On the unequal probability of a price move up or down - page 115

 
Grigori.S.B:

If you want to invent statistical arbitrage, correlation has nothing to do with it. Two highly correlated instruments can diverge for as long as you like. What matters here is cointegration, not correlation. Read for example here. And compare returns at high correlation and cointegration. Correlation and cointegration are not related and independent concepts.

And then we're back to martin again
 
b2v:

Apparently the estimated lots change slowly over the course of a day, but over the course of a week they change significantly;

yes, that's true.

 

b2v:

Apparently the estimated lots change slowly over the course of a day, but over the course of a week they change significantly;

Mikhael1983:

yes, that's right.

If you try to estimate the potential effectiveness of a single pair for a selected time period, for example: the distance travelled or corridor and/or take the tick volume and point value, correlate them, then you get just such proportions of lots depending on the selected design (the force of action should equal the force of counteraction). I use a period of 30 to 60 minutes for the calculations.
 
Renat Akhtyamov:
And then we're back to the martin again

No one has thought of it any other way:) Either we either go long on a divergence (martin) or we go long on a profitable 'leg' (anti-martin). But the end is always the same:)

 
b2v:

No one has thought of anything else:) Either we either go long on a divergence (martin) or we go long on a profitable "leg" (anti-martin). But the end is always the same:)

I was hanging the real state.

no martin, and there were only two MAs

One on the chif, one on the eu.

lots are the same

That's the whole system.

increased the deposit 22 times in 3 months

 
Renat Akhtyamov:

I was hanging the real state.

no martin, and there were only two MA's

one on the chiff, one on the eu.

I believe that on certain successful pairs one can pick up a period of wagons or other oscillators and trade for a few months.
If I had an owl on MT5, which would survive at least 2-3 years in the tester without painful optimization, that would be interesting.
 
b2v:

What the author is right about is that maths is probably hard to push on a forum. It's understandable, since not everyone has a degree in maths, physics, MIT, etc.:)

The author is pushing mathematical fallacies. Which are perceived by even less educated citizens as a revelation from above)
 
Renat Akhtyamov:

that is:

EURUSD - GBPUSD = EUR/GBP


This is correct, but it should be made clear that it is not a ratio but something else and the result will not be a ratio.

dEURUSD - dGBPUSD = dEURGBP

dEURGBP + dGBPUSD = dEURUSD

dEURUSD - dEURGBP = dGBPUSD

 
secret:
...by less educated citizens as a revelation from on high)

Less knowledgeable about the subject, to be precise. It's a bit of an insult to citizens).

 
secret:
The author pushes mathematical fallacies. Which are perceived by even less educated citizens as a revelation from above)

More and more participants are coming to an unpopular but correct conclusion.

Reason: