Williams Accumulation/Distribution: calculate WAD from price

The WAD formula

Let C be the current close, P the previous close, H the current high and L the current low. Set trueHigh = max(H, P) and trueLow = min(L, P). These are range boundaries, not an ATR average or a divisor.

WAD = previous WAD + contribution. A zero seed defines the baseline; the output is cumulative price units, not a percentage or an oscillator with fixed overbought levels. The current open is absent from the formula. TradingView documents these rules for ta.wad.

Five bars you can calculate by hand

This synthetic OHLC dataset starts both cumulative comparisons at zero after bar 0. Bar 0 supplies the previous close only. Volume is included solely to calculate the Chaikin column; changing it leaves every WAD value unchanged.

Plot WAD in TradingView with Pine Script v6

Paste this into a new indicator in the Pine Editor, save it, then add it to the chart. Blue is the built-in series. Orange dots expose the explicit calculation. Read the bar contribution in the Data Window when checking a completed candle.

//@version=6
indicator("Williams AD: built-in and explicit calculation", overlay = false)

bool hasPrices = not na(high) and not na(low) and not na(close) and not na(close[1])
float trueHigh = math.max(high, close[1])
float trueLow = math.min(low, close[1])
float contribution = 0.0
if hasPrices
    if close > close[1]
        contribution := close - trueLow
    else if close < close[1]
        contribution := close - trueHigh

// Seed at zero. Skip unavailable comparisons; do not invent prices.
float explicitWad = ta.cum(contribution)
plot(ta.wad, "Built-in WAD", color = color.blue, linewidth = 2)
plot(hasPrices ? explicitWad : na, "Explicit WAD (skip missing)", color = color.orange, style = plot.style_circles)
plot(hasPrices ? contribution : na, "Bar contribution", display = display.data_window)
bgcolor(not hasPrices and not barstate.isfirst ? color.new(color.orange, 85) : na, title = "Unavailable price comparison")

The explicit accumulator starts at zero. It hides the first point because the prior close is unavailable. A later missing required price also hides the point and adds nothing; orange shading flags the unavailable comparison. If a close is missing, the following bar also lacks its immediately previous close. Accumulation resumes only when all required inputs exist. Skipped movement cannot be recovered from this script.

This missing-data policy is deliberate. Compare the explicit result with the built-in on complete standard OHLC data; do not assume identical handling of every malformed series. The code plots values and does not detect divergence, create alerts or backtest trades. Syntax validation cannot establish those behaviors.

What a WAD disagreement can and cannot tell you

• Compare matching swing dates. A lower price low with a higher WAD low is a bullish divergence description; a higher price high with a lower WAD high is bearish divergence. Define the swing rule before examining its outcomes. Neither pattern supplies an entry, stop or exit by itself.
• Confirm when information was available. A pivot requiring three bars to its right is known three bars later. Labeling the pivot retrospectively must not make a backtest enter at the earlier date.
• Expect live values to move. High, low and close can change while a bar is open. Use completed bars for rules that require final OHLC. TradingView describes this behavior in its repainting documentation.
• Keep the calculation start consistent. Adding older history can shift the cumulative baseline. Zero is an origin chosen by the dataset, not a universal buy or sell level. A daily reset would create a modified indicator.
• Missing volume is not a WAD problem; missing or altered prices are. Session selection, feed differences and synthetic candles change the inputs. Removing volume from a formula cannot remove bad price data or prevent false trading signals.

Educational calculation, not investment advice. No timeframe, divergence rule or indicator combination here has a demonstrated win rate. Any trading test needs explicit execution rules, costs and a separate period for evaluation.