On the unequal probability of a price move up or down - page 21

 

Some pseudo-traders on the internet teach pair trading like this. Although it cannot be called paired trading, it does not smell of hedging.


 
khorosh:

I would recommend trading pairs ( EURUSD, EURGBP) or (GBPUSD, EURGBP), they have better cointegration.

As far as I understand, the idea is slightly different, not to search for cointegrating symbols on average on history, but to find some conversion coefficients such that a pair of assets, in our case EURUSD and GBPUSD, shows local cointegration on a relatively short timeframe - about several days, i.e. to enter and exit a trade. And then the coefficients are recalculated. In my opinion, a much more reasonable approach than looking for global cointegration.
 
Closed the deals. The profit was $137. The divergence has not yet closed to zero, so I have closed a little earlier than I would have. But as I said, I'm not at the computer now, and do not want to watch the deal. So I closed it already. I'll show pictures of how it all looked tonight.

The result is that the third trick succeeded as well as the first two. Whether it was an accident or not is up to everyone to decide.

 
Mikhael1983:
Closed trades. Profit is $137. The divergence has not yet closed to zero, so I closed a little earlier than I would have. But like I said, I'm not at my computer right now, and I don't want to keep track of the trade. So I closed it already. I will show you the pictures tonight, how it all looked like.

The result is that the third trick succeeded as well as the first two. Whether or not it was an accident is up to everyone to decide.

3 trades in profit is not that much of a stat to talk about some "probabilities" from the branch name.
The combinationists can show a series of 40-50 in a row in one direction.
But as a "magician", I recommend to perform 28 currency pairs in one test in order to build your confidence. Just like you did with three (two) pairs, do the same trick on 28 pairs. 14 trades will do it roughly. Then you will see if all of them are guaranteed to be in the black.
 
The probabilities in the thread title aren't about that at all. I started showing the tricks somehow by accident, out of boredom at the weekend, and the idea is that the topic of the thread is stated before the tricks started. There is a chasm from the probability considerations at the beginning of the thread to these tricks, although the latter has a direct bearing on the former.
 
khorosh:

In simple terms, the ability of the spread to return to zero on a regular basis. This is what is important for pair trading, not the magnitude of the pair's correlation. The correlation should be considered only in the sense of whether it is forward or backward. It is necessary to determine whether to enter in the same or different directions.

Once in 10 years a person appears, I mean the author of the theme, in the end it all comes down to chewing over the same thing and getting bored, but some people have hope shining through the cracks of hopelessness...

to the author of the topic, make an owl and don't sweat it.

Evra


))))))))) Best answer from old Sherlock Holmes, said it all in one sentence (:o)

 
Mikhael1983:
The probabilities in the thread title isn't about that at all. Tricks I started to show somehow accidentally, out of boredom at the weekend, the idea is that the topic of the branch is set out before the start of tricks. From the stated considerations of probabilities at the beginning of the branch to these tricks is a gulf, in fact, although the latter is directly related to the former.

Yusufkhoja was more interesting, not him you are not attracted to his formula that was the bomb 10 years ago)))

 
Martingeil:

Once every 10 years a person comes along, I mean the author of the thread, in the end it all comes down to chewing on the same thing getting boring, but some people have hope shining through the cracks of hopelessness...

to the author of the thread, make an owl and don't sweat it.


))))))))) Best answer from the old Sherlock Holmes, said it all in one sentence (:o)

the author of the topic is a bit right, in that for X/Y , Y/Z and Z/Ysimultaneously the up and down probabilities are equal only if the "masses" X=Y=Z and the instantaneous reaction are the same. But the masses are unknown and the market forms buffers - i.e. the reaction will be delayed and may not happen at all.

 
Maxim Kuznetsov:

The author is slightly correct in that for X/Y , Y/Z and Z/Y simultaneously the up and down probabilities are equal only if the "masses" X=Y=Z and the instant reaction are the same. But the masses are unknown and the market forms buffers - i.e. the reaction will be delayed and may not happen at all.

Let it add Pi = 3.14, range = (X+Y+Z)/3.14 then let it work with the golden ratio 0.618 or 1.618 - it works, and I do not know what he is thinking there, I prefer my own cockroaches)))

 
Martingeil:

Let it add Pi = 3.14 , range = (X+Y+Z)/3.14 then work with a golden section of 0.618 or 1.618 - it works, and what he does there I do not know I prefer my cockroaches))))

"Work with 0.618 1.618" in most cases is to admit that it is the sum of many random and independent quantities. And there's nowhere to dig, but nowhere to cheat (or cheat, from qualification) :-)