Standard misconceptions in trying to trade in the noise (there was a "Nightmare on MT4 Street") - page 10

New comment
 
kniff 19.12.2006 19:16
>> There will be a ban for continuing 'misunderstanding' questions.

OK, topic closed.

How about a beer to Renate for a job well done on purpose? :-))
 
Renat >> :
I'm saying on the basis of my 7 years of experience "stay away from ticks (1), don't play in the noise (2), write toasty EAs (3), don't try to build strategies on anything below NN minutes (M5, M15, to taste) (4)". But would anyone believe me? The experience of this forum suggests that they won't believe me, and every month the same questions will appear.

They won't believe me, because all this has to be experienced on its own. So, good riddance!

Renat, would you mind scribbling a few lines here about how exactly you (or your colleagues) have tried to investigate price-flow noise?

Thank you in advance.

 
I bring it up again because it has not occurred to Renat for 15 years that there is a rudimentary algorithm for perfectly cleaning up the quotes in History Center. With this algorithm, all outliers and price holes automatically disappear. A perfect tick-flow is produced.

The solution is very simple:
1. We treat a quote stream from each bank or broker as a separate stream.
2. The average world price is the price at which 50% of the brokers are on one side of it and the other 50% are on the other side.
3. if the outlier appears only at one or several brokers, the algorithm automatically filters it
4. If the outlier is in more than 50% of the brokers, it becomes a real market spike

Briefly about the algorithm:
1. Each bank or broker is split in its own stream.
2. To understand when there is a break in quotations in that flow, the average time between two ticks is calculated, for example for the last 50 ticks.
3. It is assumed that the lifetime of the last tick of the stream will be 2 times more than the average time.
4. If there is no new tick in this time, it is assumed that there is a pause in the stream, and its last price is invalid and is not used in the calculation.
5. The last valid prices of all streams form a set, from which is chosen the WORLD AIDE PRICE (50% of active brokers streams are in one side, and other 50% are in the other side of this price).
6. This price is calculated again with each new tick for each new flow.
7. The Bid and Ask prices are treated as separate streams, and separate Bid and Ask world average prices are created from them.

And the LAST and MOST IMPORTANT thing, which allows experts to make REAL HISTORY TESTS:
Bars should not be built at the Bid price, but at the average price between Ask and Bid. This solves the problem of large spreads that create non-existent dependencies in the Bid of bars. And in general in MT4 and MT5 we should add the functionality of selecting between Ask, Bid and Mid bars for visualization and for strategy tests.

This algorithm is guaranteed to clear all outliers, holes and any other problems with any broker. The output is a perfect global average price which does not require filtering. The algorithm generates calm ticks despite the huge noise generated by all the banks. It reacts quickly during the jump, but only when more than 50% of all banks have jumped. Earlier, it assumes that some of the banks have created spikes and fills them.

If History Center quotes recalculated using this algorithm, the best quotes in the world would be obtained. The same algorithm can also be applied to real time quotes. It also works perfectly. By the way, many brokers and banks around the world use exactly this algorithm, and I wonder why Renat doesn't do the same in History Center for historical bars as well as in real time on MetaQuotes demo server.

PP: Excuse me for being a bad Russian, but I'm from Bulgaria.
 
Rosimir Mateev:
I bring it up again because it has not occurred to Renat for 15 years that there is a rudimentary algorithm for perfectly cleaning up the quotes in History Center. With this algorithm, all outliers and price holes automatically disappear. A perfect tick-flow is produced.

The solution is very simple:
1. We treat a quote stream from each bank or broker as a separate stream.
2. The average world price is the price at which 50% of the brokers are on one side of it and the other 50% are on the other side.
3. if the outlier appears only at one or several brokers, the algorithm automatically filters it
4. If the outlier is in more than 50% of the brokers, it becomes a real market spike

Briefly about the algorithm:
1. Each bank or broker is split in its own stream.
2. To understand when there is a break in quotations in that flow, the average time between two ticks is calculated, for example for the last 50 ticks.
3. It is assumed that the lifetime of the last tick of the stream will be 2 times more than the average time.
4. If there is no new tick in this time, it is assumed that there is a pause in the stream, and its last price is invalid and is not used in the calculation.
5. The last valid prices of all streams form a set, from which is chosen the WORLD AIDE PRICE (50% of active brokers streams are in one side, and other 50% are in the other side of this price).
6. This price is calculated again with each new tick for each new flow.
7. The Bid and Ask prices are treated as separate streams, and separate Bid and Ask world average prices are created from them.

And the LAST and MOST IMPORTANT thing, which allows experts to make REAL HISTORY TESTS:
Bars should not be built at the Bid price, but at the average price between Ask and Bid. This solves the problem of large spreads that create non-existent dependencies in the Bid of bars. And in general in MT4 and MT5 we should add the functionality of selecting between Ask, Bid and Mid bars for visualization and for strategy tests.

This algorithm is guaranteed to clear all outliers, holes and any other problems with any broker. The output is a perfect global average price which does not require filtering. The algorithm generates calm ticks despite the huge noise generated by all the banks. It reacts quickly during the jump, but only when more than 50% of all banks have jumped. Earlier, it assumes that some of the banks have created spikes and filtrates them.

If History Center quotes recalculated using this algorithm, the best quotes in the world would be obtained. The same algorithm can also be applied to real time quotes. It also works perfectly. By the way, many brokers and banks around the world use exactly this algorithm, and I wonder why Renat doesn't do the same in History Center for historical bars as well as in real time on MetaQuotes demo server.

PP: Excuse me for being a bad Russian, but I'm from Bulgaria.

The speed of light is 300,000,000 m/s the length of the circumference along the equator of the earth is 40,075,000 m, plus telecommunication costs, in total quoting once every 1-3 seconds when dealing with the world average price. In the mt terminal, quotations are collapsing at a rate of 500-1500 quotes per minute during active trading hours.

1...345678910
New comment
Reason: