Strange observation !

 

Good afternoon to all !

When testing and optimizing the EURJPY Expert Advisor on the timeframe = H4 I ran into some incomprehension. (to put it mildly)!

Outside the optimization period on the forward/backward test the Expert Advisor has shown quite a good result. See the chart.

But run in the "back side" of the sample (outside the optimization period) was not as good as we wanted it to be.

It is because the Expert Advisor has first showed a deposit growth and then a loss. However, this loss happened at the time of the ROBOT CHAMPIONSHIP 2007!

 

I have scrolled through the history on n4 and on D1 in good faith, but have not found how the history for the end of 2007 differs from the history on other plots.

My Expert Advisor works with fractals. It has no stops. Entry and exit are based on fractals appearing.

Then I optimized the Expert Advisor on a "sinking" site. But now it yields profit here but fails on the other one!

Please, share your thoughts on this problem!

 
rid писал (а) >>

I then optimised the EA on the "plum" story. But now, he's making a profit here, but is draining on the whole other story!

Please, share your thoughts on this problem!

The market is constantly changing. Otherwise someone would be able to just run MTS and bankrupt all the other forex participants. So that no one would "kill" forex (forex is a simple market and it cannot exist without the circulation of money), it has to be as unrepeatable as possible.

At one time MTS is making profits and at another time he is draining. All in all, the entire depo is spent on spread.


Although of course it is interesting. It would be nice to find what the change in the market was. I.e. the parameter that changed and its quantitative characteristics: before, during and after the sinking period.

 
It seemed to me that the EURJPY candlesticks on n4 for October-December 2007 were a bit bigger (wider) than in other parts of the history. 2007 were a bit bigger (wider) than in other parts of the story. But maybe that's just what it looked like.....
 
rid писал (а) >>
It seems to me that EURJPY candlesticks on Н4 for Oct-Dec. 2007 were a bit bigger than on the other parts of the history. But maybe it's just my imagination.....

This only appears to be the case. In fact, all the candles are the same width. :)

But their length may be longer due to high volatility. I think you should then make a filter on the ATR indicator. If it has an X parameter higher than a certain number, do not trade. And check how the filter will change the results of trading on history, not only for the time of the Championship 2007, but also for a larger history.

If this is about increased volatility, this filter should improve your system.

 
rid писал (а) >>
It seems to me that EURJPY candlesticks on Н4 for Oct-Dec. 2007 were a bit bigger (wider) than in other parts of the history. But maybe that's just what it looked like.....

What if the organisers of the Championships have the ability to influence the market prices of the global financial market during the Championships:)

 
SK. писал (а) >>

What if the Championship organisers have the opportunity to influence the market prices of the world financial market during the Championship?)

The Championship is held not at the "world financial market prices", but at the prices of meta-quotes that they can influence. And they do. Their prices are very similar to market prices, but only similar...

 
timbo писал (а) >>

The championship is not at 'world financial market prices', but at meta-quota prices, which they can influence. And they do. Their prices are very similar to market prices, but only similar...

any data on the discrepancy?

By the way it would be good if they gave a reference to the benchmark to avoid claims by participants

and that the word "indicative" does not appear anywhere.

 
Geronimo писал (а) >> and that the word "indicative" does not appear anywhere.

Unrealistic. Every brokerage company has its own sources of quotes and its own filters. Each brokerage company is a private shop which sets its own rules. Yes, if these rules give quotes that are too different from the "correct" ones, the brokerage companies will run away from it.

 
Geronimo писал (а) >>

By the way, it would be good if they gave a reference to the benchmark to avoid claims from participants

Knowing the benchmark is where the DC gets its quotes from. It's a very good way to unbundle it.

 
Prival писал (а) >>

Knowing the benchmark is where the DC gets its quotes from. It's a very good way to take it out.

How do you do that?

As for Meta Quotes Demo, they are taken from Alpari, as they themselves have stated.

Reason: