Money Flow Index (MFI): Volume-Weighted RSI for Detecting Overbought and Oversold

The Money Flow Index is a volume-weighted RSI. It measures both the direction and the force of money flowing into and out of an instrument — rejecting price moves that lack volume support. A price spike to overbought on RSI without volume behind it is ignored by MFI. That selectivity is the edge.

Section 1: Core Mechanics

MFI takes the typical price of each bar, multiplies it by volume (producing "money flow"), then sorts those bars into positive and negative buckets. The ratio of positive to negative money flow is converted to a 0–100 oscillator using the same mathematical structure as RSI.

What Is MFI?

Developed by Gene Quong and Avrum Soudack, MFI adds the missing dimension to RSI: volume. RSI tells you a stock rallied strongly over 14 days. MFI tells you whether that rally was driven by genuine buying power or thin-air price movement. MFI above 80 on heavy volume = confirmed overbought. MFI above 80 on low volume = potentially unreliable signal.

Formula

The full calculation chain:

Step 1 — Typical Price (TP):

Step 2 — Raw Money Flow (RMF):

Step 3 — Classify each bar:

• Positive Money Flow:
• Negative Money Flow:
• Neutral (no change): excluded from calculation

Step 4 — Sum over N periods:

Step 5 — Money Ratio:

Step 6 — MFI:

Inputs

• High, Low, Close: Used to calculate Typical Price for each bar
• Volume: Weights each bar's money flow contribution
• Period (N): Default 14 bars

Parameters

Output Range

MFI always falls between 0 and 100:

• Above 80: Overbought — heavy volume at high prices
• Below 20: Oversold — heavy volume at low prices
• 40–60: Neutral zone — no strong directional bias

Visual Behavior

MFI plots as a line oscillator in a separate pane. It moves between 0 and 100. Overbought (80+) and oversold (20−) zones are marked by horizontal lines. The indicator oscillates around the midpoint of 50.

Section 2: Interpretation & Signals

Signal Zones

The Core Advantage Over RSI

RSI treats a 1-point price gain on 1 million shares the same as a 1-point gain on 100 million shares. MFI multiplies the same 1-point gain by 100× more volume in the second scenario — producing a dramatically higher money flow reading. Low-volume price moves barely register in MFI. High-volume moves dominate the oscillator.

Divergence Signals

Bullish Divergence: Price makes a lower low, but MFI makes a higher low. The second selloff was on lighter volume — sellers are exhausting. Entry on the reversal candle after MFI crosses back above 20.

Bearish Divergence: Price makes a higher high, but MFI makes a lower high. The second rally was on lighter volume — buyers are retreating. Entry on the breakdown candle after MFI crosses back below 80.

False Signals

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

Critical repaint behavior — read this carefully.

MFI's Typical Price uses (High + Low + Close) / 3. All three inputs — High, Low, and Close — update continuously during a live bar. As a result:

• MFI for the live bar changes with every tick throughout the session
• An MFI reading of 22 at 10:00 AM can become 35 at 2:30 PM on the same bar
• An MFI showing "oversold" mid-session may close at a neutral 45

This is not the same as "repaint" on historical bars (which MFI does NOT do). This is live-bar calculation updating as inputs change — completely normal behavior. But it means any MFI reading you see during a live bar is provisional and should never trigger a trade entry.

Rule: Never act on MFI until the bar has closed and the reading is confirmed. Set alerts on bar close, not on real-time value.

Section 4: Practical Use Cases

Real example: SPY during the October 2023 selloff — the daily MFI dropped to 14 on October 27, 2023, while price tested the 200-day moving average at $408. The extreme oversold MFI reading coincided with volume running 2.5× the 20-day average (heavy capitulation), and SPY reversed sharply from that level. Price rose from $408 to $479 over the next 10 weeks. The MFI oversold signal at a major support level with elevated volume was the entry indicator.

Section 5: Pseudo Code

INPUT: high[], low[], close[], volume[], period=14, ob=80, os=20

PROCESS:
  Step 1: Calculate Typical Price
            tp[i] = (high[i] + low[i] + close[i]) / 3

  Step 2: Calculate Raw Money Flow
            rmf[i] = tp[i] * volume[i]

  Step 3: Classify each bar
            if tp[i] > tp[i-1]: flow_type[i] = "positive"
            elif tp[i] < tp[i-1]: flow_type[i] = "negative"
            else: flow_type[i] = "neutral"

  Step 4: Sum over lookback period
            pos_mf[i] = sum(rmf[j] for j in range(i-period+1, i+1) if flow_type[j] == "positive")
            neg_mf[i] = sum(rmf[j] for j in range(i-period+1, i+1) if flow_type[j] == "negative")

  Step 5: Calculate MFI
            if neg_mf[i] == 0:
              mfi[i] = 100  # All positive money flow
            else:
              money_ratio = pos_mf[i] / neg_mf[i]
              mfi[i] = 100 - (100 / (1 + money_ratio))

  Step 6: Generate signals
            if mfi[i] < os and mfi[i-1] < os and mfi[i] > mfi[i-1]:
              signal = "Oversold Reversal"
            elif mfi[i] > ob and mfi[i-1] > ob and mfi[i] < mfi[i-1]:
              signal = "Overbought Reversal"

OUTPUT: mfi[], signal[]
EDGE CASES:
  - All bars in period are negative MF: neg_mf = all, pos_mf = 0, MFI = 0
  - All bars in period are positive MF: pos_mf = all, neg_mf = 0, MFI = 100
  - Live bar: flag mfi[current] as provisional until bar close
  - Zero volume bar: exclude from money flow sums, carry forward neutral classification

Section 6: Parameters & Optimization

Standard MFI Period Conventions

Threshold Adjustments

Section 7: Synergies & Conflicts

Section 8: Common Mistakes

Section 9: Cheat Sheet