Squeeze Score
0-100 rank of heavily shorted US stocks by odds of a large move
The Tapeboard Short Squeeze Score is a 0-100 rank of heavily shorted US-listed stocks (short interest of at least 15% of float) by the odds of a large price move over the next 5-10 sessions. It is not a directional call: high-scoring names have historically been more likely to jump 20% and more likely to fall 20%.
- Gate: short interest >= 15% of float (latest published FINRA settlement, split-checked). Listed common stock, price >= $1, liquid, market cap >= $25M.
- Score = 100 x (0.7 x percentile of 20-day volatility + 0.3 x percentile of 5-day float rotation), percentiles within the gated set. Borrow fee, days to cover and short-volume ratio are context only (weight 0).
- Out of sample (Oct 2025 - Sep 2026), ranks 1-10 hit +20% within 5 sessions 16.6% of the time and fell 20% 12.0% of the time, vs 4.5% / 2.3% for all stocks.
How is Squeeze Score calculated?
Eligible if SI_pct_float >= 15% (and <= 120%), settlement <= 40 days old, listed common stock, close >= $1, ADV20 >= 100k shares, median 20d dollar volume >= $500k, market cap >= $25M, >= 60 daily bars with one in each of the last 5 sessions. vol20 = population stdev of the last 20 daily log returns rot5 = (sum of the last 5 sessions' volume) / float Squeeze Score = 100 * (0.7 * pct(vol20) + 0.3 * pct(rot5)) pct(.) is the average-tie percentile rank within the day's eligible set.
Short interest is used as a gate, not a weight: it defines the population where short covering can amplify a move. Within that population the ranking uses realized volatility and float turnover, the two inputs that best separated large future moves out of sample. Percentiles replace z-scores so a single outlier cannot set the scale, and there is no decay multiplier. Weights were chosen on 2024-03 to 2025-09 data only and tested on later data.
What do traders use Squeeze Score for?
- Finding heavily shorted names where a violent move (either way) is more likely than usual. The /short-squeeze-stocks leaderboard shows the ranked set daily with historical odds per rank bucket.
- Sizing risk on a crowded short: the same names that squeeze also crash more often, so the odds cut both ways.
- Separating setups from same-day movers: stocks already running today appear in a separate In play list with their own (mean-reverting) historical odds.
Worked example
Suppose $GME carries short interest of 21.4% of float (above the 15% gate), its 20-day volatility sits at the 90th percentile of the eligible set and its 5-day float rotation at the 60th. Its score is 100 x (0.7 x 0.90 + 0.3 x 0.60) = 81. Where that lands in the day's ranking maps to a historical odds bucket (for example ranks 11-25: about 13% chance of a +20% high and 7% chance of a -20% low within 5 sessions).
Live data: /stocks/GME.
What are common pitfalls of Squeeze Score?
- Short interest publishes twice a month, about 8-9 trading days after settlement, so the gate uses a reading that is typically 2-4 weeks old.
- The score is not a buy signal. The median 5-day return of the top bucket has been slightly negative; the edge is in the size of moves, not their direction.
- Reverse splits can corrupt SI %-of-float; settlements flagged for a split, or where the float changed more than 2x since settlement, are excluded rather than guessed.
- Historical odds are frequencies from one out-of-sample year with wide confidence intervals, not guarantees.
Where this metric appears on Tapeboard
Squeeze score appears on every `/stocks/{T}` page (numeric + 0-100 bar), the /scanner/squeeze leaderboard (top 25 daily), the /short-squeeze-stocks landing, and as a screener filter on /scanner.
Tapeboard surfaces this metric as a first-class screening filter. See the comparison pages at all Bloomberg alternatives for how Tapeboard's squeeze score surface compares against Koyfin, Finviz, TradingView, and Unusual Whales.
Related terms
- Short Interest: FINRA short-interest reports settle bi-monthly with a publication lag of about 10 business days from the report date.
- Utilization (Lending): Daily-published, sourced from broker-dealer stock-loan books (Tapeboard uses IBKR).
- Days To Cover: Higher DTC means a slower unwind. Above 5 days is a meaningful liquidity constraint on short covering; above 10 is severe.
- Failures-To-Deliver (FTD): Bi-monthly SEC dataset, lag of about two weeks. Counts unsettled share volume per ticker per day.
Primary sources cited
- FINRA Short Interest Reporting: Short-interest %-of-float is sourced from the FINRA bi-monthly short-interest report. https://www.finra.org/finra-data/short-interest, retrieved 2026-05-04.
- FINRA short-sale-volume documentation: Float utilization and borrow-fee inputs route from FINRA short-sale data and IBKR stock-loan inventory. https://www.finra.org/finra-data/short-sale-volume-data, retrieved 2026-05-04.
Methodology last reviewed 2026-05-04 by Marcus Reilly, Editor at Tapeboard. Every claim on this page has a row in the citation registry. Glossary terms reverify on the Jan 15 / Apr 15 / Jul 15 / Oct 15 cron and any time the underlying primary-source publishes a methodology change. See methodology for the full fact-check process and corrections for the public correction log.
Disclaimer. This page is for educational and informational purposes only. Nothing on Tapeboard is investment advice. See the full risk disclaimer.