I've been trying to port over a Tradingview strategy for a few weeks now and struggled immensely. Something I realized is obvious to others - that the chart vs. data inputs make it difficult. So I figured out a way to get my preferred indicator to be aligned after a lot of debugging.  At the core of one part is a 14 period ADX.  

I had to build it with a manual warmup and update and rolling windows from consolidators.

So I started with a 29 period window (2*period + 1)… and the ADX values didn't match… 

After a lot of frustration I just said “ok how about 100 member window”… and it got me values within 0.01 of what I had at the same moment in tradingview… but for the sake of the system I wanted to cut it back.  60 seemed good and well above any threshold I could see… but somehow it gives me different values.

So the question, how on a 14 period ADX does a 60 member rolling window give me different results than 100? Makes no sense to me. Both work OK in backtesting and while I know the default should be "run it on the 100 as it's safer' why????