progression. The same is called for in trading to get an appropriate exit point.
I gather data periodically (frequency is dictated by market conditions) as a prescribed set. I analyze the complete prescribed set in a standard way. After that I make decisions based on what I believe the market is telling me. I take actions that fit with what is in the space.
The third part, making decisions, is a belief system template that I have created by trying to understand the workings of the market. The market seems to work consistently in the same way in each of its operating points. By understanding the market in this way, I, in effect, abide by what the market tells me to do.
So I do not get targets or % limits for profits as most ET posts suggest. What I get periodically are go/no go decisions based on complete sets of data and the corresponding analysis . A go is often a "continue". No go is more complicated; it means to act to take profits and usually, and more important, reverse to continue in the market where the market has a new operating point and more profits in sight for me.
The decision is made by looking at three consistent distinct complete facets of the market. A programmable tree so to speak.
The idea that there is an answer to your question in terms of "targets" and/or % amounts of profit would be like looking at a poker game and expecting "targets" and/or a % win. It is not possible to gin up a set of answers in playing poker at all usually. Comparing this to poker is not a great analogy, of course. But it is possible to gain a little of the necessary insight required. Set ups can be compared to hands beginning to be dealt. I a trader can keep an open mind to some extent, then it is possible to continue to play the hand instead of do the deadly deed of "targeting the hand or % profitting the hand. you may be able to get to a place to "see" every poker hand has many consistent distinct complete facets to consider. Often by the play of the hand they are presented in a series of time frames as in the market's progression. The same is called for in trading to get an appropriate exit point.
...could you comment, please, on the time frame (intraday, daily, etc.) in which you apply the approach you so eloquently described? Thanks.
I've tried both targetted exits and time-expiry exits and I found neither to be satisfactory.
The only satisfactory approach I've found is to identify the failure point for a trade, and exit there, which can be either an exit or a reverse depending on the signal. I've yet to find a way of automating it unfortunately...
Best
Natalie
The only satisfactory approach I've found is to identify the failure point for a trade, and exit there, which can be either an exit or a reverse depending on the signal. I've yet to find a way of automating it unfortunately...
Best
Natalie
...may I ask, please, if your conclusions are based on backtesting to a set of rules? No offense implied. Rigid personalities love rules. Thanks. - Mike
Quote from hypostomus:
...may I ask, please, if your conclusions are based on backtesting to a set of rules? No offense implied. Rigid personalities love rules. Thanks. - Mike
Yes. Many backtests to many sets of rules. I found no fixed targets or time exits to be satisfactory...
Best
Natalie
...that is so contrary to my own experience (admittedly limited) that I would like to ask, please, what time frame you trade in, and what instruments. It occurs to me that perhaps you may be trying to capture a larger fraction of a move than I am, which might explain your dissatisfaction with targets. Thanks. - Mike
Quote from hypostomus:
In my experience attempting to develop mechanical trading systems for intraday, most of them benefitted significantly from the addition of a "take profit" rule. Admittedly, most of them are mediocre at best (not worth the trouble to trade). However, from contemplation of those few which are a little better (they make minimum wage), I have a working hypothesis that the very best systems are optimized with a "time in trade" exit rule and are degraded by a "take profit" rule. (All this is independent of consideration of stops, the optimization of which typically must be determined iteratively with the other exit rules.)
Natalie...
...that is so contrary to my own experience (admittedly limited) that I would like to ask, please, what time frame you trade in, and what instruments. It occurs to me that perhaps you may be trying to capture a larger fraction of a move than I am, which might explain your dissatisfaction with targets. Thanks. - Mike
Hi Mike - not trying to be difficult, but these 2 statements are at odds with eachother. On the one hand you say the systems are mediocre at best, on the other you say that my findings of time and fixed target exits is contrary to your experience.
Your answer lies in your first statement. I have tested extensively with both of these methods (and combinations) and even built statistical models of the difference between exiting at an artificial exit point, and exiting upon a failure point. In all timeframes it proved to be better to exit at the natural failure points...
I think the main reasons are that with targets and times, some targets are not met and a trade will run into loss before the time exit or stop takes you out, others exit prematurely (albeit with a profit) leaving large amounts of money on the table unnecessarily.
I have methods and systems that are mediocre with fixed targets and time exits/stops etc, but become very much better when using the failure points because the winners are not capped, while the losers are unaffected so increasing the overall returns by much more. The vexing question that I am grappling with right now is how to identify/quantify the failure points in mechanical terms.
Kind regards
Natalie
Quote from Girlpower:
Hi Mike - not trying to be difficult, but these 2 statements are at odds with eachother. On the one hand you say the systems are mediocre at best, on the other you say that my findings of time and fixed target exits is contrary to your experience.
Your answer lies in your first statement. I have tested extensively with both of these methods (and combinations) and even built statistical models of the difference between exiting at an artificial exit point, and exiting upon a failure point. In all timeframes it proved to be better to exit at the natural failure points...
I think the main reasons are that with targets and times, some targets are not met and a trade will run into loss before the time exit or stop takes you out, others exit prematurely (albeit with a profit) leaving large amounts of money on the table unnecessarily.
I have methods and systems that are mediocre with fixed targets and time exits/stops etc, but become very much better when using the failure points because the winners are not capped, while the losers are unaffected so increasing the overall returns by much more. The vexing question that I am grappling with right now is how to identify/quantify the failure points in mechanical terms.
Kind regards
Natalie
Hi Natalie,
I've enjoyed your posts on the forum.
A few questions on this topic, even though your last paragraph hints that you do not have the answer, but then that seems to contradict your statement that you have tested these hypotheses.
My question: You say determing a failure point is better than a profit target or a time-exit. So how do you define/identify a failure point? Then you now say, that you have not been able to determine mechanically a failure point, so how do you know then that a failure point is a superior exit to a profit target or time-exit?
Thanks.