Williams Alligator

Three forward-shifted smoothed moving averages — the Jaw, Teeth, and Lips — used to visualize when a market is trending versus 'sleeping'.

What it is

The Williams Alligator is a trend-visualization tool developed by Bill Williams and popularized as a default indicator in MetaTrader 4 and 5. It plots three smoothed moving averages (SMMAs) of the median price (high plus low, divided by two) directly on the price chart: the Jaw (13-period, shifted 8 bars into the future), the Teeth (8-period, shifted 5 bars forward), and the Lips (5-period, shifted 3 bars forward). Williams framed the tool with a metaphor — when the three lines are intertwined the alligator is 'sleeping' (a range), and when they fan apart in order it is 'eating' (a trend). Strip away the metaphor and it is three moving averages of different lengths, each displaced to the right. That forward displacement is the detail most explanations skip: the line drawn under today's bar was actually calculated several bars ago, so the 'current' Alligator partly describes the past. The Alligator is descriptive, not predictive — it summarizes recent price behaviour and cannot forecast what comes next. This is educational material only, not financial advice, and all trading carries risk of loss.

How it works

Each line is a smoothed moving average — Wilder-style smoothing, the same running average that powers RSI and ATR, equivalent to ta.rma in Pine Script — applied to the median price (high + low) / 2. The Jaw averages 13 bars, the Teeth 8 bars, and the Lips 5 bars, so the Lips reacts fastest and the Jaw slowest. Each line is then plotted with a forward offset: the Jaw 8 bars ahead, the Teeth 5, the Lips 3. This shift is purely visual — it moves already-computed values to the right on the chart; it does not compute anything about future bars. The practical consequence matters for anyone coding or backtesting it: the Alligator value displayed under the bar you are looking at was calculated from data ending 3 to 8 bars earlier. A naive backtest that reads the plotted (shifted) line as if it were known on that bar is comparing today's price against an average that, in live trading, would have been sitting several bars to the left. When the three averages are close together, recent price has been directionless; when the Lips pulls away from the Teeth and the Teeth from the Jaw, shorter-term averages are separating from longer-term ones — the standard signature of a directional move that has already begun, since every moving average lags by construction.

How traders read it

• Intertwined lines: when Jaw, Teeth, and Lips braid around each other, the different-length averages are roughly equal — recent price has been range-bound. Williams called this the alligator 'sleeping'.
• Fanning apart in order: Lips above Teeth above Jaw (all rising) is read as an established upward move; the mirror image is read as a downward one. Because all three are lagging averages, this configuration confirms a move already underway rather than anticipating it.
• Line crossings: the Lips crossing back through the Teeth and Jaw is often read as the trend phase ending — again a lagging observation, not a forecast.
• Distance between lines: wider separation reflects a stronger recent directional push; narrowing separation reflects fading momentum in the averages.
• Mind the shift: the plotted lines are displaced 3-8 bars forward, so the value under the current bar is stale by that many bars. When coding rules, decide explicitly whether you are comparing price to the shifted (visual) or unshifted (real-time) value — they give different results.
• Treat every reading as context alongside price structure and risk management, not as an instruction to act.

Common settings

The standard configuration is Jaw = 13-period SMMA shifted 8 bars forward, Teeth = 8-period SMMA shifted 5 bars forward, and Lips = 5-period SMMA shifted 3 bars forward, all applied to median price (hl2). These are Bill Williams' original values and the MT4/MT5 defaults; they are conventions rather than optimized numbers. Some traders lengthen all three on higher timeframes for smoothness. Note that some platforms substitute an EMA or SMA for the SMMA, which produces visibly different lines from MetaTrader's — worth checking before comparing values across platforms.

Strengths

• Makes the range-versus-trend distinction easy to see at a glance without reading raw numbers.
• Built into MetaTrader by default and easy to replicate exactly in Pine Script with ta.rma, so cross-platform values can be verified.
• Uses three timescales at once, giving more context than a single moving average.
• Transparent math: despite the metaphor, it is fully deterministic and simple to audit.
• The smoothed (Wilder) averages are less jumpy than same-length EMAs, reducing visual noise.

Pitfalls to watch

• The forward shift misleads: the line under the current bar was computed 3-8 bars ago. Backtests that read the displaced value as if it were available in real time are peeking at a stale-but-shifted series and will not match live behaviour.
• Triple lag: SMMA smoothing is slow to begin with, and the visual displacement adds apparent delay on top — by the time the lines fan apart, a meaningful part of the move has typically already printed.
• Whipsaw in ranges: in sideways markets the lines braid and un-braid repeatedly, generating many false 'awakening' patterns.
• Metaphor is not mechanism: 'sleeping' and 'eating' describe line geometry, not any measured market microstructure — the vocabulary can create false confidence.
• Settings and source sensitivity: changing lengths, offsets, price source (hl2 vs close), or the smoothing type (SMMA vs EMA) changes every reading; platform implementations differ.
• No predictive power: the Alligator summarizes past averages of past prices. It does not forecast anything, and no indicator removes the risk of loss.

Pine v6 example

//@version=6
indicator("Williams Alligator Example", overlay = true)

jawLen   = input.int(13, "Jaw Length",   minval = 1)
teethLen = input.int(8,  "Teeth Length", minval = 1)
lipsLen  = input.int(5,  "Lips Length",  minval = 1)
jawOff   = input.int(8,  "Jaw Offset",   minval = 0)
teethOff = input.int(5,  "Teeth Offset", minval = 0)
lipsOff  = input.int(3,  "Lips Offset",  minval = 0)
src      = input.source(hl2, "Source")

// SMMA (Wilder smoothing) is ta.rma in Pine v6
jaw   = ta.rma(src, jawLen)
teeth = ta.rma(src, teethLen)
lips  = ta.rma(src, lipsLen)

// offset shifts the plot visually; the values are NOT known that early
plot(jaw,   "Jaw",   color = color.blue,  offset = jawOff)
plot(teeth, "Teeth", color = color.red,   offset = teethOff)
plot(lips,  "Lips",  color = color.green, offset = lipsOff)

Pro tip: When translating Alligator rules into code, drop the visual offsets and work with the unshifted ta.rma values — that is what would actually have been known on each bar. If a backtest of an Alligator idea looks dramatically better when the shifted lines are used, that improvement is lookahead, not edge. This is educational information only, not financial advice; no indicator predicts price, and all trading carries risk of loss.

Educational only — not financial advice, not a recommendation to trade. No indicator is predictive; trading involves substantial risk of loss.