From theory to practice - page 23

 
Alexander_K2:

Note what the EURJPY Ask increments are coming in this week

About the same comes for Bid.

And this data is from a real NDD/ECN account! All the same, it becomes obvious that DCs are applying some filters on the output to broadcast tick quotes.

So how to fight this? The answer is with the simplest input filters.

For example, we take the average value between two successive quotes, we obtain for the same dataset:


I am now addressing to the "doers" from DC - that's how it is, uncles, nothing will help you! :)))))) I don't think you understand either that "memory" in non-Markovian processes is not easy to destroy. Your school knowledge of physics is not enough for that!


Good afternoon again! Have you tried to do the same thing po ticularly? The thing is that dealing centres are unlikely to do the same data sampling as you do. Most likely they manipulate with ticks - I make this assumption on the basis of the fact that tick charts often show a saw - a 1-2 point movement up and down, with changes occurring every tick and usually such movements are performed without any visible benefit and only change the statistics on volumes. If you try to build a similar histogram for a sample of all ticks and in the periods of maximum market activity, for example during the American session, you may see many interesting things. Regards to your researches!

 
Yuriy Asaulenko:
Actually, masquerading is not accepted here. And it's not welcome.

Yuri, seriously - I didn't want to go back. But reading some of the posts here on the forum touched me to the core. I realised that I couldn't just quit like that. Please excuse me and let's get on with it. OK?

 
Alexander_K2:

Yuri, seriously - I didn't want to go back. But reading some of the posts here on the forum touched me to the core. I realised that I couldn't just quit like that. Please excuse me and let's get on with it. OK?

I confess, I thought the subject was over. Well, if it's not, it's not. Let's see episode two.)
 
Yuriy Asaulenko:
I admittedly thought the topic was over. Well, if it's not, it's not. Let's watch the second series).

Most accurate comment.Alexander_K2 second series )))

 
Yury Kirillov:

Good afternoon again! Have you tried doing the same thing on the web? The point is that it's unlikely that dealing centres do the same data sampling as you do. Most likely they manipulate with ticks - I make this assumption on the basis of the fact that tick charts often show a saw - a 1-2 point movement up and down, with changes occurring every tick and usually such movements are performed without any apparent benefit and only change the statistics on volumes. If you try to build a similar histogram for a sample of all ticks and in the periods of maximum market activity, for example during the American session, you may see many interesting things. With respect to your research!

But still the question of reading tick data remains.

I will repeat what I said earlier in this thread. We consider the solution of the Fokker-Planck equation, augmented with an integral term, by numerical methods. To do this we need to understand the physical meaning of EVERY parameter in it:

1.probability density W(x,t) - the probability that at a certain time t the price MAY take a certain value. And, for any sample size, price values lie in ranges defined by Chebyshev's inequality.

2.x-price is the value of Ask or Bid at time t. Moreover, itis exactly the tick quotes, i.e. if there were no deals, the price is not read and does not participate in calculations with the lapse of time.

3. time t. No, I would even sayTIME t. This is the most important parameter! I've read a lot of topics here - how to organize the reception of tick data, whether each tick is important, etc., etc. If the formula contained just W(x) - then yes, I'd have to receive every tick, and if W(x(t)) - then it would read the price with a certain frequency, regardless of whether it was a real tick or not. But, we have exactly W(x,t), which means thatboth of these approaches to receiving data are wrong. The correct algorithm for reading quotes is as follows - choose a constant t = 1 second and read tick quotes with this frequency - this will be correct.

I would like to add:

the method of reading ticks may be different - but it's important to understand WHY you are reading it this way and not otherwise. For arbitrage strategies or for trading on incremental returns the method of reading EVERY tick is preferable.

But, we are solving a serious integro-differential equation. There must not and cannot be a double interpretation of parameters! Otherwise it will lead us to defeat.

So, I insist that tick quotes are read according to a certain algorithm:

1. at equal time intervals (e.g. t=1 sec.)

2. at exponentially distributed intervals

Just read it that way, through "ragged" time intervals - no, it is not suitable for numerical methods of solving the equation, whatever you want to do with me.

 
Yury Kirillov:

Good afternoon again! Have you tried doing the same thing on the web? The point is that it's unlikely that dealing centres do the same data sampling as you do. Most likely they manipulate with ticks - I make this assumption on the basis of the fact that tick charts often show a saw - a 1-2 point movement up and down, with changes occurring every tick and usually such movements are performed without any apparent benefit and only change the statistics on volumes. If you try to build a similar histogram for a sample of all ticks and in the periods of maximum market activity, for example during the American session, you may see many interesting things. With respect to your research!

That's why I say you have to work with Last. The flippers do not have such a problem.

And why do you think DTs do it? Maybe it's a market maker in a thin market looking for which price counterparties will fall for.

When the market maker tambourines in the thin market no real trading is going on, he just moves his pending orders.

 
Yuriy Asaulenko:
I confess, I thought the subject was over. Well, if it isn't, it isn't. Let's watch the second series).

Have we solved the problem yet? Not yet. It's very serious, even though I try to stick to a clear and non-academic style of presentation.

 
Nikolay Demko:

That's why I say you have to work with Last. The flippers don't have that problem.

And why do you think that DCs do it? Maybe it is a market maker in a thin market looking for which price the counterparties will fall for.


And what is the law of distribution of intervals between Last? Uniform? Exponential? I'm sure it is neither. And you can't just use it like that.

 
Yury Kirillov:

Good afternoon again! Have you tried doing the same thing on the web? The point is that it's unlikely that dealing centres do the same data sampling as you do. Most likely they manipulate with ticks - I make this assumption on the basis of the fact that tick charts often show a saw - a 1-2 point movement up and down, with changes occurring every tick and usually such movements are performed without any apparent benefit and only change the statistics on volumes. If you try to build a similar histogram for a sample of all ticks and in the periods of maximum market activity, for example during the American session, you may see many interesting things. With respect to your research!

Let me explain a little - this small saw is an option of the server part (at least for MT4 - I don't know about MT5). It is done to provoke the "trader" that watches the price, to make a deal. It is clear that this has nothing to do with real pricing (even if pseudo-real forex).
 

Here you go, friends - please help me at least once in a while.

Please plot histograms of time intervals between ticks and post them for all to see. OK?