Module willr 

Williams %R

Oscillator on high/low/close over lookback (N):

%R = -100 * (HH − close) / (HH − LL)

where (HH = \max(high)) and (LL = \min(low)) over the last (N) bars. Range is typically [−100, 0]. Flat window ((HH = LL)): carry previous %R, else −50.

Default: period 14 (WillrParams::period_14).

---

Trading perspective

Region	Habit (classic, not a rule)
%R > −20	“Overbought” screen
%R < −80	“Oversold” screen
Cross back from extreme	Momentum resume / mean-reversion exit screen

vs other oscillators

	Williams %R	Stochastic	RSI
Scale	[−100, 0]	[0, 100]	[0, 100]
Inputs	H/L/C window	H/L/C + smooth	Closes only
Feel	Fast, inverted Stoch-like	Smoothed %K/%D	Wilder smooth, slower

Rough map: raw Stoch %K ≈ 100 + %R (same HH/LL idea; sign/offset differ). Prefer Stochastic when you want %D signal line; %R when you want a single fast line.

Pairs well with

• ADX / DI — only fade %R extremes when ADX is low (range); avoid fading when ADX is high.
• SMA/EMA trend filter — long setups only above rising MA, etc.
• Volume (OBV/MFI) — confirm oversold bounce with rising money flow.
• ATR stops — oscillator does not size risk; ATR does.

---

Engineering

WillrParams → willr / WillrState → willr_solution.
Batch uses WillrState end-to-end. HH/LL are amortized O(1) (sliding max/min).

Word problem

Highs 12, lows 10, close 11 for three bars with (N=3). What is %R on bar 2?

Expect: (HH=12), (LL=10), %R = (-100 \times (12-11)/(12-10) = -50).

use finance_solution::stocks::ta::{willr, WillrParams};
let h = [12.0, 12.0, 12.0];
let l = [10.0, 10.0, 10.0];
let c = [11.0, 11.0, 11.0];
let s = willr(&h, &l, &c, WillrParams::new(3)).unwrap();
assert!((s.willr[2].unwrap() - (-50.0)).abs() < 1e-12);

Structs

ValidatedWillr

Validated pack.

WillrParams

Williams %R lookback.

WillrSeries

WillrSolution

WillrState

Incremental Williams %R.

Functions

willr

willr_solution

Examples