How to get REAL and TRUE history data ?

 
Hi,
Is there somewhere to get the actual real account pricedata for a specific broker/account type, for offline backtesting?
My recent EURUSD 'superscalper' turns 500 into 2Mill in a year on Alpari demo, using the standard history data.
Problem is that these data is largely useless for optimizing a scalper. Same thest with same spread on a different
broker's demo and the account was lost in a month... The same would most likely happend if I ran it on Alpari Live.

So what is needed is a historical ( atleast a few months) log of the actual price data strem for a speciffic account/pair/broker/account type.
I would be most interested in data from a Alpari UK std account, or possibly a mini. Primarily EURUSD.

Are these data available somewhere, or is there some other solution to this problem ?
 
This thread may help https://www.mql5.com/en/forum/120787
I use the data on separate MT install without setting up demo account. IE, <esc> out of the first two pop-ups (open an account + login). But - is way to fake it all using dummy proxy server etc. then not have to <esc> out the two pop-ups. https://www.mql5.com/en/forum/120787
hth
.
proxy just for those times of stress
https://www.mql5.com/en/forum/121644
 
Well, the Dukascopy stuff looks interesting and its on download now.
But I assume these data will only be good for scalping on THEIR accounts, and pretty useless on Alpari which most likely have their feed filtered differently?



About disconnecting:
The way I do it is simply to click the connection indicator (lower right), choose login, and hit RightArrow and 0 and Enter. This takes less than 2 secs.
This will 'log in' to an invalid account (having an extra 0 in the account number) and we're promptly disconnected.
 
DayTrader:
Hi,
Is there somewhere to get the actual real account pricedata for a specific broker/account type, for offline backtesting?
My recent EURUSD 'superscalper' turns 500 into 2Mill in a year on Alpari demo, using the standard history data.



I use Alpari Uk as well. Check the spread using the "SYMBOL PROPERTIES" button in the strategy tester window. I just checked mine.

1.6 pips on EURUSD on the Live account

0.7 pips on EURUSD on the Demo account


So basically if you backtest on the demo account you will get better results. Do the words "FINANCIAL FRAUD" spring to mind? They are advertising a service with a defined commision rate (the spread) and then charging you more. This is not just a one off. I have monitored the spreads over days and the demo account is always better during live hours. I have no doubt that this is deliberate policy to encourage you to trade. I find it despicable.

At the weekends the spread ranges can swap over. Ths is a real pain for back testing because a strategy can work one day and then not the next. The spread can get stuck in some unrealistic value over the weekend when we are coding the hardest!


 
Actually a while ago I asked them about this spread difference, and they said that the spread on the demos comes from the Pro account... you know the 'you have to trade 1Bill/month' type of thing....
Yes I think it would be more sensible to have the mini or std sperad reflected on the demo.

Anyhow you can easily adjust the spread to whatever you want for testing, and I find this quite useful.
  But scalpers are extremely sensitive to both the spread and pricepatterns they are optimized to, essentially
  to the point where they are impossible to get working properly on a live account.
Right now I have the spreadproblem under control but are struggeling with backtesting on REALISTIC data.
 
DayTrader:
... backtesting on REALISTIC data.

You realize that ticks in the Tester are interpolated? There is nothing 'Real' about them. This is fine for strategies that do not rely on tick level data, but will greatly affect strategies that do (for better or worse...). I suspect that your strategy is the latter, in that case getting 'better' data is not going to help u.

 
DayTrader:
Actually a while ago I asked them about this spread difference, and they said that the spread on the demos comes from the Pro account... you know the 'you have to trade 1Bill/month' type of thing....
Yes I think it would be more sensible to have the mini or std sperad reflected on the demo.

Anyhow you can easily adjust the spread to whatever you want for testing, and I find this quite useful.
But scalpers are extremely sensitive to both the spread and pricepatterns they are optimized to, essentially
to the point where they are impossible to get working properly on a live account.
Right now I have the spreadproblem under control but are struggeling with backtesting on REALISTIC data.

How do you adjust the spread when testing? I can't see anything clickable in there to edit it.


I log the tick data using a script. It produces 1Mbyte of CSV files per symbol per day. There's this one particular short term trade type that I am testing that won't simulate well on invented data so I am logging the data in case I feel keen enough to do something with it.

 
dabbler:

How do you adjust the spread when testing? I can't see anything clickable in there to edit it.

You need to change the 'symbols.sel' file... Search the forum for that term.

Alternatively u can wait for the spread to change and disconnect when u like it.

 
gordon:

You need to change the 'symbols.sel' file... Search the forum for that term.

Alternatively u can wait for the spread to change and disconnect when u like it.

Thanks.

https://www.mql5.com/en/forum/113730

Wow, that is hardly straightforward! Hacking the binary symbols.sel file with its unknown file structure. Yuk.

Saving a "good" symbols.sel for the weekend is a neat trick.

Reason: