Fins in the glass - trying to understand what happened by the ticks - page 9

 
I don't understand where the local nubs get their naivety: why are they so sure that if one idiot fills a thousand contracts on the Si futures market, then the price of the rouble must change by five kopecks? - Well, it's obvious that this is nonsense. Because the SiM9 is only a small part of the rouble. There are other contracts, there is a spot and a decentralised interbank currency market. In short, a bunch of different giant markets, which are closely linked with each other. And now an idiot comes along, throws a market request, and it means that the global rouble rate will change by 5 kopecks at once because of his actions on all the markets! - This is nonsense. In reality, Sim9 in this situation was not obliged to generate a new supply or demand, it could return to the previous level without any trades and limit orders, because the general macroeconomic situation does not determine the rouble rate, but rather it is Sim9. It's time to finally understand that and close this topic.
 
Vasiliy Sokolov:
I do not understand where do the local newbies get such a naivety: why do they believe that if an idiot pours a thousand contracts on the market on Si futures, then the ruble price must change by five kopecks? - Well, it's obvious that this is nonsense. Because the SiM9 is only a small part of the rouble. There are other contracts, there is a spot and a decentralised interbank currency market. In short, a bunch of different giant markets, which are closely linked with each other. And now an idiot comes along, throws a market request, and it means that the global rouble rate will change by 5 kopecks at once because of his actions on all the markets! - This is nonsense. In reality, Sim9 in this situation was not obliged to generate a new supply or demand, it could return to the previous level without any trades and limit orders, because the general macroeconomic situation does not determine the rouble rate, but rather it is Sim9. It's time to understand that and close this topic.

Theorist, let's not get emotional - if you have a justification, say it, otherwise you look like a preacher with a mass about a fair and infallible exchange.

I have yet to see a justification for the position, and the justification could either be logic to explain the lack of limits or excerpts from exchange regulations.

The topic is an important one, at the very least it speaks to the inadequacy of testing against exchange ticks as MT5 does.

 
Don't get too excited, guys!
 
Yes, DMA gives more opportunities than conventional connectivity, that's what they pay money for; for hft strategies they rent servers in the same centres where the exchange is, hire teams of programmers and data scientist;). This is not a level for single traders. Although, the way the mixing works and how the market microstructure is organized, how hft-business is structured, you need to have an idea just to not waste your time on futile attempts to create profitable strategies at this level. Read/google #hft, #dma, #orderlog and you will understand everything.
 

like this

 
Sergey Lebedev:
In details: the limit order of a trader not using the hft connection for online t&s analysis to sell 65822 was noticed by the hft strategy, which analyzes the market, and a counter limit order was generated, which led to the registration of this transaction in the first frame of the 12th second. Without entering it in the market! That is the essence of hft-strategies: prediction of the real close price, and buying out of 'lagging' bids. That is, theft strategy evaluated the volume traded at the 11th second and 'knowing' that after such volume, the market price is in the 65870/65880 area, it generated counter limits for all 'cheap' sales. And then re-sold them at a fair 'market' price, which the strategy promptly valued. The other market participants did not see these bids in the cup at all. This is, in a nutshell, the essence of hft trading.
P.s.: By 'knew'/'estimated' you mean the result of neural networks.

As far as I know, limiters are not matched to limiters. Limiters always converge to a marker. And HFT is not about being able to pander to someone else's limiters, but about taking advantage of the speed at which you get information from the exchange and do something with your bids, whether limit or market. In other words, if you are "closer" to the exchange, you have more opportunities.

 
prostotrader:

:) It's not my strategy.

It appeared as soon as futures were introduced.

If dividends are not assigned, the dividends exhibited (counted) tend to exceed the real ones,

so there is simply no entry into the market (but the market doesn't "go on the rails", there are, not often, that there are breakdowns).

And profitability depends on market participants assessing the situation correctly.

I always put an entry into the market 3-5% above the bank deposit offer.

I always trade 56 pairs of instruments, someone always "takes a shot".

Added

I don't trade now at all, I'm all out on MOEX


Do you somehow take into account the possible volatility of the price of the asset for the period of open positions?
 
Andrey Gladyshev:

As far as I know, limiters are not matched to limiters. Limiters are always matched to a market. HFT is not about being able to match your limiters with someone else's, but about taking advantage of the speed at which you get information from the exchange and do something with your bids, whether limit or market. That is, if you are "closer" to the exchange, you have more opportunities.

You understand correctly. The author is highly imaginative, to put it mildly)

 
rjurip:

Do you somehow take into account the possible volatility of the asset price for the period of open positions?

It does not take into account at all. The positions are multidirectional (futures sold, stocks bought), i.e. volatility is irrelevant.

In an expiration, I will simply be "poured" negative stocks instead of futures (as if I sold a stock) in a quantity equal to the purchased stock

and the result of the positions will be 0. (the picture shows shares in lots in 1 lot of 10 shares)


Added

in 1 contract MOEX-6.19 100 shares -151 * 100 = -15100 shares (in the contract futures)

in 1 lot of MOEX shares 10 shares -1510 * 10 = +15100 shares

At expiration: 15100 + (-15100) = 0

 
prostotrader:

It does not take into account at all. The positions are multidirectional (futures sold, stocks bought), i.e. volatility is irrelevant.

In an expiration, I just get negative stocks instead of futures (as if I sold a stock) in a quantity equal to the purchased stock

and the result of the positions will be 0. (the picture shows shares in lots of 1 lot of 10 shares)



Yes, but if the price of the asset moves more than the expected dividend, won't you lose?