ema_stoploss

8 hours ago

Library "ema_stoploss"

What it does
A small library that builds stop-loss levels from dynamically computed EMAs. It finds EMAs strictly on the desired side of price (long: below; short: above), sorts them by distance to price, and returns the n-th nearest as your stop.

How it works

sortEMAsByDistanceStrictDyn(signal, lensArr, src, ascending)

Computes each EMA internally with the alpha formula (alpha = 2/(len+1)), so you can pass a dynamic array of lengths.

Strict side filter:

signal = 1 → only EMAs < src (below)

signal = -1 → only EMAs > src (above)

Sorts candidates by distance to src (default: nearest → farthest) and returns two arrays: EMA values and their lengths.

get_stop_loss(index) (exported)

Builds a default length array: 21, 50, 100, 200, 250, 500, 750, 1000.

Long side uses low to find the index-th nearest lower EMA.
Short side uses high to find the index-th nearest upper EMA.

Returns [sl_buy, sl_sell].

Plots

Stop-Loss Long (green): the selected lower EMA (based on low).

Stop-Loss Short (red): the selected upper EMA (based on high).

Input

Index (default 2): 0-based.
0 = nearest, 1 = second nearest, 2 = third, etc.
If there aren’t enough EMAs on the required side, the function returns na (no plot).

Why internal EMA calc?
ta.ema() doesn’t accept a series length; by updating each EMA with its alpha step every bar, the library supports arbitrary dynamic length arrays and stays bar-consistent.

Customize

Edit the list in get_stop_loss() to use your own EMA lengths.

Change ascending in sortEMAsByDistanceStrictDyn if you prefer farthest → nearest.

Use a different src if needed (e.g., close, hlc3, etc.).
The example intentionally uses low for long stops and high for short stops.

Notes

Strict side filtering: EMAs on the wrong side are ignored (no fallback).

If no EMA qualifies on a side, you’ll get na for that side.

Complexity is O(n²) for sorting, which is negligible for small EMA lists.

8 hours ago

Release Notes

v2

fix na