[zazagura]

Some of the papers I've seen are ridiculously obviously over-fitted.

For example published in 2018, but "tested" on 3 months of 2010 prices of GBP/USD, USD/SEK and USD/THB. Quality forex data is so easy to get freely, that picking 3 months from 8 years ago on one major pair and two other random minor ones just stinks.

[quirmian]

Where can I get good quality forex data from?

[capnrefsmmat]

I used TrueFX for an assignment in a statistics-for-finance class I taught: https://www.truefx.com/

[jobigoud]

They don't test the predictive power of their models?