Imagine you're using a VIP entrance to skip the line at a nightclub. You paid extra, you got the special wristband, you walk through the side door — and the bouncer is texting your arrival time to the scalpers out front, who cut in anyway. That's what's happening with "protected" order flow on blockchains. Users pay for private RPCs and native protections specifically to avoid being sandwiched — having a bot front-run and back-run their trades to extract value — and this paper shows those protections are failing at industrial scale. The committed claim: existing front-running protections across Ethereum, Solana, Tron, Base, Arbitrum, and Monad provide substantially weaker guarantees than users assume, and the authors prove it with the first multi-chain, multi-year empirical measurement. This isn't a theoretical argument. They found 28.0 million sandwich attacks on Solana alone against transactions that were supposed to be shielded. On Ethereum, 30,607 attacks hit protected order flow, with another 2,875 victims exposed through reorged blocks. Tron saw 38,567 and Base 1,889. The numbers across chains are wildly different, but the conclusion is the same everywhere: the protections leak. The detection methodology is where the engineering gets interesting. Classical sandwich detection looks for tight attacker-victim-attacker triplets within a single block. These authors had to go wider — the attacks against protected flow don't cluster tightly around victims and often span multiple blocks. They built heuristics that capture wide attacks both within and across blocks, then filtered on bot behavior patterns to separate real sandwiches from legitimate trading that happens to look similar. This is a measurement paper, not a defense paper, and the measurement instrument matters because undercounting is the default failure mode. The exposure surface differs by chain in illuminating ways. Solana's vulnerabilities are at the validator and application level — the sheer volume (28M attacks) reflects its architecture's openness to ordering manipulation. Ethereum's leaks come through order-flow auctions and block reorganizations, more sophisticated but lower volume. Tron's first-come-first-served ordering, which sounds fair on paper, fails against latency-based front-running — whoever has the faster connection wins. Base had an actual RPC bug exposing pending transactions, plus predictable victim behavior that bots exploited. Each chain's protection mechanism fails in a way specific to its design choices, which makes the cross-chain comparison genuinely useful rather than just a bigger dataset. A striking structural finding: outside Solana, a small number of entities carry out the majority of attacks. This is not a decentralized cottage industry — it's concentrated extraction by sophisticated actors who understand the protection mechanisms intimately enough to bypass them. On Solana, the attack surface is broad enough that many participants can sandwich; on Ethereum and Tron, it takes specialized knowledge of auction mechanics and block production to attack protected flow. The integrity of this work rests on its heuristic detection pipeline. The authors acknowledge this: heuristic-based detection can both over-count (false positives from legitimate trading) and under-count (attacks that don't match the pattern). They address this through bot-behavior filtering, but there's no ground truth to validate against — you can't ask an attacker to confirm. The three-year longitudinal window (covering evolving protection mechanisms) and six-chain breadth are genuine strengths, but independent replication with different heuristics would substantially strengthen the conclusions. What this paper does NOT provide is a solution. It's a diagnostic, not a prescription. The implicit milestone is end-to-end protection that provably resists these attacks — something like encrypted mempools with commit-reveal schemes or threshold encryption of transaction content. The gap between "we measured the problem" and "here's a defense that works" is where the field's real fight lives, and this paper sharpens the fight by showing exactly where each chain's armor has gaps.
