Optimise an EA and get the best of the optimised ones. - page 9

 
Serqey Nikitin:

"The postulates are flawed... But it is clear - if there is no strict system in the strategy, and there is only a "gut feeling" approach, then stability is out of the question.

But experience is a good thing and you should try it, because a negative result is useful...

The postulates may be wrong. But they agree with my experience, and no one has ever presented me TC, which would always earn, regardless of its complexity.

Second objection - how is it "there is no strict system" if the rules of applied TS are very clear, and anyone who applies them - will get the same as I did to within the spread ?

The stability in the sense of profit is excluded by the first postulate.

The third objection - about the "lack of the idea".

Again it is very strange, given the maximum precision of the wording of TS. How is it "no idea" if the principles, in my opinion, are clear even to a beginner trader? There is no "lack of idea" - there is no "silver bullet", that is, the idea that would always work. That is why I use not one, but many TSs, and what is very important - based on different principles. The trend is determined by two non-intersecting ways - either by a price crossing the moving average (in fact, its very beginning - here a lot of errors may occur, but in case of guessing - it is taken as a whole), or by break in the channel - when the beginning of the movement has been lost, but there are less errors. Entry is either along or against the trend. Finally, accompaniment is also based on the principles of little overlap.

I myself was quite skeptical, when I was told that among such simple TS there will definitely be working ones at the moment. When I wrote them and ran them - I was convinced of that. Once again I suggest to look at the chart - do you think these are "non-working systems" ?? Yes, they can stop showing such beautiful results at any time (one of the systems did so a couple of days ago), but exactly the same can do ANY, however complex TS that has "deep ideas".

The second postulate is based on the fact that simple TS take into account a small number of market parameters. As a result, they will work no matter how many other parameters you change. But even a small change of "reference points" - will cause failure of the TS. Complex TS are based on a large number of different parameters, and therefore are not afraid of small changes of each parameter. But there are too many parameters themselves. Therefore "accumulation of the critical mass of changes" in them goes much faster than in simple TS, and they will fail with the same probability. But it will take much more effort to create such a complex TS with a large number of considered factors and "deep idea". And what for?

Finally "summing strategies" - does not help with mindless random summation. If the best TCs are summed - then the sum is equivalent to diversification of TCs, which is very widely used. How can it "not help" ?

And if to remember the folk epos, we can remember the collective farm storekeeper Kuzma Gladyshev (from Voinovich's story), his "theory of the shit circle in the nature" and his fine moonshine, which was estimated by Chonkin as a "proper brew".

 
George Merts:

Again - very strange, given the utmost clarity in the wording of the TS. How is it "missing an idea" if the principles, in my opinion, are clear even to a novice trader?

Yes, it's "very strange"... Considering that ANY strategy is based on the principle of stability, it is not clear why you need to sum up the results of SEVERAL strategies? One stable strategy is enough to make a profit.

If your strategies do not allow you to obtain a stable profit, then summing up such deficiencies will further reduce this same summed profit.

There is no need to invent anything new here... It is necessary to develop ONE profitable strategy, and "do not suffer nonsense"...

 
Serqey Nikitin:

Yes, it is "very strange"... Considering that ANY strategy is based on the principle of stability, it is not clear why you need to sum up the results of SEVERAL strategies? One stable strategy is enough to make a profit.

If your strategies don't allow you to make a stable profit, then summing up such underperformance will further reduce that very summed profit.

There is no need to invent anything new here... It is necessary to develop ONE profitable strategy, and "do not suffer nonsense"...

Wrong (let's be "you").

If there was even one "stable" TS - it would have been known to everyone a long time ago and everyone would have used it. The problem is that any TS works only as long as its principles are used by a small part of traders. As soon as most traders start using these principles - the TS stops working, the TS on opposite principles starts working.

What kind of "stability at the core of any TS"...

We must sum up the results of TS to increase just the stability of the result because in adding independent random variables, the expected payoff equals the sum of these variables, but the standard deviation is the root of the sum of squares of standard deviations (in the absence of correlation - that is why I emphasize the importance of TS based on different principles for one symbol).

Let's say if we have a TS that gives 1 point profit plus minus 5 points - then using it alone - we can have -4 to +6 points of profit.

But if we take two such TS, and divide a lot between them, they will give (0,5 + 0,5) points profit plus minus SQRT(2.5^2 + 2.5^2) - so we can have from -2.54 to +4.54 points profit. By increasing the lot we get from -4 to +7.15 points of profit! And that is with the same expected payoff of 1 point.

The difference between maximal winnings (with the same loss), as you see, is even bigger than expectation.

Is this what you call "suffering nonsense" ???

 
George Merts:

Wrong...

If there was at least one "stable" TS, it would have been known to everyone a long time ago, and everyone would have used it.

Where did you get such conclusions? No one, just like that, will not replicate their profitable strategies ... Yes, there are such strategies, and it's no secret. Even statistics say about 5% of successful traders, but to share with all - is to harm yourself ...

To develop a profitable strategy is quite real, but difficult enough... And it's just stupidity to focus on simple strategies... But you can check it...

 
Serqey Nikitin:

Where is this coming from? Nobody, just for fun, will replicate their profitable strategies... Yes, there are such strategies, and it is no secret. Even statistics say about 5% of successful traders, but to share with all - is to harm yourself ...

Develop a profitable strategy - it is quite real, but difficult enough ... And the emphasis on simple strategies - it's just stupidity ... But you can check it out...

You can't hide the truth in a bag. Any profitable strategy - will always become inevitably known to all.

I am sure that there is no TC that always earns. And the only way - is a constant selection of the most stable TS, from among those who are earning now.

That's the idea I'm testing.

Above I showed as many as ten profitable TS! And this is not a tester, this is a demo trade. With the provision of an account and investment password for all who wish. When I adjust "full set of TS" with their regular updating-overoptimization - should only develop clear criteria for selection of the most stable TS, that have already shown good results not only in back and forward tests, but also on the demo, and put them in the real account. I think it is unreal to gain the deposit this way. However, to have a constant small profit is quite possible.

 
George Merts:

You can't hide the truth in a bag. Any profitable strategy will always be inevitably known to all.


The main method of strategy repetition is the analysis of profitable trading reports. But if the TS is complex enough, it cannot be repeated... And such strategies are used by successful traders. Simple strategies are written everywhere and in details, but there is no success... Your synthetic methods - combination of several simple strategies is a waste of time and effort, but if you don't test it yourself, it is impossible to change your mind... GOOD LUCK!
 
Serqey Nikitin:
The main method of strategy repetition is the analysis of profitable trade reports. But, if the TS is complicated enough, it is impossible to repeat it. And such strategies are used by successful traders. Simple strategies are written everywhere and in details, but there is no success... Your synthetic methods - combination of several simple strategies - is a waste of time and effort, but if you don't test it yourself, it is impossible to change your mind... Good luck!

It is entirely possible to change my mind. But with real arguments.

My real arguments are the ones I gave above, including the charts of the TS, and the investment password of the account.

And what are your arguments? Why would it be "impossible to replicate a sufficiently complex TS"? A processor consisting of hundreds of millions of transistors - repeatable, but the TS consisting of an order of magnitude less rules - "unrepeatable"?

If it was, all banks would have had it long ago, and would have used it. But all banks for some reason provide signals (including internal users). And they do not like Forex pairs, all of them prefer to use stock market, although there is more earning potential on Forex pairs.

The reason is simple - in the share market all participants are interested in growth of share prices, that is why price of most liquid shares grows more often than falls. And there is no such thing at Forex, they can grow as well as fall with equal probability.

There is no silver bullet. There isn't.

There is only constant adaptation to the changing situation. Which is what my "TC League" is designed to do.

 
George Merts:

It is entirely possible to change my mind. But with real arguments.

My real arguments are the ones I gave above, including the charts of the TS, and the investment password of the account.

And what are your arguments? Why would it be "impossible to repeat a sufficiently complex TS"?

If you think this strategy ( in the appendix ) deserves your attention, then try to repeat it...

Files:
temp.zip  28 kb
 
Serqey Nikitin:

If you think this strategy ( in the appendix ) deserves your attention, then try it again...

There's only a month of trading there. (Let's keep it on a first name basis, mate, we know each other quite well in absentia !)

Look at the above - any of the TS I presented has a similar chart shape, and the running time is no less.

Your(your) graph only convinces me that the quality of a TS, its running time before it starts to drain - does not depend on complexity. If the TS presented by you(you) is "very complex" - then most likely a lot of time has been spent on its development. At the same time - I have several TS, which show very similar results, and I have spent very little time on their development, I think, much less than you (you).

Three of my presented systems, each gave more than $50 profit, with a fixed lot 0.01. With your (your) lots up to 9 units - they would have yielded up to 45K each. And I have "in the black" more than a hundred systems ! If we're talking about those that have made more than a dozen trades, and in profit - there are more than 60 of them.

And now I have a question - why repeat a complex TS, which probably "hatched" more than one month, which had to think a long time, for which you had to "care and worry" - if the dumbest and most stupid TS (and not one) - produce similar results?

 
George Merts:

And now I have a question - why repeat some complex TS, which probably "nurtured" more than one month, which you had to think about for a long time, for which you had to "care and worry" - if the dumbest and dumbest TS (and not one) - give similar results?

The answer is simple: a complex TS on the right idea gives a stable profit from month to month. And a set of simple TS, of course, will work, but not stably profitable. Surely some of the months will be unprofitable ... Traders who are satisfied with this state of affairs, can live with it. But those Traders who work for Investors (Trustee), this will not stand...