1. What the system observed. On 2026-08-31, SNDK's funding rate on Coinbase International (INTX) cleared the entry threshold (+20.1% APR) with acceptable basis (+0.06%). A position opened: short the SNDK perp, long SNDK stock via Interactive Brokers. Entry prices: perp $1,491.34, IB $1,486.33.

2. Why the trade was entered. Every entry gate passed — a routine entry, nothing unusual.

3. What the model expected. Funding income with price risk cancelled between the two legs — the core premise of the strategy.

4. What changed. On 2026-09-06 at 15:04 UTC, funding flipped to an emergency level (FR = −60%). The perp leg closed immediately at $1,764.54 — SNDK had risen sharply, about 18.3%, over the six-day hold. The IB leg deferred to the next market open, about 22.75 hours later.

5. Funding, price movements, and costs. Funding earned over the six-day hold: +$27.38, the largest funding component of any trade in this batch — a longer hold earning proportionally more. Fees: −$11.50. When the IB leg finally closed on 2026-09-07 at 13:49 UTC, its price was $1,740.00 — and by coincidence (or rather, by the 2026-09-01 stale-quote fix doing its job) that was exactly the reference price recorded at the moment the perp closed. Naked-leg P&L: $0.00. The unhedged window itself cost nothing.

6. Final result. Net P&L (strategy accounting, which only counts the naked window): +$15.88 — a small nominal win. But look at the full hold: the perp leg (short) lost −$1,831.91 as SNDK's perp price rose 18.3%. The IB leg (long) gained +$1,706.73 as the stock rose only about 17.1% over the same window — a full percentage point less. That gap, multiplied across a $10,000 position, is real money: True P&L (all-in): −$109.30. The trade was a loss, once every dollar of price movement on both legs is counted — just not in the piece the strategy's own accounting tracks.

7. What was learned. This is a different failure mode from the naked-leg window covered in What happens when the perp closes before the hedge — this trade's naked window contributed nothing to the loss. The loss came from the perp and its hedge simply not moving together over the full six-day hold, while both legs were supposedly delta-neutral and hedged the entire time. This is basis drift in the fully-hedged window, not naked-leg exposure — a risk Net P&L doesn't measure at all, because it isn't designed to. It's exactly why this site always reports Net P&L and True P&L side by side: a trade can look like a small win on one and a real loss on the other, and both numbers are honest — they're just answering different questions. See When funding profit differs from total trade profit.