Torrid's tariff refund lifted the headline, not the turnaround

Run timestamp: 2026-09-05 11:30 SGT Scope: U.S. market, short only Classification: Watchlist research, Reject for current execution Primary instrument: CURV common stock, NYSE, unlevered only if every execution gate later passes

Summary: Torrid's second-quarter earnings print looked better at first glance. Net income rose to $5.2 million from $1.6 million, reported adjusted EBITDA reached $23.3 million, and the stock finished the next session higher. The primary filing shows the less convenient bridge: $11.1 million of adjusted EBITDA came from an IEEPA tariff refund, gross margin excluding the refund was 33.9%, sales fell 11.8%, comparable sales fell 6.3%, and the company carried about $304.0 million of gross debt against $22.0 million of cash. The release itself says the full-year outlook was raised to include the refund.

The short thesis is not that Torrid has no recovery path. Management reports a July inflection, a completed store-footprint reset, improving inventory, new marketplaces, and a plan for about $40 million of fiscal-year expense savings. The narrower claim is that the market may be paying for recurring improvement after a one-time cash and margin event. A late finance-feed snapshot showed $2.44 at 2026-09-04 23:25:13Z, up 8.07% from the prior close, while a separate regular-session tape showed $2.50 at 15:42 EDT. The disagreement, incomplete top-of-book data, scarce displayed borrow, and squeeze risk make the signal Watchlist only.

Opportunity Ranking

The screen used mechanism-specific searches rather than a generic earnings query:

tariff refund adjusted EBITDA debt negative comparable sales post-earnings reversal short

small-cap retailer one-time customs refund recurring margin cash interest short squeeze

recent earnings beat below-surface demand debt refinancing borrow availability U.S. short

Current-week titles, slugs, and signal symbols were scanned first. Existing September names and mechanisms were excluded from selection, including GWRE, NRGV, EGAN, GROW, ASA, LULU, DOCU, NTSK, AMBA, CIEN, CPB, WOOF, CHPT, and the OXM and RARE alternatives already used in current-week research.

Rank	Candidate	Direction	Mispricing	Evidence Freshness	Catalyst Window	Positioning	10/5 Status	Tradeability	Main Rejection Risk
1	Torrid (CURV)	Short	Q2 adjusted EBITDA included an $11.1M tariff refund, while sales and comps fell and debt stayed high	Sep. 3 SEC earnings exhibit and Sep. 4 tape	Next quarter's no-new-refund operating and cash bridge; reporting date unverified	4.10M shares short, 4.4% public float, 17.1 DTC; 10K displayed shares at 1.78% fee	Watchlist	NYSE, but late and regular quote feeds disagree; locate, depth, and exit liquidity incomplete	July inflection, cost savings, and a squeeze can outrun the accounting normalization
2	Duluth Holdings (DLTH)	Short	Q2 profit included a $16.3M tariff refund and sales fell 7.8%, but ex-refund margin improved and the balance sheet is cleaner	Sep. 3 company release and Sep. 4 tape	Second-half demand and margin follow-through	Current short and borrow evidence not verified in this screen	Reject	$4.52 late-tape reference; no current execution audit	Zero ABL debt, about $96M net liquidity, and raised EBITDA guidance make the countercase stronger
3	American Outdoor Brands (AOUT)	Short	A 44.9% session move creates reversal risk, but the latest release also showed higher sales, cash, and EBITDA guidance	Sep. 3 company release and Sep. 4 tape	First post-earnings price acceptance	Current short and borrow evidence not verified in this screen	Reject	$14.48 late-tape reference with 4.37M shares traded; small-cap liquidity is still uncertain	Normalized sales grew 4.3%, cash was $33.3M, and the company has no debt
4	Lands' End (LE)	Short	Adjusted EBITDA fell 25% and the full-year guide narrowed, but revenue, U.S. e-commerce, debt, and buyback support improved	Sep. 3 company release and Sep. 4 tape	Holiday execution and buyback activity	Current short and borrow evidence not verified in this screen	Reject	$10.94 late-tape reference; price and exit quality not audited	The new JV structure and a $100M repurchase authorization create a non-trivial support path

Selected opportunity: CURV, U.S. short research only. Why this one now: It has the cleanest issuer-specific gap between reported profit and recurring operating evidence, a near-term quarter that can test the normalization, and a capital structure that makes weak cash conversion matter. It also has enough short positioning to amplify both directions, which is precisely why execution remains blocked. What should surprise the reader: The adjustment is not a rounding item. The $11.1M refund was 47.7% of reported Q2 adjusted EBITDA and more than twice the quarter's reported $5.2M net income. The headline beat can therefore coexist with a still-shrinking sales base.

Why This Is the Best Opportunity Right Now

CURV is a better short research setup than DLTH because Torrid combines a refund-assisted print with high gross debt, a negative stockholders' deficit, and an equity value that is small relative to the debt claim. DLTH also has a one-time refund, but its latest release reported zero debt on the ABL facility, approximately $96 million of net liquidity, a 15.5% inventory reduction, and a higher adjusted EBITDA guide. AOUT's results improved on both sales and EBITDA, while LE retains balance-sheet and capital-allocation support despite weaker adjusted EBITDA.

CURV is not a simple valuation short. Using the late reference price of $2.44 and the provider market cap of $242.9 million, gross debt of about $304.0 million and cash of $22.0 million imply a rough reference enterprise value of $524.9 million before lease obligations. Subtracting the disclosed $11.1 million refund from the midpoint of the $76 million to $86 million full-year adjusted EBITDA guide gives a rough $69.9 million normalized midpoint, or about 7.5 times reference EV before leases. That multiple is not a standalone proof of overvaluation. It means the short must be won by expectations, demand, and cash conversion, not by pretending that every 7.5 times multiple is expensive.

Why This Can Move More Than 5% Soon

The stock closed the next session above the pre-release price even though the filing made the refund bridge explicit. A future quarter that reports negative comparable sales, a lower recurring margin, or weak cash after interest can force the market to remove the refund from its mental run rate. From the $2.44 reference, a $1.95 base target represents a 20.08% decline. That clears the discovery hurdle, but the adverse path is also large because 4.4% of public float is short, days to cover is high, and displayed borrow was scarce.

The catalyst is observable. Torrid guided Q3 sales to $230 million to $235 million and adjusted EBITDA to $15 million to $20 million. The next filing can show whether the July improvement survives a full quarter, whether reported margin holds without another refund, whether interest and capex consume operating cash, and whether the $304 million gross debt burden is declining. The reporting date was not verified in this run, so the holding window is the next quarterly report and the interim price-acceptance period, not a promised calendar date.

10/5 Asymmetry Gate

The hurdle is measured from a reference price because the latest observation is a late extended-hours feed and independent public pages disagree. It cannot authorize a trade.

Measure	Reference-only result	Interpretation
Reference price	$2.44	Finance-feed snapshot at 2026-09-04 23:25:13Z
Base target	$1.95	20.08% favorable decline from reference
Bottom target	$2.95	20.90% adverse squeeze from reference
Gross reward to adverse risk	0.96:1	Below the required 2:1 ratio
Probability-weighted target	$2.1125	13.42% price-only expected decline before costs
Measurement basis	Reference-only	`entry.price` is null
Status	Watchlist / Reject for current execution	Execution is blocked

The base case clears the 10% favorable-move test. The credible squeeze case is well beyond the 5% adverse bound, and the ratio is below 2:1 before borrow, spread, slippage, fees, dividends, and recall costs. The signal is therefore a research Watchlist, not a qualified short. The probability-weighted price is useful for comparing candidates, not for disguising a failed execution gate.

What Should Surprise the Reader

The tariff refund was not merely an EPS footnote. It increased adjusted EBITDA by $11.1 million, almost half of the reported Q2 figure, and the company explicitly included the benefit in its updated full-year outlook.
A smaller store base does not automatically mean a healthier demand base. Torrid ended Q2 with 457 stores versus 575 a year earlier, while total sales still declined 11.8% and comparable sales declined 6.3%.
The strongest risk to the short is not a hidden accounting trick. It is a real operational turnaround arriving while the short base is forced to cover into a thin loan market.

The Setup

Torrid is a direct-to-consumer apparel, intimates, and accessories brand serving women in North America. Its September 3 SEC exhibit reports Q2 fiscal 2026 net sales of $231.7 million versus $262.8 million a year earlier, a 6.3% comparable-sales decline, and six store closures during the quarter. The store count fell to 457 from 575 a year earlier. Torrid Q2 fiscal 2026 earnings exhibit

The reported profit line improved. Net income was $5.2 million versus $1.6 million, and adjusted EBITDA was $23.3 million versus $21.5 million. The same filing says the company received $11.4 million in IEEPA tariff benefits, including $11.1 million recorded as a reduction of cost of goods sold and $0.3 million as interest income. Gross margin was 38.7%, but 33.9% excluding the refund. Adjusted EBITDA was $12.1 million, or 5.2% of sales, excluding the refund. These are the recurring-evidence numbers that matter for a short. SEC Exhibit 99.1

The counterevidence is real. Management says sales trends improved meaningfully as the quarter progressed, July was a clear inflection point, the product assortment was being corrected, footwear inventory was strengthening, third-party marketplaces were expanding, and the store-optimization project was nearly complete. Torrid also says it remains on track for approximately $40 million of fiscal-year expense savings. Those statements are company evidence and future operating claims, not proof that the turnaround has failed. The short needs the next filing to test them.

The capital structure raises the sensitivity of that test. The release balance sheet shows $22.0 million of cash, $39.7 million of borrowings under the credit facility, $16.1 million of current term-loan debt, and $248.2 million of noncurrent debt. That is approximately $304.0 million of gross debt. Current and noncurrent operating lease liabilities add another $113.5 million. Six-month operating cash flow was $10.1 million, while cash interest paid was $12.6 million. Torrid Q2 SEC earnings exhibit and balance sheet

The Market Price

Field	Observation	Timestamp or date	Source and limitation
CURV late-tape price	$2.44	2026-09-04 23:25:13Z	Yahoo Finance quote; late or extended-hours context, not a verified executable bid
Late-tape change	+$0.18, +8.07175%	Same snapshot	Same source; the percentage is provider output
Late-tape open / high / low	$2.54 / $2.83 / $2.33	Same snapshot	Same source
Late-tape volume	3,805,298 shares	Same snapshot	Same source; volume quality and exit capacity unverified
Provider market cap	$242,892,240	Same snapshot	Same source; market-cap fields can use a different share timestamp
Regular-session context	$2.50, +12.108%, 3,572,146 shares	2026-09-04 15:42:15 EDT	ChartExchange historical tape; page was captured before the late-tape observation
Current share count	99.69M shares outstanding	August 1, 2026	Torrid SEC earnings exhibit
Regular-session price acceptance	Not verified	Post-earnings session	The late and regular feeds disagree and no independent top-of-book snapshot was available

The two price observations are close enough for research and different enough to block an order. The late feed returned $2.44 while the regular-session page showed $2.50, and the provider's prior-close basis also differs. No current bid, ask, spread, depth, venue quality, short-sale restriction state, or realistic exit-liquidity audit was available. The $2.44 figure is a reference level only.

The Mispricing

Fact: The Q2 release reports $23.3 million of adjusted EBITDA and $5.2 million of net income, but it separately identifies $11.1 million of tariff-refund benefit inside adjusted EBITDA and $0.3 million of related interest income. SEC Exhibit 99.1

Fact: Gross margin excluding the tariff refund was 33.9%, compared with reported gross margin of 38.7%. Sales declined 11.8% and comparable sales declined 6.3%. SEC Exhibit 99.1

Fact: The company raised full-year adjusted EBITDA guidance to $76 million to $86 million and says the outlook includes the tariff refund recognized in Q2. SEC Exhibit 99.1

Inference: The market may be treating the earnings beat as evidence of recurring margin repair before a no-refund quarter confirms it.

Reasonable assumption: If Q3 comparable sales remain negative, if normalized gross margin does not hold, and if interest and capex absorb operating cash, equity holders will demand a lower price even if management's store and marketing changes are directionally sensible.

Unknown: The exact share of the full-year guide attributable to the refund, the Q3 margin bridge, future tariff recoveries, current debt availability, lender covenant headroom, and the date of the next quarterly report. The company says the Q3 and full-year outlook is based on current assumptions and does not incorporate further tariff volatility.

The market is not necessarily wrong to believe July improved. The disagreement is narrower: whether one improved month and a completed store reset are enough to capitalize the next four quarters of recurring EBITDA while the revenue base is still shrinking and senior claims remain large.

The Positioning

MarketBeat reports 4.10 million shares sold short, 4.4% of public float, and 17.1 days to cover as of the August 14 settlement. This is not an extreme percentage of float, but the days-to-cover measure is high relative to the stock's ordinary volume. MarketBeat CURV short interest

The same-day Interactive Brokers snapshot displayed 10,000 shares available at a 1.78% borrow fee as of 2026-09-04 05:59:55 EDT. A separate public loan page later showed about 8,000 shares at roughly 1.8%. These are indicative public snapshots, not a broker-level locate or a promise that shares remain available. ChartExchange CURV borrow fee, CurvedTrading CURV borrow context

Current positioning score: 4/5. The short base is large enough to amplify a weak next print, and the displayed loan supply is small enough that a good operational update can create a squeeze. The score is not a bullish or bearish conclusion. It is a path-risk warning. Direct holder concentration, options dealer gamma, intraday short volume, recalls, and actual locate terms are not verified.

The Catalyst

Step	State	Timing	Observable test	Failure mode
Q2 earnings print	Filed	September 3, 2026	Separate the $11.1M refund from Q2 EBITDA and gross margin	The market capitalizes the headline and ignores the bridge
First regular-session acceptance	Unverified	Next regular session	Check whether price holds the post-earnings gain on ordinary volume, spread, and depth	A late mark fades, or a thin squeeze extends
Q3 operating report	Company guide filed; report date unverified	Next quarterly report	Test sales of $230M to $235M, adjusted EBITDA of $15M to $20M, comps, margin without a new refund, and cash after interest	July inflection does not persist, or reported EBITDA again needs a special item
Debt and liquidity bridge	Partly filed	Next 10-Q and lender updates	Reconcile revolver draws, term-loan balance, cash, interest, covenant headroom, leases, and capex	Cash flow remains too small to reduce senior claims
Store and customer reset	Management evidence	Q3 and Q4 operating updates	Test traffic, conversion, repeat customers, marketplace economics, inventory age, and closures	Fewer stores reduce cost but also shrink the revenue base

The cheapest falsification sequence is the next Q3 comp and revenue table, the ex-refund gross-margin and EBITDA bridge, operating cash after interest and capex, the debt and revolver balance, and a fresh broker-level locate and exit audit. A new marketing initiative or marketplace logo is not enough.

The Payoff

These are analyst scenarios, not sourced price targets or valuation floors. They use one reference price and one holding window: the next quarterly report plus the intervening four to eight weeks.

Top case: $1.60 if the no-refund quarter shows renewed negative comps, normalized margin pressure, weak cash after interest and capex, or a higher revolver balance, while the post-earnings gain reverses.
Base case: $1.95 if Q3 confirms a still-shrinking sales base and only partial recurring margin repair, but the company remains liquid enough to avoid an immediate restructuring.
Bottom case: $2.95 if July's improvement becomes positive comparable sales, normalized EBITDA tracks the guide, cash and debt improve, or short covering runs into scarce stock.

Using the $2.44 reference price:

0.25 x $1.60 + 0.50 x $1.95 + 0.25 x $2.95 = $2.1125

The price-only weighted target implies a 13.42% decline before costs. Net probability-weighted EV cannot be computed responsibly because the reference quote is not an executable entry and the current spread, locate, recall, depth, exit liquidity, and carrying costs are unverified. The debt and lease obligations also need a common timestamp with the market data for a defensible equity valuation bridge.

Price Target and Probability Map

Scenario	Probability	Target / Level	Return / Payoff	Horizon	Conditions	Evidence Quality
Top Case	25%	$1.60	34.43% favorable decline	Next quarterly report and four to eight weeks	Negative comp or margin without refund, weak cash after interest and capex, debt pressure	Medium
Base Case	50%	$1.95	20.08% favorable decline	Same window	Sales remain weak, recurring EBITDA is below the headline run rate, no immediate liquidity event	Medium
Bottom Case	25%	$2.95	20.90% adverse squeeze	Same window	July inflection becomes durable, cash and debt improve, or short covering dominates	Medium
Invalidation	n/a	Q3 shows durable positive comps and recurring cash improvement, or the evidence gap is resolved in favor of the countercase	Thesis broken, not a price stop	Before or during the catalyst ladder	The next filing provides the missing bridge	High when filed

Probability-weighted expected value: $2.1125 price-only reference calculation, or a 13.42% expected decline before costs. Net EV is not computable with aligned executable data. Current market level and timestamp: $2.44 at 2026-09-04 23:25:13Z, late-tape reference. Primary instrument: CURV common stock, NYSE, unlevered only after all gates pass. 10/5 favorable base move: 20.08% reference-only decline. 10/5 credible adverse move: 20.90% reference-only squeeze. 10/5 measurement basis: Reference-only, not verified entry. 10/5 status: Watchlist / Reject for current execution. Confidence: Medium on the refund, margin, sales, debt, and guidance facts; low on executable market structure and the timing of the next report.

The Kill Shot

The strongest counterparty argument is that Torrid's second quarter may be the trough of a real turnaround. Management says July was a clear inflection point, the product mix has been corrected, footwear inventory is stronger, the store reset is largely complete, and customer acquisition and retention programs are improving. The company generated $10.1 million of operating cash in the first half, reduced inventory year over year, and guided to $15 million to $20 million of Q3 adjusted EBITDA. If those improvements are recurring, the $2.44 reference price may already discount too much distress.

The capital structure does not automatically defeat that counterargument. A rough 7.5 times reference EV to refund-normalized full-year EBITDA midpoint is not a distressed multiple. The company has $22.0 million of cash and $74.4 million of total liquidity including revolver capacity, and it says the store optimization is producing savings. A short at $2.44 can lose even if Q2's refund is non-recurring, because the market only needs to believe that the next quarter is improving.

The load-bearing assumption is that negative demand and high senior claims will matter more to the next price-setting marginal buyer than the company's July and cost-reset narrative. The cheapest test is a filed Q3 bridge that shows comparable sales, recurring gross margin, cash after interest and capex, and debt together. Until then, this is a plausible expectation mismatch, not a completed short.

What Could Go Wrong

Operational recovery: Positive comparable sales and stable ex-refund margin can support a rerating before debt reduction is visible.
Positioning squeeze: 17.1 days to cover and only 10,000 displayed shares in one loan snapshot can turn a modest beat into a discontinuous rise.
Borrow and recall: The fee is not a guarantee of availability. A recall, higher fee, or forced buy-in can create losses without a thesis change.
Further tariff recovery: A new refund or favorable tariff ruling could again lift gross margin and cash, although the company did not include further tariff volatility in its outlook.
Liquidity and gaps: A small-cap NYSE stock can gap through a stop, halt, or re-open at a materially different price. A 5% stop does not control that risk.
Debt flexibility: Lenders may extend or amend facilities, or the company may obtain new liquidity, delaying the equity response to weak operating cash.
Instrument mismatch: A put may cap loss but can lose to implied volatility, time decay, wide spreads, or a delayed catalyst. No current option-chain audit was completed.

What Would Prove This Wrong

The short thesis is invalidated by evidence, not by a single intraday bounce. The following would remove it:

The next quarterly filing shows positive comparable sales or a clear sales inflection, recurring gross margin that does not depend on a new refund, and adjusted EBITDA in the guided range with operating cash positive after interest and capex.
The debt bridge shows falling revolver and term-loan balances, adequate covenant headroom, and no equity issuance or new senior claim needed to fund ordinary operations.
Inventory, traffic, conversion, repeat behavior, and marketplace economics confirm that the July improvement was not only promotional timing or a smaller store base.
A fresh regular-session tape holds above the $2.95 stress level with verified depth and exit liquidity while borrow remains available. This is an execution cancellation and squeeze-control condition even if the accounting thesis remains unresolved.

Risk Audit

Information risk: The Q2 release is primary evidence, but the report date for Q3, the exact refund allocation inside the full-year guide, and future tariff treatment are not verified.

Gap risk: The reference quote is late-tape context. A new session can open above the modeled bottom before a short can be located or covered.

Liquidity risk: Current bid, ask, spread, depth, venue quality, short-sale state, volume quality, and exit liquidity are missing. The $2.44 reference is not a promise of fill quality.

Borrow risk: Public snapshots show scarce displayed supply but do not establish a broker locate, recall terms, fee stability, or borrow duration.

Balance-sheet risk: Gross debt is about $304.0 million and operating lease liabilities about $113.5 million. Cash of $22.0 million and total liquidity of $74.4 million do not equal durable common equity value.

Thesis risk: The company may be right about July. A store reset, lower promotional activity, better inventory, and expense savings can repair margins faster than the sales decline damages them.

Path risk: A correct medium-term short can lose money if the stock squeezes first, the next report is delayed, or a financing amendment buys time.

Execution cancellation: No locate, no current top-of-book, no aligned regular-session quote, no realistic exit, or no acceptable borrow means no short. The signal must remain entry.price: null and execution.can_execute: false.

Best Trade Strategy

Current decision: No executable trade. Keep CURV on a Watchlist.

Primary expression: Unlevered CURV common stock only after a fresh regular-session quote, broker-level locate, current borrow fee and recall terms, acceptable bid and ask, spread, depth, venue quality, volume quality, and realistic exit liquidity are verified. The evidence trigger is a no-refund Q3 bridge showing weak comps, normalized margin pressure, and cash that does not cover interest, capex, and debt needs. That trigger is research language, not an order.

Reference-only levels: Top $1.60, base $1.95, bottom $2.95. The invalidation state is recurring operating and cash improvement, not a mechanical price stop.

Horizon: Next quarterly report and four to eight weeks around the operating update. The report date was not verified, so there is no calendar promise.

Do not trade when: the only quote is extended hours; feeds disagree; the loan is unavailable or recall terms are unknown; the spread or depth is wide; the price gaps above the $2.95 stress level; the next filing confirms recurring cash and debt improvement; or an option expression lacks a verified live chain, open interest, implied volatility, spread, expiry, and maximum-loss calculation.

Alternative: A defined-loss put or put spread could cap squeeze exposure, but no live option chain or pricing audit was available. It is not approved as a substitute for the missing stock and borrow checks.

Sources

Torrid Q2 fiscal 2026 SEC Exhibit 99.1 - primary results, refund bridge, margin, guidance, cash, debt, liquidity, shares, and cash flow, September 3, 2026.
Torrid September 3 Form 8-K - filing that furnished the earnings release.
Yahoo Finance CURV quote - late-tape price and market-data snapshot, 2026-09-04 23:25:13Z.
ChartExchange CURV historical tape - regular-session price and volume context, 2026-09-04.
MarketBeat CURV short interest - public short-interest and days-to-cover context, August 14, 2026 settlement.
ChartExchange CURV borrow fee - indicative Interactive Brokers availability and fee snapshot, September 4, 2026.
CurvedTrading CURV borrow context - additional public loan context, September 4, 2026.
Duluth Holdings Q2 fiscal 2026 results - ranked alternative with refund, sales, liquidity, and guidance evidence, September 3, 2026.
Duluth Holdings September 4 Form 10-Q - ranked alternative filing and share-count context, September 4, 2026.
American Outdoor Brands Q1 fiscal 2027 results - ranked alternative with normalized sales, cash, and EBITDA guidance evidence, September 3, 2026.
Lands' End Q2 fiscal 2026 results - ranked alternative with adjusted EBITDA, revenue, JV, and repurchase evidence, September 3, 2026.

Research Quality Scorecard

Dimension	Score	Evidence
Primary operating evidence	5/5	SEC exhibit gives refund, ex-refund margin and EBITDA, sales, comps, guidance, cash, debt, liquidity, and shares
Market-price freshness	3/5	Late-tape and regular-session observations are current but do not reconcile to one executable quote
Positioning evidence	4/5	Short interest and same-day public borrow context exist; locate, recall, dealer, and holder data are incomplete
Catalyst observability	4/5	Q3 sales and EBITDA ranges are filed; next report date and ex-refund bridge are not verified
Cash and capital-stack bridge	4/5	Debt, leases, cash, liquidity, interest paid, and operating cash are visible, but current state can change before the next filing
Execution quality	1/5	Entry, bid, ask, spread, depth, venue, and exit-liquidity audit are missing
Total	21/30	Strong research case, failed execution case

Canonical classification: Watchlist. The reference-only base decline is greater than 10%, but the credible adverse squeeze exceeds 5%, gross reward to adverse risk is below 2:1, and required short-execution checks are unverified. This is not Trade-qualified.

Bottom Line

CURV's Q2 print contains a real operating improvement and a large one-time refund. The mispricing opportunity is the risk that the market capitalizes the refund-assisted headline before recurring demand, margin, cash, and debt evidence catches up. The short is attractive as a research question, not as an order today. Wait for the no-refund Q3 bridge and a verified borrow and liquidity state. Until both arrive, the honest signal is watch, with null entry and execution blocked.

AI Illustration Prompt

A realistic editorial financial-magazine illustration of a small urban apparel retailer storefront labeled "Torrid" under sharp studio light. Above the entrance, a bright gold customs-rebate receipt labeled "$11.1M tariff refund" inflates a transparent glass balloon marked "Adjusted EBITDA $23.3M". Beneath the glass, the building's foundation is visibly made of heavy steel debt plates labeled "$304M gross debt" and a narrow cash drawer labeled "$22M cash". Through the window, several mannequins stand beside a fading sales ledger marked "Sales -11.8%" and "Comps -6.3%", while one lit sign reads "July inflection" to show the credible counterargument. In the foreground, a short-sale borrow ticket shows a nearly empty availability gauge and a small warning light, symbolizing squeeze and locate risk. Composition is restrained, analytical, and tactile, with slate blue, charcoal, muted burgundy, and one controlled amber accent; no people, no generic candlestick chart, no neon, no hype. Wide horizontal composition suitable for a premium markets publication cover, subtle readable watermark text "The Mispricing Desk" at bottom right.