Where is the line between fitting and actual patterns? - page 59
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Very interesting. I would like to get a feel for it, specialprogramme.....
2Gerasimm: Yes, really interesting... I'd love to see it too...
2Gerasimm: Yes, interesting indeed... I'd love to see it too...
there's a 30 day trial )
Write back afterwards with your impressions, please.
there's a 30-day trial )
Write back afterwards with your impressions, please.
Where is the line between fitting and real patterns?
Looking at the market we see that possibly existing patterns cannot be parametrically constant. Every system has a level of fit and a level of regularity of one or more events.
And the preponderance towards the second level is responsible for the rationality of the trading idea itself.
Thinking abstractly. The thoughts of others would be of interest.
It's been a good thread... Lots of opinions expressed. There are grains of truth in many of the comments.
But the question remains open.
I will try to answer it.
--------------------------------------
The line is in the length of the testing/optimization period.
Let's take for an example a TS based on 2 MA crosses, TF = H1, lot constant, optimization in three months easily gives us parameters for a "profitable" TS, which has
I have 125 trades during 3 months, each of them is profitable, MO = 10 spreads, FS = ~ 4.5. In short - a poet's dream...
Now let's try to apply the same trick to a period of 10 years.
And? It does not work.
The best set of parameters resulting in MO=2 spreads, FS < 1, trades 969 (i.e. ~24 for three months), drawdowns of a year and a half duration. In short - it sucks.
So, the optimization fails to solve the seemingly elementary task assigned to it. Why?
Maybe it's because we've crossed the line?
It was a good thread... Many opinions have been expressed. In many of the comments there are grains of truth.
But the question remains open.
I will try to answer it.
--------------------------------------
The line is in the duration of testing/optimization.
Let's take for example a TS based on 2 MA crossings, TF = H1, constant lot, three months optimization easily gives us parameters for a "profitable" TS with
For three months 125 trades, each of three months profitable, MO = 10 spreads, FS = ~ 4.5. In short - a poet's dream...
Now let's try to apply the same trick to a period of 10 years.
And? It does not work.
The best set of parameters with the result MO=2 spreads, FS < 1, trades 969 (i.e. ~24 for three months), drawdowns of one and a half years. In short, it sucks.
So, the optimization has not coped with this seemingly elementary task. Why?
Maybe because we have crossed a line?
Was it fault of optimization in this case? I doubt it.
The reason is in the primitiveness of the TS based on 2 MA crossings.
The longer the period - the less chances to adjust the parameters of 2 MAs while keeping the MO.
If we make the period longer, the MO will probably be even lower.
In this case such TS is a Procrustean bed.
As for the boundary - it seems to be really passed for this type of TS.
Jingo:
The market is often in a buoyant uncertainty after rallies or corrections of different forces.
I think the principles of entry and exit should be different at all.
I think IgorM is right when he says that "-exit from the market on a reverse entry signal". This principle should reflect the regularity within the TS itself. If such a principle does not lead to profits in the long run, then the entry signals should also be considered wrong.
As for the boundary - it seems to be really passed for this type of TC.
The criterion (or boundary) sought in this thread should not depend on the TC type.
I cited this TC only as a clear example for everyone.
.......................
Ok. Let's set a problem from the opposite direction:
You need any available TS by any optimization in the tester (even most severe overoptimization) finally produce the following results:
1) Number of transactions in 1 year at least 250-300.
2) mathematical expectation of at least two spreads.
3) Let the recovery factor be equal to four (minimum).
.............
Who can present a tester report with these results?
Immediately I see a forest of hands...
Ah, yes I completely forgot:
4) Testing range all available history from 1999 to 12.2010 (12 years)
=====================
If anyone can show something like this,
it would be appreciated.
PS And probably surprised. ))
But the question remains open.
I will try to answer it.
--------------------------------------
The line is in the length of the testing/optimisation period.
Let's make it more formal. The line itself needs to be articulated. Not what it is. And if you end up with a semblance of a formula, that's fine. It's an interesting idea in principle.
This is the edge, that for 12 years of history the market has been in different states, phases, etc..
And to fit such a test period, so to speak, a one-strategy TS is extremely difficult ( without self-interest or fanaticism, of course)
And if the results over 12 years are as set and stable, then we have found and use some kind of pattern.
Otherwise, it's a fit.
The criterion (facet) is clear.
.............................
Well, and certainly not as in the next branch the comrade is selling: all ticks, four years in the middle with no profit, etc.