Ultimate Oscillator: Multi-Timeframe Momentum Confirmation

The Ultimate Oscillator solves the single-period problem: RSI gives different signals depending on which period you choose. The Ultimate Oscillator runs three RSI-like calculations simultaneously (periods 7, 14, and 28) and combines them with weighted averaging. The result is a 0–100 oscillator that is more stable, generates fewer false signals, and requires multi-timeframe consensus before triggering.

Section 1: Core Mechanics

The Ultimate Oscillator was created by Larry Williams in 1976 and published in Stocks & Commodities magazine in 1985. Williams observed that RSI signals varied dramatically based on period selection and that traders spent too much time optimizing a single parameter. His solution: make the oscillator inherently multi-timeframe by baking three periods into the formula.

Formula

For each bar, calculate intermediate values:

Then for each period (7, 14, 28):

Finally, the weighted combination:

The weights 4, 2, 1 give double emphasis to the shortest period, creating responsiveness while the longer periods prevent false signals.

Why Buying Pressure Uses True High/Low

Standard oscillators use only the current bar's high and low. If a stock gaps down, the standard low misses the gap's impact. By anchoring True High and True Low to the previous close, BP and TR capture gap behavior — a downward gap reduces BP (buyers had to overcome the gap to push price up) and increases TR (the effective range is wider).

Inputs

• High, Low, Close for each bar, plus previous close
• Periods: 7, 14, 28 (Larry Williams' fixed design — these are not meant to be optimized arbitrarily)

Parameters

Output

UO outputs 0–100, like RSI. Above 70 = overbought. Below 30 = oversold. The 50 line is the neutral centerline. Unlike RSI, UO rarely spends extended time at extremes because three periods must simultaneously agree that momentum is extreme.

Section 2: Interpretation & Signals

Signal Zones

Williams' Divergence Rules

Larry Williams specified exact conditions for valid divergence signals on the Ultimate Oscillator. These rules are more selective than generic RSI divergence — and that selectivity is the edge.

Bullish divergence signal (4 conditions, all required):

1. UO falls below 30 (oversold extreme is reached)
2. Price makes a lower low — UO makes a higher low (standard bullish divergence)
3. UO rises above the highest high it reached between the two price lows
4. Enter long when UO breaks above that intermediate high

Bearish divergence signal (4 conditions, all required):

1. UO rises above 70 (overbought extreme is reached)
2. Price makes a higher high — UO makes a lower high (standard bearish divergence)
3. UO falls below the lowest low it reached between the two price highs
4. Enter short when UO breaks below that intermediate low

The key difference from RSI divergence: Williams requires UO to actually reach the extreme zone (below 30 or above 70) before the divergence is valid. Many RSI divergence setups trigger at RSI 40 or 60 — nowhere near an extreme. Williams' approach filters for genuinely exhausted momentum.

Best Market Conditions

The Ultimate Oscillator excels in markets with clear trending bias on at least two of its three timeframes. It is less useful in tightly ranging markets where the three-period consensus creates conflicting readings. In strong momentum moves, UO can remain above 60 or below 40 for extended periods — do not fade this without a complete divergence setup.

Section 3: Pass vs. Live — Real-Time Reliability

UO calculates BP and TR from the previous close and current bar data. Once a bar closes, all values for that bar are fixed. No historical values change. The live bar's UO updates as current high, low, and close evolve. The 28-period sum means intraday swings have a smaller percentage impact on UO than they do on a single-period RSI — more stable on the live bar.

Section 4: Practical Use Cases

Real example — Apple (AAPL), early 2023: After AAPL's bear market decline, the daily UO dropped to 26.4 (below 30) on January 5, 2023, at a price of $125.02. A second low on January 19, 2023, printed $132.12 — higher price low — while UO registered 28.1 (still below 30 but higher than 26.4). This met Williams' first three conditions. UO then broke above its intermediate high of 38.5 on January 20. Entry at the close of January 20 captured a move from $135 to over $170 by the end of February.

Section 5: Pseudo Code

INPUT: high[], low[], close[], short=7, medium=14, long_=28, ob=70, os_=30

PROCESS:
  Step 1: For each bar i > 0:
            true_high[i] = max(high[i], close[i-1])
            true_low[i] = min(low[i], close[i-1])
            bp[i] = close[i] - true_low[i]  # Buying Pressure
            tr[i] = true_high[i] - true_low[i]  # True Range

  Step 2: For each valid bar i >= long_ - 1:
            avg7 = sum(bp[i-6:i+1]) / sum(tr[i-6:i+1])
            avg14 = sum(bp[i-13:i+1]) / sum(tr[i-13:i+1])
            avg28 = sum(bp[i-27:i+1]) / sum(tr[i-27:i+1])

  Step 3: Ultimate Oscillator:
            uo[i] = 100 * (4 * avg7 + 2 * avg14 + avg28) / 7

OUTPUT: uo[] — values in range 0-100
EDGE CASES:
  - tr[i] == 0 for all bars in a period window: avoid division by zero;
    if sum(tr) == 0 for a window, set that average to 0.5 (neutral)
  - First bar (i=0): no previous close — skip; UO starts from bar index 1
  - Fewer than 28 bars of data: return NaN
  - NaN in high, low, close: propagate NaN

Section 6: Parameters & Optimization

Williams' Period Relationships

The 7, 14, 28 periods are in a 1:2:4 ratio. Each period is exactly double the previous. This geometric relationship means each timeframe is measuring a fundamentally different cycle length — not overlapping. The weights (4, 2, 1) invert this ratio, giving the most weight to the shortest period while the longer periods anchor the reading.

Parameter Impact

Section 7: Synergies & Conflicts

Section 8: Common Mistakes

Section 9: Cheat Sheet