Stable MTS - page 25

New comment
 
Сергей:
Forex is a system managed on every tick. Such systems do not lend themselves to statistical analysis.
Sorry Sergey, my knowledge is lacking here. Please explain in more detail, I really do not understand. What does it mean controlled on every tick? Why are such systems not subject to statistical analysis (which lends itself to everything in the world)?
 
Сергей:
This can easily be checked through Mat Lab. Yuri wrote about this above when he talked about autocorrelation
He was talking about something else. And he was just suggesting to use a standard statistical test.
 
Oleg Shenker:
He was writing about something else. And he was just suggesting to use a standard statistical test.
which worked in forts but failed in forex)
 
Yuriy Asaulenko:

Of course there is such data.

By the way, the indicator I suggested was quite good and makes physical sense.

You offered a great indicator!!! I am being absolutely sincere. You will laugh, but it has the physical meaning of economic expectation, which is calculated as the product of the profit value multiplied by the probability of profit minus the loss value multiplied by the probability of loss.
 
Сергей:
which worked in Forts, but failed in Forex)

Perform a statistical analysis of the series

79; 0; 67; 15; 60; 14; 59; 16; 53; 17; 40

This is just an example, so we can understand each other later.

 
Комбинатор:

Do you think the owner of such an algorithm would show a working system to just anyone? You think he's really only interested in the money?

You're in the middle of nowhere. And you're only interested in the money.

I think it's not about having or not having an algorithm that meets the requirements. The requirements are quite loyal. But it is not clear how organisationally, technically and legally the author proposes the cooperation process.
 
Oleg Shenker:
You suggested a great metric!!! I am speaking absolutely sincerely. You will laugh, but it has the physical meaning of economic expectation, which counts as the product of profit value multiplied by profit probability minus loss value multiplied by loss probability.

I've already spun this all around. So far, I have called it the direct cost-recovery ratio, conventionally.

It made me smile to see that, just by analogy with the uncertainty principle, the proposed expression works. )

 
Oleg Shenker:

Sergei, I, for example, got a lot out of you. So you don't think this is a lecture for nothing.

Estimating the accuracy of matrix expectations is done through confidence intervals. Why are these mathematical models not applicable to FOREX? Explain why!

To put it in perspective, you have an Expert Advisor with positive mathematical expectation. You shift the testing period and it fails. This suggests that the expectation spread (accuracy) is 100%, i.e. there is no forecast ......
 
Oleg Shenker:

As an algorithm that earns a little bit, the main thing is that the drawdowns should be non-catastrophic (within 10%). I've heard stories about them, I've seen pictures of capital charts, but I've never managed to test one.

Ok, I'll ask a more specific question, has anyone seen here in the Market algorithm, which when tested, at least a little closer to the ideal, with a max. drawdown <10%, and a recovery time after DD at least 3-4 months.

https://c.mql5.com/6/724/mover11_eurusd_newopt.JPG

If initial deposit is increased by 5 times while keeping lot, it will be just 10% drawdown and 11% p.a. I trade with max 50% drawdown and 50% p.a.. Well, that's on paper, of course. ;)

 
Alexey Burnakov:

https://c.mql5.com/6/724/mover11_eurusd_newopt.JPG

If you increase the initial deposit by 5 times while keeping the lot, it will be exactly 10% drawdown and 11% p.a. And I am trading at max 50% drawdown and 50% p.a. Well, this is on paper, of course. ;)

Good afternoon, can you test on a 5 minute period with a spread of 22? Very interesting
1...1819202122232425262728
New comment