On the unequal probability of a price move up or down - page 9
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
My advice to you is not to overindulge in New Year's Eve drinks.
Thank you.
It's not quite like that. More precisely, not at all. Let me show the trick in real time.As of 00:00 Moscow time on 03.01.2020 let's plot the last half of the day in the M5 tf, namely the Eurodollar (ED) and the Pounddollar (PD).
In some shamanistic way we also plot two additional curves: EDq and PDq. Having the following properties: their correlation coefficient corr(EDq, PDq) = 1. Strictly, let us note, not approximated, but bluntly exactly 1.
Thus, they can serve as reference curves (give us a point of reference, so to speak). What we see: that ED is more downwardly deviated from EDq than PD is from PDq. I won't write the numbers, you can see everything on the charts.
Conclusion: we open a pair of deals: buy EURUSD on volume 8, and simultaneously sell GBPUSD on volume 3 (why - because the volatilities of EDq and PDq are so correlated). And we can easily wait for profits.
P.S. I foresee an explosion of cries that there will be no profit. But the trick, let's note, is shown in real time. Let's see.
And if we construct some "shamanistic" way the reference curves in such a way, that such divergence in data is not observed? And then what about the objectivity of the estimate? What is the point of creating reference curves? Everyone can construct them as they want, and see what they want to see.
bzdyn - for two arbitrarily taken majors (maybe with crosses too, but I fiddle with majors) you can build such a curve, that abs(Pearson) 0.75-0.9 almost always . But not forever and it's sudden :-)
Actually all topics are whirled to reduce this "almost" to "exactly" that one and to take the profit out of correlations in large periods that two. Judging by the abundance of topics it is a real nut
PS/ and there is great suspicion that the approach and formulas of correlations are not suitable. Either the discreteness of the values or the time, but something is wrong somewhere :-)
Maxim Kuznetsov:
... something is wrong somewhere :-)
Waiting. State of play as of 3:10am Moscow time.
Went to bed around 7am Moscow, only came back to look at the market now.What do we see (not half a day, but a day):
The misalignment did not decrease, but increased. Well, no one has forbidden that, of course. Let's open once again in the same direction: at a ratio of lots 8 to 3, respectively, EURUSD - buy, GBPUSD - sell, and wait for profits.
P.S. Strictly speaking, the profit was already there - just a small one - because the mismatch was small when opening deals - and it was possible and necessary to close, but I was sleeping at that time. This moment (when the black curves came close to the red ones corresponds to the reference i approximately = 180 on the charts above).
Went to bed around 7am Moscow, only came back to look at the market now.What do we see (not half a day, but a day):
The misalignment did not decrease, but increased. Well, no one has forbidden this, of course. Let's open once again in the same direction: at a ratio of lots 8 to 3, respectively, EURUSD - buy, GBPUSD - sell, and wait for profits.
P.S. Strictly speaking, the profit was already there - just a small one - because the mismatch was small when opening deals - and it was possible and necessary to close, but I was sleeping at that time. This moment (when the black curves came close to the red ones corresponds to the reference i approximately = 180 on the charts above).
It's necessary to make trades on the maximum divergence and close them on a convergence
Of course. And the magnitudes of the "discrepancies" are statistically clear, it's not hard to represent all this directly in terms of percentages of previously observed discrepancies. Say, to enter the market at discrepancies such that 80% of the time on the history they are smaller. However, I'm not trying to maximize profit; my aim is to show the trick. The main thing is not the optimality, but the very fact of profit. I opened the first pair of deals shortly before I posted here describing the reasons for doing so. Then I filled it up. Then I just opened another pair and saw a significant divergence. At the moment my six trades look like this (I also use stoploops about 100-150 pips 4 signs, which I shift from time to time as prices move):
So the important thing is that there will be a profit, not how optimal the market entry was)
P.S. The server time in the screenshot is 1 hour less than Moscow time.
We should make trades on the maximum divergence and close them on the convergence
Convergence does not guarantee profit, it can still be minus. Also an interesting point about swaps, but not about that
You have to make trades at the maximum divergence and close at the convergence
If you can determine the moment of maximum divergence, then what prevents you from determining the maximum or minimum of the price - the tasks are almost identical, then pair trading will not be necessary).
If you can determine the moment of maximum divergence, then what prevents you from determining the maximum or minimum of the price - the tasks are almost identical, then pair trading will not be necessary).