Real work on MT5 NDD - page 16

New comment
 
Interested in any market research.
 
Anybody? All right, then. Here's an example of the surge in public interest in action-price strategy. Where it's coming from again and how to make it all work. It's a perfectly normal topic of research. The point of the game is to determine how widespread these strategies are and at what levels and how profitable it is to play against them. Or here, I'm not the only one, but many have already noticed that companies have started paying dividends on shares. Before, many did not pay, they invested their profits in development, but now they have started to do it. Why would that be and where the market is headed in the long term.
 

Unfortunately, I have hardly ever been involved in stock market analysis, apart from statistical arbitrage. Therefore I can't even squeeze anything out at this stage on the topics you have touched upon.

Moreover, in your formulation it all comes down to beating other players like you. For some (cockroach in the head) reason I don't like it. So I analyze the primitive FOREX.

Having traded stat. arbitrage, having synthesised synthetic stakes came to the simple conclusion that price is a probabilistic event. Yes, it is an old idea, but in case of noisy statistical arbitrage it is clearly seen in practice.

Correspondingly, I investigate the dependences of the probability of the appearance of the price at a certain level in the course of astronomical time (and whether these dependences exist) and with the flow of ticks (the number of ticks as a measure of time). So far I come to the conclusion that thick tails are a problem of out-of-noise analyses. And since you can make a good noise due to your own aggregation, it makes sense to examine it.

I also practice scalping, but it is a special case of statistical arbitrage, where the same principles apply to higher timeframes.

In general, I chose static arbitrage only because of its superprofitability and low risks, i.e. purely for pragmatic reasons. Of course, other topics are also of interest.

 

Also the topic of determining the lag-overrun of different feeds is very interesting. For example, you synthesise a stack and it is not clear from which source the prices are most relevant and which are already from the past. I.e. who is leading. At the same time, the presenters may switch roles. It is also interesting that for Bid there may be one master, and for Ask - absolutely different one.

From the mathematical point of view the problem is reduced to the analysis of price BPs. From the practical point of view it is interesting at least for the simple reason. It is necessary to choose the broker, who gives the actual prices before others. The lags occur up to 200-300 msec (already taking into account ping), i.e. very decent.

It is interesting to discuss the algorithms themselves for determining the master and slave.

Unfortunately, all this is so specific that it is unlikely anyone will do it. Because you either have to be just an enthusiast, or understand perfectly well that you can squeeze quite a lot of money out of it.

 
hrenfx:

It is interesting to discuss the master and slave identification algorithms themselves.

Unfortunately, this is all so specific that it is unlikely that anyone will do it. Because you either have to be an enthusiast, or understand perfectly well that you can squeeze a lot of money out of it.

No, all this has long been known and practiced. Once again I draw your attention to usdak. A synthetic tumbler is a faint semblance of a left-handed nasdaq. So, there's the notion of "axe" - the current leading MM and there are methods of identifying the axe (they're hiding, the bastards, they know it's dangerous to stick out in the open) and playing with it. There are a lot of scalper (and not only) techniques on the tumbler, more precisely on Level2, etc. And there are both manual and automatic strategies there. I am sure that in the currency departments of banks (no one will let us there), which are majors in forex, there are exactly such strategies and approaches, because they have their own synthetic tumblers. All in all, my opinion, you need to look there. But, again, you have to remember the difference between synthetic betting and nasdaq quota-mounting.
 

The trick with FOREX, as opposed to NASDAQ, is that some market participants are unaware of each other's existence. They are scattered in different places, have their own local platforms and so on. At NASDAQ, everything is centralized. Even classical arbitrage is practically impossible because everything is consolidated in one place. And only the one who brings it together can use this technical insider, i.e. NASDAQ itself. Therefore, synthetic stakes in FOREX are far less polished and have far more inefficiencies.

Regarding algorithms, please share the links.

 
hrenfx:

In general, I chose statistical arbitrage only because of its super-profitability and low risk, i.e. purely for pragmatic reasons. Of course, other topics are also interesting.

As I was assured by people who say, that they are in the subject, almost all trendy HFT strategies - the essence of statistical arbitrage on scalping.


In general this is certainly an interesting subject. Here is a picture for example.

US stock futures, German stocks and the Eurodollar. You can see that they move consistently and there seems to be arbitrage opportunities. But that is not the question, the question is what forces cause these financial instruments from different areas to behave in such a way. In fact, it is a question of the nature of statistical arbitrage and hence its stability.

 
hrenfx:

The trick with FOREX, as opposed to NASDAQ, is that some market participants are unaware of each other's existence. They are scattered in different places, have their own local platforms and so on. At NASDAQ, everything is centralized. Even classical arbitrage is practically impossible because everything is consolidated in one place. And only the one who brings it together can use this technical insider, i.e. NASDAQ itself. Therefore, synthetic stakes in FOREX are far less polished and have far more inefficiencies.

On the subject of algorithms, please share the links.

I don't know, at Nasdaq arbitrageurs and scalpers are quite alive, not extinct as a class. There MM has one more duty besides obligations to stand on the spread's edge - to sell/buy for their clients, and all kinds of possibilities appear right there. In principle, the axe is the one who has either foolishly (I doubt it) decided to move the market or has to place a large bid.

And I can't help you with links to algorithms. They do not advertise them, although sometimes you can find something on English-speaking forums. In our mamba trying to do the same thing, but our market is essentially manipulative, so it is unlikely fit. In general, I do not know how to help in this matter, as I myself do not practice it.

I remember, there was an old book on Nasdaq, a translation, about 20 years old, by van Tharp. It might be a good introduction, but if I'm not mistaken, it was manual trading, and it was outdated.

 

You seem to be unfamiliar with the practice of statistical arbitrage, since you talk about questions of its stability (and nature).

To be honest, it would probably be much easier to arbitrate on exchanges with their huge number of characters. Alas, lazy and conservative. In terms of algorithms everything should be the same.

Can't tell anything from the picture above at all, it's too crude - M1. I would collect ticks with lifetime from 50ms. That's where I would look.

The very search for bins for stat arbitrage lends itself to automation.

 

When I said no classic arbitrage on NASDAQ I meant Ask < Bid. On exchanges it is impossible in principle because of pricing rules identical to ECN. On the FOREX, there is STP - a feature that none of the stock exchanges has. I myself have not traded on exchanges, but logically there positive slippage at pre-existing limits should be absent as a class.

1...9101112131415161718192021
New comment