The Weighted Moving Average assigns linearly decreasing weights to older bars - the most recent bar has weight N, the next has N-1, and so on down to 1.

Formula

WMA[t] = (N*close[t] + (N-1)*close[t-1] + ... + 1*close[t-N+1])
       / (N + (N-1) + ... + 1)

Params

period - N in the formula above.

Output

Single column named after your indicator (e.g. 20-bar WMA).

When to pick WMA

You want more responsiveness than SMA without the recursive smoothing of EMA.
You want predictable, bounded weight decay (linear rather than exponential).
You're porting a TradingView script that explicitly uses WMA.

In practice EMA is more common in crypto + futures strategies. WMA shows up in classical TA stacks (Hull MA, Welles Wilder smoothing).

Pitfalls

WMA only knows the bars inside its window - it has no recursive memory the way EMA does. Drop-off effects are real but smaller than with SMA because the dropping bar already had the lowest weight.