Full article
Hidden transactions and unexploitable transactions are different
DefiLlama’s October 1, 2026 analysis compares private routing with protocol-level ordering. Moving a transaction away from a public pending pool changes who can observe it before execution. It does not, by itself, answer what the remaining participants can do with it. Our analytical framework separates three questions: who sees the request, who can influence its position, and whose behavior must be trusted. Calling a path private provides an incomplete answer to those questions. The useful security description is the actual distribution of information and control.
Sandwiching targets the conditions of execution
A sandwich strategy uses positions around a target transaction to seek profit from the execution conditions it faces. Consider a hypothetical swap allowing a two-percentage-point deviation from its estimate. That tolerance is an acceptance boundary, not an instruction to lose two percent on every trade. Someone able to move the price before execution may find room within that boundary. Tightening the limit changes the trade-off: it can also cause a transaction to fail during ordinary movement. The relevant comparison includes both adverse execution and the chance that a legitimate request cannot complete.
Private routing changes the audience
Imagine announcing a purchase in a public square versus submitting it to a designated desk. The second arrangement reduces the audience, while making the desk’s handling of information more important. In a hypothetical comparison, one route forwards to a single processor and another to several. They could offer different opportunities for completion and different exposure of the request. This is a comparison model, not an allegation about any provider. Assessing a real route requires knowing its forwarding arrangements and information practices instead of inferring every security property from its name.
Ordering rules change the economic calculation
The source uses THORChain’s ordering and streaming-swap mechanisms as a case study in attack economics. This introduces a different question from secrecy: even when a target is visible, can an adversary obtain the positions needed for a profitable strategy? Consider a hypothetical gross gain of five units. If necessary fees and price exposure cost more than five, an apparent price gap is not enough to make the strategy attractive. This is an illustration of net economics, not a measured result for THORChain or an estimate of returns available to any participant.
A local defense does not establish universal safety
Making a particular ordering strategy unattractive does not answer every question about an exchange service. Availability, changing liquidity, software behavior, and settlement across assets are separate dimensions. Our interpretation is that private channels and protocol rules can each alter some attack conditions, with effectiveness depending on implementation and transaction context. The strongest explanation identifies the behavior being constrained and the assumptions required for the constraint to work. It should also identify what happens when those assumptions change. That is more informative than extending a narrow mechanism claim into a promise that an entire platform is risk-free.