Market etiquette or good manners in a minefield - page 97
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
I have a question, but don't kick it... :)
So we have a simple reversal EA working on first differences of pt vs last count. Of course, such EA has no stops - it just reverses against the last pt count and that's it. What will happen if in case of a loss in a previous transaction the next transaction will be opened with a double lot? I am aware of the fact that this is almost obscene flirting with martingale, but still?
No, no, Fedor, don't confuse martingale with martingale. What you wrote is martingale.
And what can be martingale in a random vector process?
No, no, Fedor, don't confuse martingale with martingale. What you wrote is martingale.
And what can be martingale in a random vector process?
A wallet state depending on that random process-vector?
Yeah, I've had a bit of a 'blunder' recently too. Said flipping a coin was a martingale. That's not true. A coin flip is a stationary process and it has a mathematical expectation. A martingale, on the other hand, does not have a mathematical expectation.
https://ru.wikipedia.org/wiki/%D0%9C%D0%B0%D1%80%D1%82%D0%B8%D0%BD%D0%B3%D0%B0%D0%BB
It's all there.
No, no, Fedor, don't confuse martingale with martingale. What you wrote is martingale.
Yeah, "Babel and Hegel are completely different people" kind of joke :o).
https://ru.wikipedia.org/wiki/%D0%9C%D0%B0%D1%80%D1%82%D0%B8%D0%BD%D0%B3%D0%B0%D0%BB
it's all there.
Yes, of course, the player's state as a function of the number of games is a martingale. But I meant a little differently - there is a mathematical expectation of the number of throws before "tails" appear and it equals 2.
No, no, Fedor, don't confuse martingale with martingale. What you wrote is martingale.
Got it, thanks... -:) I thought Karl Marx and Friedrich Engels were husband and wife.
But it turns out they're four different people :o
It's happening! I have four different grids with the results of statistical modelling of EURUSD 1h trades, the number of training epochs is 2000. But first I will show you the data already posted above, for the number of epochs N=1000:
The figure on the left shows profitability as an average of pips per trade (we averaged over 20 independent numerical experiments). In red, the performance of a single linear neuron as a function of the number of inputs is shown (abscissa axis). Blue - non-linear NS with one hidden layer and two neurons in it (output of 1 neuron - Buy/Sell). Black - with four neurons in hidden layer and lilac - with 8. You can see that with increasing number of inputs, profitability slowly increases for all configurations, and with increasing number of neurons in hidden layer the stability of NS-based NS slightly increases. To the right are the normalized variance plots for the training sample (with index P) and for the test sample (with index E). The normalisation was performed on the variance of the input data. A value of <1 indicates that the network is "trained". For all configurations, the training sample length P was assumed to be P=w^2/d. The fact that for NS with small number of neurons there is a strong divergence of variance in training and test samples, indicates a small sample length according to this estimation, and on the contrary, at the number of neurons more than 4 in the hidden layer, we observe the convergence of these indices, which indicates an overestimate of the length.
Let us now compare this data with the new data, where the number of training epochs is doubled:
It can be seen that there is no radical improvement and is not foreseen. Probably, the situation with accurate prediction can only be improved by figuring out the optimal length of the training sample. I will provide a set of statistics for the case when P=16*d.
So we have a simple reverse EA working on first differences of RT vs. the last reference. Of course such EA has no stops - it just rolls vs. the last RT reference and that's it. What will happen if in case of a loss in a previous transaction the next transaction will be opened with a double lot? I am aware that this is almost an obscene flirtation with martingale, but still?
Fedor, a proper trading robot can be broken down into two main components - analysis unit - decides on the direction of a position to be opened and maximizes profit in pips, and MM unit - maximizes profit in the currency of the account. From this point of view, what you offer is a two-in-one. You have "wrong" TS, the results of which you correct by analyzing the past events (transactions) and on top of all this you put a primitive MM - lot doubling depending on the outcome. In principle it will work somehow, but all this together, it looks like putting your trousers over your head. Why?
So, is the right trading robot different from the wrong one by separating the analysis block and the MM block? Or is there something wrong with the analysis block?
Right now I am carefully examining in Matkadec the tick charts with PT constructions at different H. I would like to find some other ways of working besides the one described by Pastukhov, because the latter still leads to a pipsarian, no matter how hard it is. Here it is necessary either to recognize pipsarian as a quite adequate expert, which I don't want to do for aesthetic and not only aesthetic reasons, or... take from this method only that which it can give - namely, the idea itself, the basis of which is H and look for ways to bypass the inevitable pipsing in the classics of this method.