Why "Delta-Neutral" Positions Still Get Liquidated.

Delta-neutral positions get liquidated because neutrality is a property of the combined position while liquidation is decided leg by leg, by margin engines that cannot see the other leg. A long on venue A hedged by a short on venue B nets to zero exposure on your spreadsheet, but venue A’s engine sees only an unhedged long against A’s mark price, and venue B’s sees only an unhedged short against B’s. Neither engine knows the hedge exists, neither can draw on the other’s collateral, and either can liquidate its leg alone. The construction is neutral; the legs are naked.

The failure sequence is always the same shape. A fast move arrives. The two venues’ mark prices, computed by different engines from different indexes per mark price versus index price, diverge from each other, the behavior covered in what is cross-venue mark price divergence, and divergence is widest precisely during the violent moves the hedge was supposed to make irrelevant. One leg’s mark crosses its liquidation threshold while the combined position is still roughly flat. That leg is forcibly closed, at the venue’s quote plus a penalty, in a fast market. The survivor is no longer a hedge: it is a naked directional position you did not choose, at full leverage, mid-move, and the loss that follows usually dwarfs the funding the construction was built to collect.

Each structural cause suggests its own defense. Margin segregation means collateral cannot flow to the leg under pressure, so the defense is buffer: run leverage low enough that each leg independently survives the worst divergence you can model, then more, since the divergence that gets you is the one you did not model. Mark divergence means your survival math must be done against each venue’s mark, not against the spot chart, with liquidation distance on each leg sized per liquidation distance versus funding capture, where the asymmetry between basis-point capture and percentage-point liquidation is laid bare. Independent engines mean monitoring is two jobs, both marks, both margin ratios, especially around the settlement times when flows concentrate. And thin markets amplify everything, wider divergence, worse forced-exit prices, per the low-volume problem, which is why the screener-flagged opportunities on fresh listings are the most likely to produce exactly this failure, the anatomy in the 1,500% APR that doesn’t exist.

The sturdier alternative exists and is worth naming: the same-venue construction, spot plus perp under one cross-margined engine per cash and carry versus funding arbitrage, where the hedge is visible to the only engine that matters and the spot leg cannot be liquidated at all. The cross-venue version pays better on capital precisely because it carries this failure mode, and the honest way to hold it is to treat the liquidation-of-one-leg scenario not as a tail risk but as the cost of the strategy, sized so that when it fires, it stings instead of ends you.