From theory to practice - page 45

 
Mikhail Dovbakh:

We know what increment is. We also know what sampling thinning is.
And what does averaging of quotes mean in this case?

Can you describe formally what is taken as the discredit interval in Beautiful's histogram?

о)

Artificially introduced bias, usually

 
Nikolay Demko:

The forum allows you to attach files to your messages.

.xls cannot be attached, but they can be zipped.

For clever people I personally do not spare ANYTHING. Use it!
Files:
EURJPY.zip  6826 kb
 
Renat Akhtyamov:

An artificially induced error is usually


I haven't seen your histograms yet, so I don't know what kind of error you are artificially injecting into them.
That is not what we are discussing. Not speculation, but fanoman.

 
Mikhail Dovbakh:

I haven't seen your histograms yet, so I don't know what kind of error you're skillfully entering into them.
That's not what we're discussing. It's not speculation, it's fanomania.

It's not me who's inputting, it's the author.

the whole phenomenon is that at exponential time intervals the exponent is in the form of a histogram.

nothing phenomenal at all.

Moreover, if the price is flat, the histogram will be a bell; if it is in a trend, it will look more like a half of a bell

analyzing the EURJPY quote, it is now flat

As it needs to be proved - the probability of error is close to 100%

 
Mikhail Dovbakh:

We know what incremental sampling is. Sampling thinning too.
But what does averaging of quotations mean in this case?

Can we formally describe what is taken as a discredit interval in Beautiful's histogram?

о)


Pulse generator gives a signal to read quotes in exponential intervals. If a new quote has come, fine, we will use it in our calculations, if not, we won't.

Then we simply take the average value between two consecutive quotes.

I assure you - no one, no broker is able to distort the picture we need with this method of reading.

 
Alexander_K2:

The pulse generator gives a signal to read quotes at exponential intervals. A new quote arrives - fine, we use it in our calculations, if not, we do not.

Then we simply take the average value between two consecutive quotes.

I assure you - no one, no broker is able to distort the picture we need with this method of reading.

So the "increment" is the difference between these averages between exponentially distant quotes?
 
Mikhail Dovbakh:
i.e. the "increment" is the difference between these averages between quotes exponentially removed in time?
Exactly right.
 
Alexander_K2:
Absolutely right.

Thank you.
I don't agree with Renat that this will obviously automatically give the right distribution.
But it's worth thinking about.
And in my remarks about the strobe, I'm just stating my case... At the limit are we waiting for the next flash (new value), in a few years?)
Or will it all re-initialise somehow?

 
Mikhail Dovbakh:

Thank you.
I don't agree with Renat that this will obviously automatically give the right distribution.
But it's worth thinking about.
And in my comments about the strobe light, I'm only getting approved.... At the limit are we waiting for the next flash (new value), in a few years?)
Or is it somehow re-initialised?

It has a frequency there, meaning it adds up roughly the same quotient and most likely averages the result again

With that in mind....

That's why I don't see the point in making a histogram, because I have already drawn it in my head.

Of course from the bar chart you can understand where the price went.

BUT!

It is very difficult to find the turning point in the most profitable place in such calculations.

And the approach using ticks becomes clear.

IOW

At large sample sizes and time intervals the error is off the charts

 
Renat Akhtyamov:

I'm not the one who's putting it in, it's the author.

the whole phenomenon is that in the exponential time intervals the exponent will appear as a histogram

absolutely nothing phenomenal actually

If the price is flat, the histogram will be a bell; if it is trending, it will look more like half a bell

I'm analyzing the EURJPY quote, it's flat

as it needs to be proved - the error probability is close to 100%

Provide an illustration of what has been said. Fortunately the history of these plots is easy to pull up.
By a bell you mean Student's, Cauchy's or Gauss and who else?
They're all bell-like as it were. I'm also curious where you saw"looks like half a bell".

Now that's definitely an artefact phanomenon. And the original source of the data here would also be useful for analysis...
Please attach it.