Obtaining a stationary BP from a price BP - page 11

 
Svinozavr >> :

Well... on the back of "You can close the forum too..." I guess not.

In my opinion, in the background.

All you have to do is to read botanical books on econometrics and identify their flaws, i.e. do everything they advise, only in reverse.

the man's been convinced enough already

 
Reshetov >> :

In short, we divide the story into three parts.

In the first part we feed NN inputs with residuals from the price difference and (I won't say what). The neural network is lagging. We obtain weight coefficients. We extrapolate.

In the second part we check for extrapolation. We see that there is a fit. Subtract first NN's extrapolation from the remainders, i.e. we get one more BP of the remainders (errors of the first NN). We feed the inputs of the second NN.

On the third part the second NN corrects errors of the first NN. Forward is successful.

This is a well-known technique. I think many here know the NeuroPro website, in particular the Articles page. What's important for this discussion is this one - Optimisation of experts by their learning curves. I trust this author and know for sure he's not a nerd (as some people call each other ;-) ).

Of course BPs are different and it's not a fact that you can get a positive result on any quotes. But the fact that Reshetov voiced the idea once again in application to Forex, only confirms that it deserves attention here.

 


Reshetov писал(а) >>

Форум тоже можно закрывать - мне здесь ловить больше нечего. На всех инструментах, которые успел проверить в тесте, есть профит на третьем участке.

В общем, если я здесь больше не появлюсь, то значит ловить в этом флуде действительно больше нечего.

Засим прощаюсь с местной публикой

That's funny.

See you, Yura... in a month's time :-)

HideYourRichess >>:

Nothing like that.

What is being discussed, in the broadest sense, is getting from a non-stationary series of prices, a stationary series of profits.

Well said - precise and concise.

By the way, there are direct proofs of the possibility in principle of such transformation under non-stationarity of generating series and, what is the most important, taking into account commission costs. Here is the result of the PRICE->PUTS transformation on more than 2000 counts:


This is the locked-in account of the legendary Better on Onyx.

It seems (let the mathematicians correct me by some criterion of evaluation) that such BP dynamics is not accidental and reflects the fact that we really have something to strive for. If we attach an optimal MM to such a TS, we will get the output of the DREAM.

 

Profit factor 1 and mate expectation is just over $4. With an initial deposit of 3K.

Apparently pips and optimal MM.

Not impressed. ;)

 
Sorento >> :

Profit factor 1 and mate expectation - a little over 4 dollars. With an initial deposit of 3K.

Apparently pips and optimal MM.

Not impressed. ;)


Here, Sorento, it seems to me that everything is not so simple and does not always boil down to the hackneyed "profit factor and mate expectation" to say "Not impressed" in such a simple way. Or it does, but you need to understand all the subtle points. It is easier to simulate on a computer the given trading algorithm (or rather its probable model) and look at its implementation. At a rough guess (a rough estimate) we can assume that increments of points on the trading account have normal distribution with the MR of 25 points and variance of 500 points. The picture on the left represents one of several hundred test realizations of this process. Compare it with the picture above. I think they match satisfactorily.

Now let's load an optimal MM on this case and see the account balance after 3000 transactions as a function of the leverage size (TP). The figure on the right shows relative account deposit increments obtained by averaging over 200 independent realizations for each TP value. The whiskers show the characteristic value of the statistical variation of 1/e.

We can see that there is an optimal TP value maximizing the account balance for the presented TP. At TP=25-30 deposit growth for three years amounted to about 40 times (not %, but times!) with a scatter from 30 to 60 times.

Are you not impressed by this, Sorento?

 
marketeer >> :

It's a known trick. {...}

If you read and study too much you will never make a discovery :-).

 
Neutron >> :

By the way, there is a direct proof of the possibility in principle of such transformation under non-stationarity of the generating series and, most importantly, taking into account the commission costs. Here is the result of the PRICE->Points transformation on more than 2000 counts:


This is the locked-in account of the legendary Better on Onyx.

It seems (let the mathematicians correct me by some criterion of evaluation) that such BP dynamics is not accidental and reflects the fact that we really have something to strive for. By screwing an optimal MM to such a TS we will get the output of the DREAM.

Interesting account. Unfortunately, I don't know the intricacies of the strategy, so I can't say anything. There are suspicions that it is a regularly adjusted or self-adjusting system.

 
jartmailru >> :

If you read and study too much, you'll never make the discovery :-).

IMHO, you have to have some basic knowledge that could be combined to make a Discovery. ;-) Besides, why reinvent the wheel when much has already been done before us, and even better than we can afford (in terms of research etc.)?

 
marketeer >> :

IMHO, you have to have some basic knowledge {...}

That was a joke. And the bike will be crooked...

 
marketeer >> :

IMHO, you have to have some basic knowledge that could be combined to make the Discovery. ;-)

>> Yes. Of course. But that's where

Besides, why reinvent the wheel when much has already been done before us, and even better than we can afford (in terms of research etc)?

you can argue. Letting the basics pass you by - without doing the mental work of deriving the concept of the bike, you can stay forever within that concept, for the main thing disappears: what's it all for? And all you will be doing in the future is perfecting the bicycle.

===

What a load of crap. Sorry - haven't quite woken up yet. I have a fight with someone and then I will go back to normal. )))