Money Flow Index (MFI): Formula, 80 and 20 Levels, MFI vs RSI, and a Real S&P 500 Example

The Money Flow Index (MFI) is an oscillator from 0 to 100 that compares the money flow of the bars on which the typical price rose with that of the bars on which it fell, over the last 14 bars. Created by Gene Quong and Avrum Soudack, it is, as StockCharts puts it, a volume-weighted RSI: the RSI formula applied to price times volume instead of price changes. Readings above 80 are usually called overbought and readings below 20 oversold.

How the Money Flow Index is calculated

Each bar has a typical price, the average of its high, low and close, and a raw money flow, typical price times volume, which StockCharts notes is essentially dollar volume. The flow counts as positive when the typical price is higher than on the previous bar and negative when it is lower; when the typical price is unchanged, the flow is not used. The money flow ratio is the sum of positive flows over 14 bars divided by the sum of negative flows, and the RSI formula turns it into an index. StockCharts calls 14 the setting recommended by the creators.

How to read the MFI

StockCharts gives the three basic signals Quong and Soudack identified: overbought and oversold readings, bullish and bearish divergences, and failure swings at 80 or 20. It adds that the creators suggested expanding the extremes to 90 and 10 to qualify signals, since moves beyond them are rare, and that MFI can stay overbought while prices keep rising in a strong uptrend, or oversold while they keep falling.

The two volume signals combine in practice. In a bullish divergence, price makes a lower low while the MFI makes a higher low: less money flowed on the falling bars of the second decline than of the first. The failure swing often comes with it, and StockCharts calls the pair a stronger signal than either alone. Failure swings are one of the three signals StockCharts attributes to Quong and Soudack, and it describes the bullish one this way: the MFI dips below 20, bounces, holds above 20 on its next pullback, then rises through the peak it made in between. The bearish versions mirror these at 80. The diagram below draws the bullish case.

The MFI on a real chart

The chart below the introduction shows MFI(14) on the daily S&P 500 ETF from December 19, 2024 to June 13, 2025, with dividend-adjusted prices. The window opens at an MFI of 33.5 and a close of 572.88. MFI went below 20 on a single day, March 11, 2025, at 19.9 and a close of 545.21, and was back at 25.6 the next day. The ETF rose 3.8% over the next 10 days, then fell to a close of 488.38 on April 8, 10.4% below the oversold reading. During that April slide, the MFI never returned below 20: its low was 23.6 on April 7, as the 14-bar window still held enough rising bars to keep it above 20. It did not go above 80 either; its high, 78.6, came on May 20, at a close of 583.18. At the end of the window, the ETF was at 587.26, 7.7% above the March reading.

The RSI(14) on the same bars, below, read the April slide differently. It went below 30 on March 11 and 13, then again on April 4, 7 and 8, with a low of 21.6 on April 8, the day of the lowest close. Over those April days the RSI called the market oversold while the MFI stayed above 20: on April 7, the 14-bar MFI window still held five rising bars, March 19, 24 and 25 and April 1 and 2, worth about 164 billion dollars of money flow against 529 billion on falling bars: a ratio of 0.31, above the 0.25 that puts the MFI at 20. The MFI weighs each bar by its dollar volume, not by the size of its price move, while the RSI took the 4.9% and 5.9% falls of April 3 and 4 at full size. Neither line marked the turn by itself; the RSI was simply nearer its own extreme.

Is MFI better than RSI?

They mostly say the same thing. On the daily S&P 500 ETF from February 1993 to October 2026, the correlation between MFI(14) and RSI(14) was 0.753; on daily Bitcoin since August 2017, it was 0.791. MFI reaches its extremes far more often: on the ETF it was above 80 on 4.0% of the days and below 20 on 1.3%, while RSI was above 80 on 0.4% and below 20 on 0.1%. The two rarely agreed at the extremes: they were both below 20 on 5 days, MFI alone on 106 days and RSI alone on 4. MFI below 20 was about as rare as RSI below 30 (1.3% of days against 1.4%), and MFI above 80 half as frequent as RSI above 70 (4.0% against 8.4%): the choice between them is largely a choice of thresholds. On Bitcoin, keep in mind that the volume of our series is that of one exchange, not of the whole market.

Best MFI settings

14 bars is the setting recommended by Quong and Soudack, according to StockCharts. With 80 and 20 on the ETF since 1993, MFI(14) entered overbought territory 107 times, about 3.2 times a year, and oversold territory 50 times, about 1.5 times a year. The stricter 90 level was reached 20 times, and the 10 level twice, on November 4, 1993 and December 24, 2018.

Does it work?

Our bench tests chart patterns detected by rule against random entries, not indicators (see our methodology). We measured what followed MFI(14) extremes on the daily S&P 500 ETF from February 1993 to October 2026, without costs. Over the 10 days after MFI entered oversold territory, the ETF rose 64.0% of the time, with a mean change of +0.29%, against 61.9% and +0.46% for all days. After it entered overbought territory, the ETF rose 70.1% of the time, with a mean of +0.48%: overbought readings were followed by gains more often than usual. On daily Bitcoin since August 2017, the 10 days after entering oversold averaged −1.29% and those after entering overbought +4.65%, against +1.51% for all days. On neither market did overbought readings precede declines on average. These are descriptive averages, not a trading test.

How to repeat the count: compute MFI(14) on daily bars as in the formula above, with the typical price of each bar compared to the previous one. An entry into overbought territory is a day with the MFI above 80 after a reading at or below 80 the day before; an entry into oversold territory is a day below 20 after a reading at or above 20. The change is measured from that day’s close to the close 10 trading days later, and the reference is the same 10-day change from every day with an MFI value. The ETF series uses dividend-adjusted prices, from January 29, 1993 to October 28, 2026; the Bitcoin series starts on August 17, 2017.

The MFI is the volume-weighted form of the RSI. For other volume-based oscillators, see the Chaikin Money Flow and the Chaikin Oscillator, or the on-balance volume. All momentum tools are on our page of momentum indicators.