Sadly it repaints. I accidentally deleted the before pic but where you see the lines there have been a confirmed closed oscillator signal, then you refresh the chart and it changes everything kind of like a tma

Mescalito wrote: Mon May 12, 2025 12:06 pm Sadly it repaints. I accidentally deleted the before pic but where you see the lines there have been a confirmed closed oscillator signal, then you refresh the chart and it changes everything kind of like a tma

he is just having trouble making it NRP

I think I got it now. Actual regressions between degrees dev. can start if so as next.

ZigZag wrote: Mon May 12, 2025 7:29 pm I think I got it now. Actual regressions between degrees dev. can start if so as next.

it's still not good.

you need to pass it throught the strategy tester, and then put another instance on the backtested chart and compare both

both instances have to be exactly identical
once you got that it's all good

by the way, are you sure the regression is causal ? because it's in the same type of non causale regressions like Nadaraya-Watson and whatnot

Meanwhile, something to read.

https://trendspider.com/learning-center ... works-cnn/

I actually realized today metatrader does not have matrix support exist, but for for machinelearning there is a way, to use .CSV file to train Phyton Panda module for kernel regressions and metatrader can read this .CSV file.