I've exported hundreds of times, perhaps its my brain dead in front of this computer the last two nights.... files arents saving....
Emini divergence journal
- Thread starter no_pm_please
- Start date
- Status
I've exported hundreds of times, perhaps its my brain dead in front of this computer the last two nights.... files arents saving....
He also exits on the close. In the code above you would be holding overnight if neither stop or target is hit by the close.
Also, divergence is a very subjective thing. I have yet to see code for divergence that I thought represented what I believed to be divergence 100% of the time in real time analysis. While I think you should test as much as possible, you will not be able to replicate no_pm's trades with this. I bet if you compare his realtime entries with the test that they would rarely be the same. So the test may or may not have any relevance to the method here.
I don't presume to speak for no_pm here, but I think the key to his results are in his money management discipline. His winners are bigger than his losers, and his entries seem to put him in the market better than random (based on his actual results and not the attempt to back test it hypothetically).
I have known other traders that made money using simple techniques that to anyone who has tested thousands of systems would have seemed impossible. There is some intangible there that can't be explained by a mechanical system test.
Also, divergence is a very subjective thing. I have yet to see code for divergence that I thought represented what I believed to be divergence 100% of the time in real time analysis. While I think you should test as much as possible, you will not be able to replicate no_pm's trades with this. I bet if you compare his realtime entries with the test that they would rarely be the same. So the test may or may not have any relevance to the method here.
I don't presume to speak for no_pm here, but I think the key to his results are in his money management discipline. His winners are bigger than his losers, and his entries seem to put him in the market better than random (based on his actual results and not the attempt to back test it hypothetically).
I have known other traders that made money using simple techniques that to anyone who has tested thousands of systems would have seemed impossible. There is some intangible there that can't be explained by a mechanical system test.
Good post...
Some of the realities about backtesting methods that test strategies used by traders that have discretion as part of the methodology.
Trading only certain times of the day, ignoring FED days or afternoons, not trading on 1/2 market days prior to a holiday weekend, not trading the dead zone, closing positions by a specific time period, not taking positions until a specific time period,
the above is common place of discretionary traders.
Therefore...backtesting results will always be different than actual trading results until there's a way to incorporate the things that cause traders to ignore signals or take unclear signals.
Although backtesting its fun to play around with.
Yet...backtesting has more validity if the trader was using a mechanical system.
Just as true as one traders actual results will be different than another traders actual results...
sometimes extremely different.
P.S. I gave up using info from backtesting discretionary strategies a long time ago.
I know prefer putting on a very small position (1 to 3 contracts) and test someting via real trading after paper trading it in a realtime simulator.
That would be my best indicator of the validity of any discretionary strategy with strict rules and some subjectivity.
NihabaAshi
Well all I can say if anyone trades the above technique, good luck to you.
Unfortunately, at less than 30% accuracy, there are much better methods of fading the market. (I have no idea how no_pm jumps a method from 28% accuracy to over 70% accuracy) If in fact he does, his market skills are the reason for his success, not the methodology.
One can have a great money management plan, but if the plan is backed by a negative expectancy, the mm plan doesn't matter.
If you close out all the trades at the close, the results are worst.
Divergence doesn't have to be a subjective thing, but I guess as discretionary traders go, they have an uncanny ability to judge WHICH ONES are going to work. I don't know about you, but my feeling is after 14 years of market experience, my intuitions are up there with the best. The last thing I would put my intuition to the test would be a methodology that has 2.8 winners for every 10.
We've tested many mechanical and discretionary methodologies, and I just have to state there is absolutely no way this methodology will make money over time... while it has good risk reward, the reward CHANGES as the profit moves up or down so in very real terms, the risk:reward is not as concrete as you might think.
Putting more rules into the methodology only makes it worst. Times, Fed meetings, etc....
Not testing a methodology or system before putting money behind is irresponsible. Backt-testing to judge results, whether purely mechanical or discretionary gives the trader confidence, experience, and an understanding of how it works.
Successful traders, whether discretionary, mechanical, chart readers, or quants all do the same process. They go over as many instances of a pattern to allow them to understand what the probabilities are for the future.
Finally, and I have to say again, anyone who trades this method will lose over time, especially newbies...
The accuracy is way to low:
a) the risk:reward is not truly defined, the risk will ALWAYS BE 1 UNIT while the reward can FLUCTUATE (while it is better than most because it is a strict plan).
b) there is no statistical edge in fading the market on these parameters when ADX is above 30
c) there is no statistical edge AT ALL in the chaiken finding market turns at these parameters
d) there is no statistical edge in determining when the market decides to play ping pong on the 5 minute chart
No matter how discretionary you may get, no matter how great you are at trading the market, why would any trader want to trade a methodology where each parameter of the methodology has no statistical edge?
There are better, simpler ways.
Unfortunately, at less than 30% accuracy, there are much better methods of fading the market. (I have no idea how no_pm jumps a method from 28% accuracy to over 70% accuracy) If in fact he does, his market skills are the reason for his success, not the methodology.
One can have a great money management plan, but if the plan is backed by a negative expectancy, the mm plan doesn't matter.
If you close out all the trades at the close, the results are worst.
Divergence doesn't have to be a subjective thing, but I guess as discretionary traders go, they have an uncanny ability to judge WHICH ONES are going to work. I don't know about you, but my feeling is after 14 years of market experience, my intuitions are up there with the best. The last thing I would put my intuition to the test would be a methodology that has 2.8 winners for every 10.
We've tested many mechanical and discretionary methodologies, and I just have to state there is absolutely no way this methodology will make money over time... while it has good risk reward, the reward CHANGES as the profit moves up or down so in very real terms, the risk:reward is not as concrete as you might think.
Putting more rules into the methodology only makes it worst. Times, Fed meetings, etc....
Not testing a methodology or system before putting money behind is irresponsible. Backt-testing to judge results, whether purely mechanical or discretionary gives the trader confidence, experience, and an understanding of how it works.
Successful traders, whether discretionary, mechanical, chart readers, or quants all do the same process. They go over as many instances of a pattern to allow them to understand what the probabilities are for the future.
Finally, and I have to say again, anyone who trades this method will lose over time, especially newbies...
The accuracy is way to low:
a) the risk:reward is not truly defined, the risk will ALWAYS BE 1 UNIT while the reward can FLUCTUATE (while it is better than most because it is a strict plan).
b) there is no statistical edge in fading the market on these parameters when ADX is above 30
c) there is no statistical edge AT ALL in the chaiken finding market turns at these parameters
d) there is no statistical edge in determining when the market decides to play ping pong on the 5 minute chart
No matter how discretionary you may get, no matter how great you are at trading the market, why would any trader want to trade a methodology where each parameter of the methodology has no statistical edge?
There are better, simpler ways.
Maybe you could start your own journal or thread and mention one of these
simple ways to take money out of the
the market on a semi consistent
basis, if you think there is such a thing.
Personally, and this may only be my
opinion, I believe your usefulness to this thread is over now. You have stated, over and over, that you don't think no pm's method works. Ok, we understand. Now what is needed
is for you to churp up with suggestions
of how no pm can improve his method
for better performance over time,
using your 14 years of trading
experience to help us. Otherwise,
what's the point of your continued
posting of displeasure with this
journal's method?
jd
i like offshore's comments, but i like to hear contrary opinions on anything, helps me get a more rounded view afterall we all have limited mental capacities and others musings are required to "see" more clearly if you know what i mean having said that the answer to your second part of question #1 is he is *clarifying* his point even further i like this whats the harm in further elucidation of a problem why are you so concerned over his posts you have a personal steak in this
Quote from randynutts:
i like offshore's comments, but i like to hear contrary opinions on anything, helps me get a more rounded view afterall we all have limited mental capacities and others musings are required to "see" more clearly if you know what i mean having said that the answer to your second part of question #1 is he is *clarifying* his point even further i like this whats the harm in further elucidation of a problem why are you so concerned over his posts you have a personal steak in this
no...no personal state.
It just seemed a bit over the top.
But I'm not the moderator. If you guys
want this guy's comments, fine by me.
Could be I over reacted to the
constant negative tone. But if someone
has 14 years of trading experience
and tells us he has better systems,
I'd sure like to hear him talk about them, not that he has any obligation to do so of course.
It's Sunday and sunny here in Seattle. It's a happy day.
regards,
jd
with a handle like jackdaniel every day is happy
ps my other job is field tester for jackdaniels
Since it seems that no_pm doesn't post on weekends, perhaps it would be best to wait until he has a chance to offer his observations on Offshore's posts, rather than jumping all over each other Personally I don't care for Offshore's tone, and I don't really see how criticism of this sort can help an ongoing trading journal, but I understand if he feels the need to warn us if he sees a problem with the overall methodology. Which, if I read it correctly, is not really represented properly by the code he posted.
By the way, steak = meat! Stake = "an interest in".
Just my moment of levity
Mark
Personally I don't care for Offshore's tone, and I don't really see how criticism of this sort can help an ongoing trading journal, but I understand if he feels the need to warn us if he sees a problem with the overall methodology. Which, if I read it correctly, is not really represented properly by the code he posted.By the way, steak = meat! Stake = "an interest in".
Just my moment of levity
Mark
- Status