The short squeeze predictor nobody is using — Asquith-Pathak-Ritter 2005 applied
Why "high short interest" alone misses the setup, what the 2005 paper actually says, and how the three-factor confluence (short ratio + momentum + options flow) avoids the usual false positives.
"Short interest is 20% of float, this is going to squeeze" is one of the most-repeated and most-misleading takes in retail equity research. High short interest alone is a reason to avoid a stock, not to buy it — sophisticated money went short for a reason. What the 2005 Asquith-Pathak-Ritter paper actually showed, and what DeepVane's SHORT_SQUEEZE_SETUP pattern operationalises, is that short interest combined with two other specific conditions is the actual tradeable setup.
What Asquith-Pathak-Ritter 2005 actually found
The paper, published in the Journal of Financial Economics, ran a long-short portfolio sorted on short interest ratio (days-to-cover) from 1988-2002. Key finding: the top-decile shorted cohort underperformed the market by about 15% annualised. In other words, a naive "buy what's heavily shorted" strategy loses substantial money.
The squeeze trade appears only under specific conditions. Within the high-short-interest cohort:
- Stocks where price had already started rising over the past 30 days outperformed by ~15% over the following 60 days
- Stocks where momentum was flat or declining continued to decline
- The effect strengthens further when derivatives markets position bullishly
The intuition is mechanical: short sellers cover shares to close positions. When the stock is already trending up, each cover pushes the price higher, triggering more forced covers, accelerating the move. When the stock is trending down, shorts simply stay short because their thesis is working. Short interest by itself is fuel without a fire.
The three-factor confluence
DeepVane operationalises this as a three-condition pattern. All three must fire simultaneously:
- Short interest elevated. The crowded-short condition — measured as short interest ratio or days-to-cover high enough that a meaningful fraction of float is short. Our score_short_interest is inverted so a low score = high actual short interest, and we require the score in the bottom quartile of the universe.
- Momentum already rising. The ignition condition. Price must be up over the past 1-3 months. We require score_momentum in the top half of the universe — specifically above the 60th percentile.
- Options flow bullish. The confirmation condition. Put/call ratios and IV skew indicating call-buying pressure (what Pan-Poteshman 2006 identified as informed-money-positioning). score_options above the 60th percentile.
When all three fire, the pattern tags the ticker as SHORT_SQUEEZE_SETUP with a composite override above the raw factor sum. When any one fails — high short interest but flat momentum, or crowded short + rising price but bearish options — the pattern doesn't fire and the composite reflects the underlying factor mix without the squeeze premium.
Why most "squeeze calls" fail
Retail commentary usually looks at short interest in isolation. This produces a high false-positive rate because most heavily-shorted stocks are shorted with justification: deteriorating fundamentals, accounting irregularities, dilution risk. The short thesis is correct; the shorts don't cover until fundamentals show improvement, at which point the short-covering is a gradual unwind, not an explosive move.
The three-factor screen rejects the false positives mechanically. Without rising momentum there's no ignition; without bullish options flow there's no derivatives-side confirmation. The 2020-2021 memestock setups (GME, AMC, BBBY in their squeeze windows) all exhibited the full three-factor confluence before the squeeze move; many heavily-shorted 2024 names that never squeezed failed one or both of the confirming conditions.
Tail-alignment: how confident should we be?
Beyond the trigger, DeepVane adjusts pattern confidence using the tail-dependence between participating factors. The specific question: do short_interest, momentum, and options_flow actually co-move in the upper tail historically? If yes, the pattern fires with high confidence. If the three factors happen to align on this ticker but historically show no joint upper-tail dependence, we damp the confidence.
What to watch after the pattern fires
Setup doesn't mean certainty. Post-trigger watch items:
- Insider activity. If insiders are selling into the move, the squeeze is short-lived even if the three-factor setup is clean.
- Earnings calendar. Squeezes within 5-10 days of scheduled earnings compress into the report; position sizes should reflect the binary event.
- Sector context. Squeezes ride the macro tape. A single-name squeeze against a declining sector usually fails within 2 weeks.
- Short interest update. Bi-weekly SEC filings; watch for short interest declining rapidly — that's the covering happening, which is usually where the move exhausts.
Where to see this live
The SHORT_SQUEEZE_SETUP pattern page lists every ticker currently firing the three-factor confluence, refreshed daily. The pattern appears on individual stock pages as the SQUEEZE SETUP badge above the factor breakdown. Full mechanism is in Methodology, and our Track Record page describes how we'll measure per-pattern hit rate once forward returns accumulate from May 2026.
The pattern is bullish when it fires cleanly. It's also one of the patterns we're most careful about for exactly the reason discussed here: the same three-factor setup without academic grounding produces many false positives in retail commentary, and credibility depends on the willingness to withhold the label when any leg is weak.