Basic Research and statistical analysis example for Noobs

[anon81929759]

Wow, what a ‘piece’ - must be important as there is a shed load of data here

yeah i know IT’S WEIRD HUH…
Normally you’d expect to see a wall of text like this from me

@_bob
you’ve done good mate
i haven’t had a chance to read all this, but i will, but @Jacqui appreciates is so you’ve done well… good job

[Falstaff]

Wow, what a ‘piece’

Yes it is important that the newbies know how to do these things.

Thank you @_bob

I downloaded a load of 1 minute data from dukascopy, back to 2003 and I find there are some minutes missing (Data issues, not weekends etc)

I’m looking to write a macro probably, to put in new lines and duplicate previous lines, until the data returns (usually 1-4 minutes) - do you have ay issues with missing data from your source ?

[eddieb]

@_bob
Great thread bob, well set out in bite-sized chunks as well so its not overwhelming for newbies

[_bob]

@Jacqui

Thank you for sharing a valuable resource/insight!

Thanks @Jacqui. When we talk of hard work this is what we talk about. But like our trading it is personal therefor subjective. We each have to tailor our research to our own needs. Hopefully this well help the chosen few on their way.[quote=“Falstaff, post:22, topic:135938”]
do you have ay issues with missing data from your source
[/quote]

Can’t comment to much on this bro. Yes I have also noticed this missing data and there is even examples of here

Donchian Channel Trading

Just smash this out on the EURUSD. It’s getting late here so just the results and a bit of light discussion. All tests on the 1 hour chart at 1% risk unless stated otherwise. First chart break set at 20, sl at 10 to trail immediately [image] Next a tweak to 24 for the break and 9 sl set to trail immediately [image] Next again a break of 24, but this time back to sl set at 2x ATR. Trailing stop set at the 9 channel and trails once the level has passed the stop loss. [image] Now this lo…

and a chart here as well

Donchian Channel Trading

A picture paints a thousand words [image] Thats just a gap in my data. 2 days out of 2000 trading days. Hard to think that would effect the results too significantly.

In my observation data errors occur more in the earlier years than recent history. Statistically speaking IMO it’s an error of less than 0.1% making it insignificant to our results. But then most of my work is done on the EURUSD so I haven’t examined the data to much with other pairs.

Also IMO it isn’t really worth investigating back much bast 2011. Again this is my view, I believe the world markets changed after 2008. New technologies, algo trading, and HFT entered. Therefor data analysis any further back is probably not reflect of future market conditions.

@eddieb

Great thread bob,

Thanks my friend. You left a gold standard yourself, pity no-one has been active there for a while.