Discussion of high-frequency trading on MT5 - page 64
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
MT4 + Level 2 in terminal (fx+++n) time M1 and level 2 based FDK (fx+++n), compared to the implementation with FDK, the stack in MT4 is a slow turtle.
As a result the volumes are different, i.e. it makes no sense to use any level 2 based TS in MT4.
MT5 fx+++n doesn't have it, sorry not to compare, but most likely the situation won't be particularly advantageous.
the glass in MT4 is a slow turtle.
But it's not native, it's a third-party product. the mt4 has nothing to do with it.
The programmers may have set it to 500 ms in the debug stream and then forgot to reduce it.
But it's not native, it's a third-party product. MT4 has nothing to do with it.
The programmers may have set the sleep in the debug stream to 500 ms and then forgot to reduce it.
it's a little weird.
MT4 server + terminal stand locally. i.e. as you have 0 ping.
Yes, locally.
Up to 400 is including sending to the pvp. i.e. 150-200 for sending and executing on the check. depending on the number of reply packets and ping.
Yes, locally.
450 is the execution time of OrderModify function, i.e. nothing is sent to STP.got it. i tested two MT4 servers. on different servers and both had ~150 without sending to STP.
Maybe they are such powerful physical servers?...but i don't think so.
Tested different MT4 servers, never got even close to 150ms. Maybe, I've got into heavily loaded trading servers, but unlikely.
In fact, on LIVE in F***N on MT4 it's ~450ms (sometimes about a second (no re-connect, I guess)), FDK ~15ms. And you should also take into account that prices in FDK are also coming faster than in MT4. Therefore the real trading difference is even bigger.
Tested different MT4 servers, never got even close to 150ms. Maybe, I've got into heavily loaded trading servers, but unlikely.
In fact, on LIVE in F***N on MT4 it's ~450ms (sometimes about a second (no re-connect, I guess)), FDK ~15ms. And you should also take into account that prices in FDK are also coming faster than in MT4. Therefore the real trading difference is even bigger.
...
2. the influence of large bids on price behaviour. large bids can be a way of manipulation on one hand (in which case it is worth trading against such bids), and on the other hand, an inefficient change in position size, revealing to the market information about someone's expectations (in which case it is worth fronting such bids).
If I understand correctly, in some approaches (based on Obizhaeva-Wang model or model of Alfonsi, Fruth and Schied) optimal execution (trading for large volumes) implies exactly the execution of relatively large orders at some points in time. If so, the question arises of how widely these approaches are used and how they are masked by other algorithms using fake bids (for example, as in Chameleon in Cortex iX).
Again, maybe I'm missing something. It all turns out rather complicated.
Guys, why are you raving about some stupid article. It's all for I don't know who, geeks probably. You want to know what algorithms the big uncles actually use, look at how bids were placed and what was going on during the darkpool crash - what they were doing is even easier than chintz pants at rub twenty. Black and white, it must have been said ten times in this thread that the advantage there is in iron, in pings, but not in matan.
Uh, you just blew it.)
you're ruining the whole circus.)