I have received a reply from Oanda's help desk regarding this, which I quote below for forum members' information. "The partialFill = DEFAULT value currently serves a limited purpose. It indicates that partial fill is possible if the order is marked ...

What is "partialFill" on Oanda's v20 API?

Are you 100% sure about that? 'All positions on your account at the end of each trading day (5pm ET)' would seem pretty explicit about open positions at 5pm ET and not any position that had been open in the previous 24 hours but closed by 5pm ET. ...

OANDA Discussion

Just happened on the end of this thread while looking for something else and was surprised by complaints about financing rates. Have any of the complainers actually read the .pdfs for each instrument? I quote: "All positions on your account at the ...

OANDA Discussion

On the matter of data mining bias, I'd like to draw the board's attention to url and a blog about it here -> url . It would seem that this approach would be a valuable tool to have.

Machine Learning with algoTraderJo

I am pleased that this thread has now moved on to NN as this is an area I am particular interested in. As a hopefully useful contribution I would draw readers' attention to url where a pdf file is available with some really useful tips with regard ...

Machine Learning with algoTraderJo

A quite simple introduction to the idea of DMB tests is summarized at url The tests alluded to in previous posts are closest to the test called "method 3" in the link.

Machine Learning with algoTraderJo

Starting from page 33, some more good reading on data mining bias url

Machine Learning with algoTraderJo

A good read on testing for data mining bias here:- url

Machine Learning with algoTraderJo

Well, it would depend on what you want to test, but generally a bootstrap with replacement would be used to measure the variability of a statistic of interest, e.g. confidence intervals. See url for a more detailed description.

Machine Learning with algoTraderJo

Imagine a bag with 1000 marbles in it, where each marble is a single return. Bootstrap with replacement is taking one marble out, recording its value, and then putting it back in the bag, and repeating this another 999 times. Random permutation is ...

Machine Learning with algoTraderJo

An interesting online article here: url ; substitute data mining bias for correlation it then shows how difficult things can be when trying to account for it.

Machine Learning with algoTraderJo

PipMeUp asked Well, the two offered by Aronson in his book are: The expected return from a "best performing system" equals zero The "best performing system" is devoid of predictive power, i.e. is randomly correlated with future market behaviour ...

Machine Learning with algoTraderJo

Because somehow I can't upload images to the forum and I don't want to hijack this thread, I have written a short post about data mining bias on my my blog at url . I hope readers find it interesting.

Machine Learning with algoTraderJo

I have created a repo on Github at url with MATLAB/Octave code for various data snooping tests, as well as uploading some academic papers on the subject.

Machine Learning with algoTraderJo

Rather than go on a lengthy exposition about the tests, I would refer you to my earlier links and also add White's original paper, url , and the R ttr-tests package url which implements it as well as some other tests. I also have some zipped ...

Machine Learning with algoTraderJo

Conduct a White's Reality Check test and a Monte Carlo Permutation test, both of which are explained in this book - url and companion website here - url Basically what you would do is detrend and randomly re-order your time series data set and ...

Machine Learning with algoTraderJo

A very interesting thread, which I have just spent the last few hours reading. I particularly like the move to hourly bars to avoid the complications of being limited by the constraints of daily bars. Also, the idea of using excursions from the open ...

Machine Learning with algoTraderJo