Positive Volume Index (PVI): Following the Crowd on High-Volume Days

Paul Dysart didn't wake up one morning and decide to build a single indicator. He built a system -- two cumulative indices that split the market into its noisy days and its quiet days. The Positive Volume Index tracked what happened when volume rose. Its twin, the Negative Volume Index (NVI), tracked what happened when volume fell. Together, they were designed to answer a question that still matters today: are the informed players and the crowd moving in the same direction, or are they quietly diverging?

The theory behind PVI rests on a behavioral finance observation that has held up remarkably well across nearly a century of market data. When volume surges, the increase is disproportionately driven by retail and less-informed participants. News breaks, CNBC runs a segment, social media lights up, and suddenly everyone wants in -- or out. These high-volume sessions capture the emotional, reactive side of the market. The crowd buys into rallies because prices are going up, and sells into panics because prices are going down. It's reflexive, not analytical.

Dysart's original work on the New York Stock Exchange in the 1930s was largely overlooked by the trading community. Volume analysis was still in its infancy, and most market practitioners focused on price action alone. It took four decades for Norman Fosback to resurrect the concept, backtest it rigorously, and publish the results in Stock Market Logic. Fosback's contribution was quantitative: he didn't just theorize about crowd behavior -- he proved that PVI's position relative to its moving average carried statistically significant predictive value for bull and bear market regimes.

What makes PVI conceptually interesting is its selective attention. Most indicators process every single bar of data regardless of volume context. PVI deliberately ignores low-volume sessions. It's not that quiet days don't matter -- they matter a great deal, which is exactly why the NVI exists. But PVI's job is narrower: it captures what happens to price specifically when participation is elevated. If price rises on a high-volume day, PVI rises. If price falls on a high-volume day, PVI falls. If volume is flat or declining, PVI does absolutely nothing. This conditional design means PVI only reflects the market's behavior when the crowd is most active.

The practical implication is that PVI acts as a sentiment proxy for the majority of market participants. When PVI is trending upward, the crowd is making money -- buying into an uptrend and being rewarded for it. When PVI is trending downward, the crowd is losing -- either buying at the wrong time or panic-selling into declines. Neither state is permanent, but extended periods of PVI weakness tend to coincide with distribution phases where institutional sellers are quietly offloading positions to eager retail buyers.

One important nuance: Dysart and Fosback never claimed that the crowd is always wrong. During sustained bull markets, the crowd is right for long stretches -- prices rise, volume expands, and PVI climbs steadily. The crowd only becomes consistently wrong at turning points: at major tops where euphoria peaks and at major bottoms where panic reaches its maximum. PVI's value isn't in labeling the crowd as stupid. It's in showing you when crowd behavior is confirming a trend versus when it's diverging from what quieter, more informed activity suggests.

This is also why PVI is rarely used alone. Its real power emerges when paired with the NVI, which we'll explore in detail shortly. Think of PVI as one lens in a pair of binoculars -- useful by itself, but far more powerful when both eyes are open.

The math behind PVI is refreshingly simple. No exponential smoothing, no complex normalization -- just a conditional update rule that any spreadsheet can handle in about thirty seconds.

Here's the logic in plain English: look at today's volume. Is it higher than yesterday's volume? If yes, update PVI by the percentage change in price. If no, do nothing. That's it. The entire indicator boils down to one if-then statement.

Formally, the calculation works like this:

If Volume(today) > Volume(yesterday): PVI = PVI(previous) + [(Close(today) - Close(yesterday)) / Close(yesterday)] x PVI(previous)

Or equivalently: PVI = PVI(previous) x (1 + Price Percentage Change)

If Volume(today) <= Volume(yesterday): PVI = PVI(previous) -- no change.

The starting value is conventionally set to 1,000. This is arbitrary -- you could start at 100 or 10,000 and the indicator's behavior would be identical, just scaled differently. The 1,000 baseline simply provides a readable reference point.

Let's walk through a concrete example to make this tangible:

Notice Day 3: price dropped, but volume also dropped, so PVI stayed frozen. The indicator doesn't care what price did -- only what price did when volume was elevated. Day 5 is the interesting one: volume surged but price fell, so PVI declined. The crowd showed up in force and lost money. That's exactly the kind of information PVI is designed to capture.

A subtle but important point: PVI doesn't use volume in its actual calculation formula. Volume only serves as a gate -- a binary yes/no switch that determines whether the price change gets recorded. Once the gate is open, the magnitude of the volume increase doesn't matter. A 1% volume increase and a 200% volume spike both trigger the same update. This is by design. Dysart was interested in the direction of crowd activity, not its intensity.

The signal line adds the analytical layer. A 255-period exponential moving average (EMA) of PVI values is the standard, representing roughly one trading year on a daily chart. Some platforms use a simple moving average instead -- the difference is marginal in practice since PVI moves slowly enough that the choice of average type barely affects the output.

When PVI crosses above its 255-period EMA, the interpretation is that crowd buying on high-volume days is accelerating -- a bullish signal. When PVI crosses below, crowd activity is producing negative returns on high-volume sessions -- a bearish signal.

Fosback's backtested results on U.S. equity data from 1941 to 1975 produced specific probability estimates: when PVI sat above its one-year moving average, the probability of being in a bull market was approximately 79%. When PVI fell below, bear market probability rose to roughly 67%. These aren't trading signals in the scalping sense -- they're regime filters that tell you whether the broad market environment favors being long or cautious.

On most charting platforms, PVI is available as a built-in indicator. TradingView lists it under the volume category with the standard 255-period EMA. MetaTrader 5 users may need to download a custom version from the MQL5 marketplace, as PVI isn't included in the default indicator set. When configuring it, the only parameter you'll typically adjust is the signal line period -- and for most traders, the default 255 is the right choice.

Using PVI without NVI is like reading only the even-numbered pages of a book. You'll get some of the story, but you're missing half the plot. The two indicators were designed as a pair, and their combined interpretation is considerably more powerful than either one alone.

Here's the core framework. NVI tracks price changes on days when volume decreases -- the quiet sessions where institutional and informed traders are believed to operate. PVI tracks price changes on days when volume increases -- the loud sessions dominated by crowd behavior. When both indicators agree on direction, the market has broad consensus. When they diverge, something interesting is happening beneath the surface.

Four scenarios define the PVI/NVI relationship:

The two divergence scenarios -- rows three and four -- are where the real analytical gold lies. When PVI falls while NVI rises, the crowd is getting shaken out on high-volume days (panic selling, failed breakouts, stop hunts), but on quiet days, price is creeping higher. This pattern suggests informed players are accumulating positions while retail traders are bailing out. It's a classic bottoming signature.

The opposite divergence -- PVI rising while NVI falling -- is the topping equivalent. The crowd is buying enthusiastically on high-volume days, pushing PVI higher, but on quiet sessions, prices are slipping. Institutional money is using the crowd's enthusiasm as liquidity to distribute their holdings. By the time the crowd realizes the trend has turned, the smart money is already out. This is the pattern that often precedes the nastiest corrections, because the exit door gets very narrow once retail buying dries up.

A practical approach for combining both indicators: plot PVI and NVI on the same daily chart, each with its 255-period EMA signal line. The strongest buy signals occur when NVI crosses above its signal line while PVI is either flat or below its own signal line. This configuration means smart money is bullish while the crowd is either neutral or pessimistic -- a setup that historically precedes the early stages of new bull markets when skepticism is highest and institutional buying is quiet but persistent.

The strongest sell signals invert this: PVI crosses above its signal line (crowd euphoria) while NVI drops below its own signal line (smart money withdrawing). The crowd is all-in, the smart money is checking out, and the party is about to end.

Fosback's research quantified this pairing. Using NVI alone, the probability of correctly identifying a bull market when NVI was above its one-year average was approximately 96% -- a remarkably high hit rate. PVI's standalone numbers were lower (79% bull market probability when above its average). But the combination of both indicators in agreement produced the most reliable regime identification in his backtesting data.

One caveat worth emphasizing: these probability estimates come from a specific test period (1941-1975) on U.S. equity data. Modern markets have different characteristics -- algorithmic trading, passive index fund flows, and the democratization of information through social media all affect the volume dynamics that PVI and NVI try to capture. The indicators still provide useful directional guidance, but the exact probability figures should be treated as historical context rather than fixed constants.

For forex traders using tick volume, the PVI/NVI pairing loses some of its theoretical grounding since tick volume is a proxy rather than true transacted volume. The indicators still produce readable signals on major pairs like EUR/USD and GBP/USD during liquid sessions, but the smart money versus crowd money distinction is less clean than on exchange-traded equities where volume data is centralized and standardized.

The bottom line: always plot both. PVI tells you what the crowd thinks. NVI tells you what the quiet operators think. When they disagree, pay attention to the quiet operators -- historically, they've been right more often.

Here's the thing about crowd behavior that contrarian-minded traders often miss: the crowd is right most of the time. During trending markets, following the herd is profitable. The crowd only gets destroyed at extremes -- at major tops and major bottoms. In between those turning points, which can be separated by months or years, the crowd's momentum-chasing tendency actually works.

PVI captures this reality directly. During a healthy bull market, high-volume days tend to produce price gains because new buyers are entering the market, institutions are filling large orders on the buy side, and overall demand exceeds supply. PVI rises steadily, stays above its 255-period signal line, and the message is clear: the crowd is making money on the days it's most active, which means the trend is intact.

This makes PVI a surprisingly effective bull market confirmation tool -- not for timing entries, but for answering a more fundamental question: should I be positioned long right now, or should I be defensively cautious?

The confirmation framework is straightforward:

Bull market confirmed: PVI above its 255-period EMA. According to Fosback's data, this state corresponded with bull markets approximately 79% of the time. The crowd is participating on high-volume days and being rewarded, which means buying pressure is genuine and broad-based.

Bull market in question: PVI below its 255-period EMA. Bear market probability jumps to roughly 67%. High-volume days are producing negative returns on average, suggesting the crowd is buying into weakness or selling into strength -- neither of which supports sustained upward price movement.

The practical application is less about individual trades and more about portfolio posture. When PVI confirms a bull market (above signal line), traders can afford to be more aggressive: larger position sizes, wider profit targets, and willingness to hold through normal pullbacks. When PVI signals a bearish regime (below signal line), reducing exposure, tightening stops, and favoring capital preservation over growth becomes the priority.

A concrete example from equity markets illustrates this well. During the 2020-2021 bull run following the pandemic crash, PVI climbed steadily above its signal line by mid-2020 and stayed there for over a year. High-volume days consistently produced gains as retail participation surged (think stimulus checks, zero-commission trading, and meme stock enthusiasm). The crowd was right -- aggressively, persistently right -- for 18 months straight. Traders who dismissed the rally as "retail-driven" and therefore unsustainable missed one of the strongest bull markets in recent history.

PVI started flattening in late 2021 as high-volume days began producing smaller gains or outright losses. By early 2022, PVI had crossed below its signal line on several major indices, coinciding with the onset of the bear market that dragged equities lower through most of that year. The crowd was still showing up on high-volume days, but they were losing money -- PVI's decline was the tell.

The signal timing isn't precise -- PVI is a lagging indicator by nature, and the 255-period EMA further smooths the signal. Expect the crossover to occur weeks or even a few months after a genuine trend change has begun. This lag is acceptable when using PVI as a regime filter rather than an entry trigger. You don't need to catch the exact top or bottom. You need to know whether the prevailing environment favors offense or defense.

Combining PVI's bull market confirmation with a trend-following entry method creates a two-layer system. For instance: only take long entries from a 50/200-day moving average golden cross when PVI confirms a bullish regime. This filter eliminates many false crossovers that occur during distribution phases when price temporarily rises but crowd participation (as measured by PVI) is deteriorating.

One last thought on bull market confirmation: PVI works best as a long-term filter on daily charts applied to broad market indices (S&P 500, NASDAQ, DAX, FTSE). Individual stocks can have idiosyncratic volume patterns that muddy the crowd-behavior interpretation. A single stock might see a volume surge because of an earnings report or a merger announcement, not because of broad crowd participation. Index-level PVI smooths out these firm-specific events and captures the aggregate crowd sentiment that the indicator was designed to measure.

Let's be honest about what PVI can't do, because overselling any indicator is the fastest way to lose money with it.

The most fundamental limitation is the assumption itself: that high-volume days are dominated by uninformed, crowd-following traders. In modern markets, this isn't always true. High-frequency trading firms, quantitative funds, and algorithmic market makers generate enormous volume without any retail involvement. An earnings announcement might trigger massive institutional rebalancing that dwarfs retail order flow. Volume surges driven by index fund rebalancing at quarter-end have nothing to do with crowd sentiment. PVI's theoretical foundation -- high volume equals crowd, low volume equals smart money -- was more accurate in the 1930s and 1970s than it is in the era of algorithmic trading.

The 255-period signal line creates significant lag. By the time PVI crosses below its one-year moving average, a bear market may already be several months old. The 2022 drawdown in equities was well underway before PVI's bearish crossover confirmed on some charting platforms. If you're looking for timely exit signals, PVI alone won't save you. It's a confirmation tool, not an early warning system. Think of it as the indicator that tells you the house is on fire -- after the smoke alarm (price action, momentum indicators) has already gone off.

Forex traders face an additional structural problem. The forex market is decentralized -- there's no single exchange reporting total volume. What your broker shows you is tick volume (the number of price changes per period), which is a proxy at best. PVI's entire logic depends on comparing today's volume to yesterday's volume. If that volume data is incomplete or inconsistent between brokers, the gate mechanism -- the core of PVI's calculation -- triggers at different points for different traders looking at the same currency pair. This doesn't make PVI useless on forex, but it does mean you should trust it more on EUR/USD during London hours than on NZD/CHF at 3 AM.

The binary nature of PVI's volume gate is another design limitation. A day where volume rises by 0.5% and a day where volume doubles both trigger the same PVI update. The indicator can't distinguish between a modest volume uptick (possibly meaningless noise) and a genuine participation surge. Indicators like On Balance Volume (OBV) or the Volume Rate of Change offer more granular volume analysis when you need to assess the intensity of participation rather than just its presence.

PVI also provides no information about where within a session the volume occurred. A day might have elevated total volume, but if 80% of that volume occurred in the first 15 minutes due to an overnight gap and the rest of the session was quiet, PVI treats it the same as a day where volume was steadily elevated throughout. Intrabar volume distribution -- something tools like Volume Profile and Order Flow capture -- is invisible to PVI.

Backtesting PVI strategies requires caution. Fosback's original probability estimates (79% bull market accuracy when PVI is above its average) come from a specific 34-year test window (1941-1975) on U.S. equity indices. Market structure has changed dramatically since then. Passive investing didn't exist. Retail traders couldn't access markets from their phones. ETFs, derivatives, and dark pools were decades away. While PVI's core logic remains sound in principle, the exact numerical probabilities should not be taken as gospel for modern markets.

Finally, PVI generates very few signals. On a daily chart, weeks or even months can pass without a meaningful crossover. For active traders who want regular trade setups, PVI is frustratingly quiet. It's not an action-generating indicator -- it's a background filter that occasionally says something useful about the market's broader regime. If you need frequent signals, look elsewhere. If you want a patient, long-term lens on crowd behavior that complements your existing analysis, PVI earns its place on the chart.

The sensible approach: use PVI as one input in a multi-indicator framework, never as a standalone trading system. Pair it with NVI for the complete Dysart system. Layer in momentum indicators (RSI, MACD) for timing. Use price action for entries and exits. PVI tells you the weather forecast -- whether it's bull market sunshine or bear market storms. Your other tools decide when to step outside.