The Bullish Percent Index Manual
Research note. Computed at render from our daily Point & Figure scan of ~4.8k US common stocks, 2010+. Forward returns are SPY price returns. See How we checked it.
What is the Bullish Percent Index?
The share of stocks whose Point & Figure chart is on a buy signal. Abe Cohen of ChartCraft built it in 1955 on NYSE stocks, and its charm is what it ignores: P&F charts filter out every move smaller than the box size, so each stock is unambiguously “on a buy” or “on a sell” with no wiggle in between. That makes the BPI the slowest, most deliberate of the breadth gauges — where the McClellan oscillator measures day-to-day breadth velocity, the BPI counts standing commitments. Above 50% the bulls hold the field; the classic bands read above 70% as overbought and below 30% as oversold — a claim this Manual grades rather than repeats.
The latest reading
As of August 21, 2026: 55.9% — 2,682 of 4,799 stocks on buy signals, which is the bull field: majority on buys, no extreme. Since 2010 the index has spent only 310 days below 30 and 15% of days above 70. The live gauge updates the reading, the buy/sell counts and the zone after every close.
How the number is built
Each stock’s daily closes feed a Point & Figure chart with a 1% box and 3-box reversal. A stock flips to a buy signal when a rising column of X’s exceeds the prior X-column’s high, and to a sell when a falling column of O’s breaks the prior O-column’s low — double-top and double-bottom breaks, the plainest P&F rules. The BPI is simply buys ÷ (buys + sells) across our ~4800-stock common-stock universe (ETFs, preferreds and pseudo-tickers excluded). Cohen’s original and the widely followed StockCharts versions run the same logic on narrower index memberships; the broader universe changes levels at the margin, not behavior, and the difference is disclosed wherever it matters.
The sub-30 washouts — every episode since 2010
The folklore’s stronger half. Sub-30 readings mean fewer than three stocks in ten hold a P&F buy — a washed-out tape. Our record contains 25 episodes (21-session gap rule, graded from each episode’s first sub-30 day): SPY averaged +4.1% over the following quarter against a +3.2% baseline, and was higher six months on in 19 of 24 cases, averaging +7.6%. The exceptions are the lesson: 2022 produced three consecutive sub-30 washouts that kept falling — January, April and September — and 2021’s November washout preceded the bear market rather than a bottom. Washouts marked accumulation zones in range-bound and recovering tapes; in a genuine bear trend the index just kept washing out.
| First sub-30 day | Low | SPY +21s | +63s | +126s |
|---|---|---|---|---|
| 2010-02-05 | 22% | +7.3% | +4.3% | +5.4% |
| 2010-05-07 | 14% | -4.2% | +1.0% | +9.9% |
| 2011-06-13 | 26% | +3.2% | -8.6% | -1.3% |
| 2011-08-03 | 7% | -4.1% | -3.3% | +5.2% |
| 2011-11-22 | 19% | +5.1% | +14.9% | +11.2% |
| 2012-05-18 | 24% | +4.6% | +9.6% | +7.2% |
| 2012-11-15 | 28% | +5.9% | +12.9% | +23.0% |
| 2014-10-07 | 20% | +4.7% | +4.7% | +8.1% |
| 2015-08-21 | 15% | -1.9% | +5.5% | -2.7% |
| 2016-01-07 | 13% | -4.4% | +5.4% | +9.6% |
| 2018-02-08 | 27% | +8.1% | +5.6% | +10.7% |
| 2018-10-10 | 17% | +0.8% | -6.9% | +4.3% |
| 2018-12-11 | 9% | -1.9% | +6.4% | +9.6% |
| 2020-02-27 | 3% | -14.8% | +1.8% | +16.8% |
| 2020-09-24 | 28% | +6.9% | +13.6% | +22.4% |
| 2021-11-30 | 23% | +4.5% | -3.9% | -10.1% |
| 2022-01-24 | 24% | -4.1% | -2.6% | -11.1% |
| 2022-04-26 | 13% | -4.5% | -3.6% | -7.5% |
| 2022-09-19 | 8% | -4.5% | -1.4% | +2.7% |
| 2023-03-10 | 20% | +6.2% | +11.4% | +16.2% |
| 2023-09-21 | 19% | -2.4% | +8.5% | +20.8% |
| 2024-04-18 | 29% | +6.0% | +9.9% | +16.6% |
| 2024-12-19 | 29% | +4.0% | -1.8% | +3.6% |
| 2025-03-04 | 8% | -2.1% | +3.3% | +11.6% |
| 2026-03-20 | 25% | +8.6% | +14.8% | open |
The overbought myth
The folklore’s weaker half fails outright. “Above 70 is overbought and due for a correction” appears on nearly every reference page — and across our 32 over-70 episodes since 2010, SPY’s forward returns were above baseline: +4.1% over the next quarter (28 of 32 positive) and +7.5% over six months. Broad participation is what strong markets look like — the same finding our 90% up days and NYSE A/D studies keep producing. The practitioners’ defense is that the raw level was never the signal — Cohen’s successors read reversals from the extremes (bull alert, bear confirmed and the rest of the vocabulary). Fair, and it concedes the point most readers take from the bands: on sixteen years of daily data, selling because the BPI crossed 70 was a losing rule.
Where it will mislead you
Four traps. The overbought fade — covered above; the 70 line describes strength more often than fragility. Washouts in a bear trend — sub-30 is an average edge and no floor; 2022 hit it three times on the way down. Universe dependence — an NYSE-only, index-only and all-market BPI disagree at the margin, and none is the referee (the same lesson as the Zweig Breadth Thrust’s near-misses). And slowness cuts both ways — P&F states change only on 1%-box breaks, so the BPI ignores noise and also lags fast turns; it confirmed the 2020 bottom weeks after the velocity gauges did.
How we checked it
Everything comes from one pipeline artifact: the daily per-stock P&F state scan (1% box, 3-box reversal, double-top/double-bottom signals) aggregated to the BPI, with SPY closes attached. Band episodes use a 21-session gap rule and grade forward price returns from each episode’s first qualifying day, only where the full window exists. The baseline is every day’s forward 63-session return over the same years. Overlapping episodes share regimes and are not independent samples; the 2010+ start means the record contains no 2008-style collapse, and we say so rather than extrapolate.
Frequently asked questions
What is the Bullish Percent Index in simple terms?
The share of stocks whose Point & Figure chart is on a buy signal. Because P&F charts ignore small moves entirely, each stock is simply "on a buy" or "on a sell" — the BPI counts the bulls. Above 50% the bulls hold the field; the classic bands call above 70% overbought and below 30% oversold.
What is the Bullish Percent Index right now?
55.9% as of August 21, 2026 — 2,682 of 4,799 common stocks on Point & Figure buy signals, which is the bull field (above 50%). The live reading updates after every close on our Bullish Percent Index tool page.
Is a BPI below 30 a buy signal?
On average it has paid to lean that way, with real exceptions. Across 25 sub-30 episodes since 2010, SPY averaged +4.1% over the following quarter (baseline +3.2%) and was higher six months later in 19 of 24 cases. The exceptions cluster: 2022 produced three consecutive sub-30 washouts that kept falling.
Is a BPI above 70 a sell signal?
Our data says no. The 32 over-70 episodes since 2010 preceded ABOVE-baseline forward returns — +4.1% over the next quarter and +7.5% over six months. Broad participation is how strong markets behave; treating 70% as a fade signal has been a losing rule on sixteen years of data.
Who invented the Bullish Percent Index?
Abe Cohen of ChartCraft (later Investors Intelligence) in 1955, applied to NYSE stocks. Later practitioners — notably Earl Blumenthal and Mike Burke — added the reversal-based signal vocabulary (bull alert, bull confirmed, bear alert and so on) built from the index's own P&F chart.
How is our BPI different from the StockCharts version?
Same construction — 1% box, 3-box-reversal Point & Figure buy/sell state per stock — applied to a much broader universe: roughly 4800 US common stocks rather than a single index's members. Broader coverage makes the reading less hostage to one index's composition; the level bands behave the same way.