What is a distribution day, and how many is too many?

By Daniel, The Philosopher Investor · updated October 2, 2026

Why does a down day on higher volume matter?

Prices fall on light volume when buyers step back; they fall on heavy volume when holders sell. The second is what a distribution day captures. Institutions cannot exit a position in a day, so they sell into strength over weeks, and the tell is the index closing lower on more shares than the day before. A single day can be noise. A pattern of them is a fund manager reducing exposure.

The 0.2 percent threshold exists to exclude flat days where volume happened to be higher. The volume comparison is to the prior session, never to the average, so a distribution day can occur on a below-average volume day if the day before was quieter still.

How do you count them?

Keep a rolling window of about 25 trading sessions, roughly five weeks, and count the distribution days inside it for each major index separately. Investor's Business Daily runs the count on the S&P 500 and the Nasdaq. When either reaches five or six the market outlook shifts to under pressure. Any higher and it is treated as a correction.

A day drops out of the count in two ways. It ages out after 25 sessions. Or the index rises five percent or more above that day's close, which means the selling was absorbed and no longer matters. This is why a strong rally clears the count quickly.

What is stalling?

A variant that catches selling into strength without a down close. The index rises slightly, less than about 0.1 percent, on higher volume than the prior day, after a run of gains, and closes in the lower half of its range. Heavy volume that produces no progress means supply is meeting demand at the highs. IBD counts stalling days alongside distribution days, and they often come first.

Stalling days are the subtle version. A chart that shows three up days each closing lower in its range on rising volume is a chart where somebody is selling into every rally. That is distribution with better manners.

How many is too many?

Five or six inside the window has been the threshold since O'Neil's original work, and it has held up. Uptrends can absorb two or three. Once the count reaches five, the historical record shows a large share of uptrends topping within weeks. Not all, and the count has produced false alarms in long bull markets where volume patterns were distorted.

Read the count with the leaders. A high count while leading stocks hold their moving averages is a warning. A high count while the leaders break down, on their own heavy volume, is the top being made in front of you.

What does the count not tell you?

How far the correction will go, or how long. A cluster of distribution days precedes both five percent pullbacks and forty percent bear markets. The count is a sign that the uptrend is under pressure and that new buys are risky. It is a defensive tool: raise cash, tighten stops, stop adding. It is not a forecast of the depth of what follows.

It also lags on gap-driven markets. A single news shock that takes the index down four percent in a day registers as one distribution day, the same as a quiet 0.3 percent decline. The count is best on the slow-turning tops it was built for, and weakest on the sudden ones.

Common questions