My EA's test results - what's the catch and illusion?

 

Good afternoon.

I guess a lot of people are already going on holiday etc. But I hope some experienced TRADERS will be able to comment critically on the testing results of my Expert Advisor for 2007-2013. I'm sceptical about it myself, expecting experienced people to advise - what's the catch of the next variant. I think that's the end of my attempts to invent the "grail" :)

The lot size did not change within the testing period. Stepwise increase of lot size leads to "flight into space" - tens of millions of dollars :)))

Thanks in advance to all who will write on the merits.

 
concord99:

Good afternoon.

I guess a lot of people are already going on holiday etc. But I hope some experienced TRADERS will be able to comment critically on the testing results of my Expert Advisor for 2007-2013. I'm sceptical about it myself, expecting experienced people to advise - what's the catch of the next variant. I think that's the end of my attempts to invent the "Grail" :)

The lot size did not change within the testing period. Stepwise increase of lot size leads to "flight into space" - tens of millions of dollars :)))

Thanks in advance to anyone who will write on the merits.



Not much data on TC. At the moment, there is only one place that can help:

https://www.mql5.com/ru/forum/133408

 
you are most likely using a very close stop loss (or close trailing stop). Minutes are already inadequate if the stop is less than 10p (4 digits), and tick emulation is irrelevant.
 
Avals:
you are probably using a very close stop loss (or close trailing stop). Minutes are already inadequate if the stop is less than 10p (4 digits), and tick emulation is irrelevant.

Thanks. I too think the reason is the stop loss on the tester. I use hourly charts but, the stop loss is only 1 pip + spread.
 
concord99:

Good afternoon.

Thank you in advance to everyone who will write on the merits.

From what can be seen without psychics:

1) The last year and a half TC has actually earned nothing. Not a good trend.

2) There is a period of about a year where the TS just plummets, will you wait a year and suffer losses on a real account?

3) Profitable trades are less than plum ones, maybe not a big deal, but 44K trades, out of none really giving profit (stepping up) no more than 50. Could be a fit and a "happy" coincidence

4) Well and get rid of mismatch errors (re-pump history, recalculate frames, check holes, etc.)

concord99:

I use hourly charts, but, stop loss is only 1 pip + spread.
Well you can't test this in the tester at all, unfortunately....
 
Figar0:

From what can be seen without psychics:

1) For the last year and a half, TC has actually earned nothing. It's a bad trend.

2) There is a period of about a year where the TS just plummets, will you wait a year and suffer losses on a real account?

3) Profitable trades are less than plum ones, maybe not a big deal, but 44K trades, out of none really giving profit (steps up) no more than 50. Could be a fit and a "happy" coincidence

4) Well and get rid of mismatch errors (re-pump history, recalculate frames, check holes, etc.)

Well you can't test such things in the tester at all, unfortunately....

Thanks for informative reply. Please clarify - why cannot I test iton hourly charts with a stop loss of only 1 pip + spread ? What parameters should be maintained?
 
concord99:

Thanks for the informative reply. Please clarify why I cannot test onhourly charts with a stop loss of 1 pip + spread ? What parameters need to be maintained?

Not on the hourly chart, but just with such short take and stops. There are no Bids and Asks in the story, no spreads accordingly. There is only OHLC, and :

1) the tester itself generates a tick sequence, and the stop is tick-sensitive.

2) It generates according to a quite strictly specified algorithm. Some time ago I accidentally created the Grail using tester ticks, because a simple NS could partially catch this algorithm

3) to test on old data, when the spread, for example, EURUSD was 2-4 points at 4 (20-40 at 5) with the current 7-10 points - is incorrect. And the tester takes the current values of the trading environment. In 2007, you simply could not place a stop at 1 point. Stops were much larger there, and the quotes were correspondingly slightly different. There is no unified quote in Forex, the brokerage companies deliver what they can afford according to their trading conditions.

The tester is a good tool, but for more "thick-skinned" strategies that don't touch ticks. I will use it to analyse the trade history of the market and also to analyse the market perceptions.

 
Figar0:

Not on the hourly chart, but just with such short take and stops. There are no Bids and Asks in the story, no spreads accordingly. There is only OHLC, and :

1) the tester itself generates a tick sequence, and the stop is tick-sensitive.

2) It generates according to a quite strictly specified algorithm. Some time ago I accidentally created the Grail using tester ticks, because a simple NS could partially catch this algorithm

3) to test on old data, when the spread, for example, EURUSD was 2-4 points at 4 (20-40 at 5) with the current 7-10 points - is incorrect. And the tester takes the current values of the trading environment. In 2007, you simply could not place a stop at 1 point. Stops were much larger there, and the quotes were correspondingly slightly different. There is no unified quote in Forex, the brokerage companies deliver what they can afford according to their trading conditions.

The tester is a good tool, but for more "thick-skinned" strategies that don't touch ticks. The minimum you should lower in the tester is minute OHLC


Thank you very much for the information. I think I'm unlikely to come up with anything else.

It's impossible to beat the casino and you'll have to accept that idea. Sooner or later you have to finally give up the illusion of "flying" :(

 
concord99:


Thank you very much for the information. I think it's unlikely that I'll think of anything else.

It's impossible to beat the casino, and I'll have to accept that idea. Sooner or later you have to give up the illusion of "flying" :(

It's not about the "casino", it's about the discrepancy between the tester quotes and the real ones.

Where in the tester are the requotes, slippages, "no prices", "no connection"?

 
concord99:


Thank you very much for the information. I think I am unlikely to come up with anything else.

It's impossible to beat a casino, and you have to accept that idea. Sooner or later you have to give up the illusion of "flying" :(

IMHO, one of the axioms in this business - do not take the word of anyone. Check everything and draw your own conclusions.

About your system - try loading real ticks into tester and test on them - there will be more sane results. Otherwise you can't test a pip tester.

 
Dima_S.:

IMHO, one of the axioms in this business is not to take anyone's word for it. Check everything and draw your own conclusions.

About your system - try to load real ticks into the tester and test on them - there will be more sane results. Otherwise the pip tester won't test.



The thing is that I'm not a pipsitter in the sense of loss-profit ratio. My P&L ratio is about 2 (it depends on spread) / 150-1500, i.e. Pipsing is out of the question. Please, advise me, where can I get real quotes?