A breadth thrust is a sudden, broad surge in market participation off a low — a moment when the proportion of advancing stocks flips from washed-out to overwhelmingly positive in a compressed window. The premise is that durable bull moves begin with a stampede of buying across the whole list, not just the index leaders, so a thrust off a bottom is read as a market-initiation signal rather than a fadeable bounce. The Zweig Breadth Thrust (ZBT), devised by Martin Zweig (Winning on Wall Street), is the canonical formulation. Its defining tension: the precise trigger is so demanding that it is genuinely rare — which is exactly what gives the "perfect track record" folklore around it almost no statistical weight.

How it's calculated / formed

The Zweig Breadth Thrust indicator is the 10-day exponential moving average (EMA) of [Advancing issues ÷ (Advancing + Declining issues)], traditionally computed on NYSE breadth data. The raw input is a daily ratio bounded between 0 and 1 (or 0–100%); the 10-day EMA smooths it.

The classic signal fires when that 10-day EMA:

• starts below 0.40 (≈40% — an oversold, washed-out reading where declines dominate), and
• rises above 0.615 (≈61.5% — broad upside participation), and
• does so within 10 trading days.

The two thresholds and 10-day window are Zweig's original specification, as carried in Greg Morris' Encyclopedia of Breadth Indicators and replicated by StockCharts/Arthur Hill (sources below). The 0.615 number is not arbitrary; it is the specific level Zweig identified as marking a true thrust rather than ordinary recovery.

A note on the math worth flagging: the widely used StockCharts/Morris version applies an EMA, but some implementations and earlier descriptions use a simple average of advances/(advances+declines). EMA vs. simple smoothing changes the exact day a signal trips and is one reason historical signal counts differ slightly between sources. Treat any precise count as implementation-dependent.

How it's used in practice

The ZBT is used almost exclusively as a bull-market initiation / "all-clear" confirmation, not as a timing oscillator:

• Off a bottom: a fresh signal is taken as evidence that a correction or bear phase has likely ended and that the rally has broad backing — practitioners use it to shift from defensive to risk-on positioning.
• As breadth confirmation: because it requires the whole list to participate, it screens out narrow, mega-cap-only rallies that index price alone can hide. A thrust that fires with the index breaking out is read as higher-quality than price strength alone.
• It has no native sell signal. The classic ZBT only triggers long; exits are imposed externally. Modern recodes (e.g. TrendInvestorPro's S&P 1500 version) bolt on an ATR-based exit because Zweig's original gives none.

The swing-level operational mechanics — exact entry, stop placement, position sizing, hold period — are not part of this node. Defer those to the Swing Trading branch; the ZBT here is a regime/initiation read, not a trade-entry recipe.

Related thrusts

• Whaley Breadth Thrust (Wayne Whaley, 2010 Charles Dow Award, Planes, Trains & Automobiles): a family of thrust tests over a 5-day window — advance thrusts (5-day advances/total issues crossing above ~75% or below ~25%), volume (up-volume/total volume) thrusts, and price thrusts. Broader and more frequent than Zweig's, with explicit bearish as well as bullish variants.
• "90% up days" (Lowry Research / Paul Desmond): days where up-volume and up-issues each exceed ~90% of the total; clusters of 90% up days after a 90% down day are read as a thrust confirming a bottom.
• Generic advance-decline / McClellan thrusts: surges in the McClellan Oscillator/Summation Index or in the A-D line are also used as thrust proxies. The Zweig signal is essentially the most stringent member of this larger breadth-thrust family.

Standing & evidence

This is where the topic demands real honesty, because the popular framing is statistically weak.

• The "100% win rate" claim is a tiny-sample result. Carson Group, citing the count back to WWII, states that the ~19 prior signals with a year of subsequent data all showed higher prices 12 months later, with a median ~13% gain at 6 months and an average gain above ~23% at 12 months. ISABELNET/Ned-Davis-style charts circulate similar figures (a frequently cited ~24.8% median 12-month gain). These are real, but they rest on only roughly a dozen-to-twenty signals since 1945 — far too few to constitute a robust base rate. "Every signal was followed by gains" with n≈12–19 is folklore-grade evidence, not a dependable edge; one or two failures would collapse the headline. The forward-return numbers should always be read as attributed claims on a tiny sample, not as a measured probability.
• The signal is genuinely rare — only on the order of a dozen-plus classic NYSE triggers since 1945. The NASDAQ version reportedly has not fired since the 1980s. Rarity is the whole problem: it is precisely what makes the track record unfalsifiable in any statistical sense.
• Modern triggers are disputed. The 2023 signal (and others) drew arguments that it was computed on a looser or different basis than Zweig's original — e.g. EMA vs. simple average, S&P 1500 vs. NYSE-only breadth, or thresholds reached without the full thrust dynamic Zweig intended. The 2023 signal also underperformed the folklore: the S&P 500 traded sideways for roughly two months after the spring trigger before topping mid-year — a "win" on a 12-month horizon but not the immediate liftoff the legend implies.
• The McClellan critique. Tom McClellan, extending the history back to 1928, has argued the buy signals have closer to a ~50/50 success rate when "success" is defined as the market promptly making new highs — directly contradicting the perfect-record narrative.
• Methodology is dated. NYSE-only breadth omits the now-enormous NASDAQ universe and is contaminated by non-operating-company issues (closed-end funds, preferreds, ETFs) on the NYSE list, which critics argue distorts the modern advance-decline numbers Zweig's thresholds were calibrated on.

Strengths & limitations

• Strength: when it genuinely fires off a deep low, it captures real, broad participation and has historically clustered near durable bottoms — a useful confirming input alongside other evidence.
• Limitation — sample size: the marquee "always up afterward" stat is the single most-overcited and least-robust claim about it. Do not treat it as a probability.
• Limitation — no exit, no bear signal: the classic version only tells you to get long; it is silent on when to leave.
• #1 misuse: quoting the 100%/24.8% forward-return figures as a reliable edge while ignoring the n≈dozen sample — and treating a modern, looser-recipe trigger as if it carried the original's (already weak) pedigree.

System relevance

Sibling breadth nodes in this branch — the McClellan Oscillator/Summation Index, advance-decline line, up/down volume, new highs–new lows, and 90%-day analysis — are the natural cross-links; the ZBT is the most stringent member of that breadth-thrust family and should be read alongside them, not in isolation. For Delvantic, a ZBT is best consumed as one regime/initiation input to the regime engine and Augustus — flagged explicitly as a rare, small-sample signal whose forward-return statistics are folklore-grade, not a measured base rate. Whether any given thrust "works" is Augustus's call at decision time with live data and Cairn's record, not a verdict this node should pre-bake. Swing entry/stop/target mechanics belong to the Swing Trading branch.

Sources

• StockCharts / Arthur Hill, "The Zweig Breadth Thrust Triggers, but the Indicator is Missing Something" (2023) — confirms 10-day EMA of NYSE Advances/(Advances+Declines), 0.40→0.615 within 10 days, citing Greg Morris' Encyclopedia of Breadth Indicators; notes NYSE-only / missing-NASDAQ limitation.
• StockCharts / Arthur Hill, "Two Ways to Use the Zweig Breadth Thrust" (2025) — ATR-based exit, S&P 1500 recode (ZBT1500).
• Tom McClellan, McClellan Financial, "Watching for a Zweig Breadth Thrust Signal" — history to 1928, ~50/50 success critique, simple-average vs. EMA note.
• Carson Group, "Houston, We Have A Zweig Breadth Thrust…" — ~19 signals since WWII, 100% higher 12 months later, ~13% median 6-month / ~23%+ average 12-month forward returns (attributed, small-sample).
• ISABELNET, "S&P 500 Index Returns – Zweig Breadth Thrust Signals Since WWII" — circulated ~24.8% median 12-month gain figure (Ned-Davis-style).
• Wayne Whaley, Planes, Trains & Automobiles: A Study of Various Market Thrust Measures (2010 Charles Dow Award) — Whaley Breadth Thrust definitions (5-day advance/volume/price thrusts).
• Investopedia / TrendSpider / Babypips — corroborating definitions of formula and thresholds.

Flags: forward-return figures (100% hit rate; ~13%/~23–24.8%) are attributed, tiny-sample (n≈12–19) folklore, not a robust base rate; EMA-vs-simple-average and NYSE-vs-S&P 1500 implementation differences mean signal counts/triggers vary by source; modern (e.g. 2023) triggers are disputed as looser than Zweig's original.