Negative Volume Index (NVI): Fosback’s Formula, the 255-Day EMA Rule, and What It Did on the S&P 500 ETF

The Negative Volume Index (NVI) is a running index that changes only on days when volume is lower than the day before. StockCharts traces it to Paul Dysart in the 1930s and to Norman Fosback, author of Stock Market Logic, who adapted it; TradingView dates Fosback’s version to the 1970s. The assumption behind it is that informed, “smart” money is active on quiet days and the crowd on busy days, so the price changes of low-volume days should show where the informed money is going. Its companion, the Positive Volume Index (PVI), moves only on days when volume rises.

How the Negative Volume Index is calculated

Dysart’s original version added net advances, advancing issues minus declining issues, on days when volume fell. Fosback substituted the percentage price change, and that is the version charting platforms use. StockCharts gives the steps: start at 1,000, add the percentage price change when volume decreases, leave the index unchanged when volume increases, and apply a 255-day EMA as a signal line, about one year of trading days.

How to read NVI: Fosback’s rule

Fosback compared NVI with its one-year average. According to StockCharts, he found the odds favor a bull market when NVI is above its 255-day EMA and a bear market when it is below, with asymmetric odds: a 96% chance of a bull market above the EMA, but only a 53% chance of a bear market below it. In other words, the rule is mostly a bullish filter. StockCharts adds that the indicator was designed for broad market indices and exchange volume, that it does not always act as expected on single stocks and ETFs, and that even a 255-day EMA gives plenty of whipsaws.

NVI on a real chart

The chart shows NVI and its 255-day EMA on the daily S&P 500 ETF from October 30, 2025 to October 28, 2026, with dividend-adjusted prices. For this chart we start the index at 1,000 on January 4, 2021, the first bar of the chart's data series, which leaves enough history for the 255-day average. Over the window, NVI rose from 5,057.6 to 7,406.0, a gain of 46.4%, while the ETF rose 14.8%, from 672.57 to 772.31. NVI stayed above its EMA on all 250 days, including March 30, 2026, when the ETF made its lowest close of the window, 628.78. The Positive Volume Index did the opposite: it fell from 389.3 to 305.3 and stayed below its own EMA on all 250 days, as the second chart shows on the same bars.

The gap is a matter of arithmetic. Of the 249 day-to-day changes in the window, the 136 made on lower volume compounded to +46.4%, and the 113 made on higher volume to −21.6%. The ETF’s losing days tended to come with rising volume, so the index that skips those days kept climbing.

Does it work?

Our bench tests chart patterns detected by rule against random entries, not indicators (see our methodology). We measured Fosback’s rule on the daily S&P 500 ETF from February 1993 to October 2026, without costs, with the EMA available from January 31, 1994.

All days had a mean 250-day change of +12.30% and rose 81.4% of the time; days with NVI above its EMA, +11.67% and 80.7%; the 313 days at or below it, +27.69%, and every one of them was followed by a rise.

NVI was above its EMA on 96.2% of the 8,241 days, a share of time that is not Fosback’s 96% probability, though it happens to be close. The 313 days below came in 21 separate dips; 213 of those days fell in 2002 and 2003, the rest in 1994, 1998, 2008 and 2009, so they are few independent cases. Over the full period, the ETF’s 4,307 lower-volume days compounded to 6,317 times their starting value and its 4,187 higher-volume days to 0.005 times. On Bitcoin, from 2018 to 2026, with volume from a single exchange, NVI was above its EMA on 66.5% of the days; the 809 days below it with a full year of data after them were followed by a mean 250-day change of +146.1%, against +11.0% for the days above. These are descriptive averages over overlapping periods, not a trading test.

How to repeat it: start NVI at 1,000 on the first bar of the daily series and multiply it by the day’s close divided by the previous close on every day whose volume is below the previous day’s; build the PVI the same way on days with higher volume. The signal line is a 255-day EMA, seeded with the simple average of the first 255 values. Each day with an EMA value and 250 later trading days is sorted by whether NVI was above its EMA, and the change runs from that day’s close to the close 250 trading days later. The windows overlap, so neighboring days share most of their future.

The On Balance Volume line adds volume instead of price changes, and the moving average page explains the EMA used as the signal line. See how trading volume behaves on its own, or browse all volume indicators.