Answer · confidence: partial
Do people using private routes for trades lose less to sandwich attacks after fees?
Research question: Do retail users who route transactions through private RPCs or protected order flow services suffer measurably lower sandwich losses, and do measurement studies quantify the net effect after fees and failed inclusions?
The evidence is mixed. A 2023 study found that private RPCs appeared to completely eliminate adverse slippage, with only 3 of 8,294 private USDC swaps showing negative adverse slippage above $5 13. Private routing is not a reliable shield, though: private-channel sandwiches caused about $409,237 in user losses in November–December 2024 13. No source reports a measured sandwich-loss reduction for Flashbots Protect itself 24, and the study tracking victims' migration to private routing does not report net savings after fees or failed inclusions 7. On costs, the 2026 documentation says Flashbots Protect includes only non-reverting transactions, so users do not pay for failed ones 23, while a 2025 benchmark found that two-thirds of some providers' transactions were excluded from blocks because they were less profitable after fees 8. The newest fact is dated 19 August 2026, about two months before October 2026 2223.
Checked 2026-10-18. Only documented and supported facts are used.
Dig deeper on this questionSources
[1] supported
Private routing does not reliably protect users from sandwich attacks, and it can concentrate exploitable order flow.
"private routing, while increasingly adopted, does not offer reliable protection against sandwich attacks and can concentrate ex-"Sandwiched and Silent: Behavioral Adaptation and Private Channel Exploitation in Ethereum MEV
[2] supported
Private and encrypted mempools hide transactions to stop sandwich attacks, but the pair, the direction, and a coarse size range can still leak, so they do not hide everything.
"atransaction’spair,direction,andacoarserangeforitssizecanstill"How Much Must a Private Mempool Hide? Exact Leakage Thresholds for Sandwich Attacks
[3] supported
In November and December 2024, sandwich attacks on private transactions caused about $409,237 in user losses while attackers earned about $293,786.
"with aggregate user losses of $409,236.97 and attacker profits of $293,785.95."Sandwiched and Silent: Behavioral Adaptation and Private Channel Exploitation in Ethereum MEV
[4] supported
One operator (0xae2f...ae13) accounted for about 65% of private frontruns, indicating strong centralization in private-path exploitation.
"a single operator (0xae2f...ae13) accounts for about65%ofprivatefrontruns,pointingtostrongcentralizationinprivate-path"Sandwiched and Silent: Behavioral Adaptation and Private Channel Exploitation in Ethereum MEV
[5] supported
Private routing is an alternative path rather than an absolute shield, since inclusion policies, cross-domain leakage, and intermediary interactions can still enable adversarial ordering.
"private routing is an alternative path for order flow, not an absolute shield: inclusion policies, cross-domain leakage, and strategic interac-"Sandwiched and Silent: Behavioral Adaptation and Private Channel Exploitation in Ethereum MEV
[6] supported
Sandwich attacks on Ethereum remain common, and they are not confined to simple Uniswap v2 paths.
"Sandwiching remainscommononEthereum[28,25]andisnotconfinedtosimpleUniswapv2paths"How Much Must a Private Mempool Hide? Exact Leakage Thresholds for Sandwich Attacks
[7] supported
After a public sandwich, 37.2% of victims migrate to private routing within 60 days, rising with repeated exposure, though the study does not report net savings after fees or failed inclusions.
"exposed. After a public sandwich, 37.2% of victims migrate to private rout-"Sandwiched and Silent: Behavioral Adaptation and Private Channel Exploitation in Ethereum MEV
[8] documented
In a specific example provided in the study, two-thirds of Merkle and Blink's transactions were excluded from blocks due to lower profitability relative to the fees, whereas Flashbots and MEV Blocker flows were favored due to fewer additional costs.
"The example shows 2/3 of Merkle and Blink’s transactions were excluded due to lower profitability. Flashbots and MEV Blocker flows were favored due to fewer additional costs."Private MEV Protection RPCs: Benchmark Stud
[9] supported
Flashbots Protect does not have a fee mechanism to monetize order flow, meaning builders do not need to account for RPC fees when deciding whether to include transactions.
"Flashbots RPC does not have a fee mechanism in place to monetize orderflow, and therefore builders do not need to take anything into account when treating their txs."Private MEV Protection RPCs: Benchmark Stud
[10] documented
Flashbots provides gas rebates to order flow providers by redistributing excess gas fees, but the calculation for how much is rebated is not known upfront.
"Additionally, Flashbots provides gas rebates to order flow providers by redistributing excess gas fees that are attributable to the orderflow originators every once in a while. However, it is important to note that the calculation on how much “excess gas fee” you will be rebated is not known upfront."Private MEV Protection RPCs: Benchmark Stud
[11] supported
Merkle and Blink operate as permissioned Order Flow Auctions where only approved searchers can access the private mempool and backrun user transactions, unlike the permissionless model of Flashbots.
"Launched in late 2023, Merkle and Blink are private RPC services designed to prevent MEV attacks by implementing a permissioned Order Flow Auction (OFA). In both models, only approved searchers can access the private mempool and backrun user transactions."Private MEV Protection RPCs: Benchmark Stud
[12] documented
Private MEV protection services offer gas rebates to order flow providers by redistributing a portion of the subscription fees paid by builders.
"Both services offer gas rebates to order flow pro viders by redistributing part of the builder-paid subscription fees."(PDF) Private MEV Protection RPCs: Benchmark Stud
[13] documented
Private RPCs appear to completely eliminate adverse slippage, as only 3 out of 8,294 private USDC swaps and 12 out of 8,629 private PEPE swaps experienced negative adverse slippage greater than $5.
"In fact, out of the 8294 (8629) private interface swaps in USDC (PEPE), only 3 (12) have negative adverse slippage that is larger than $5. That is, private RPCs seem to completely eliminate adverse slippage."Don’t Let MEV Slip:The Costs of Swapping on the Uniswap Protocol
[14] documented
The use of private RPCs does not significantly increase collision slippage, indicating that potential latency issues do not negatively impact average slippage for these transactions.
"Furthermore, the coefficient on collision slippage is not significant, meaning that the potential increase in latency has no effect on average collision slippage."Don’t Let MEV Slip:The Costs of Swapping on the Uniswap Protocol
[15] documented
The effect of order size on adversarial slippage is stronger for public transactions than for private ones.
"We find that the effect of order size on adversarial slippage is stronger for public transactions, yet gas price interacted wit"Don’t Let MEV Slip:The Costs of Swapping on the Uniswap Protocol
[16] documented
Previous research has shown that transactions sent to private RPCs may incur higher average transaction costs compared to public transactions.
"[ 37 ] show that transactions sent to private RPCs may have on average higher transaction costs than public transactions."Don’t Let MEV Slip:The Costs of Swapping on the Uniswap Protocol
[17] documented
Flashbots Protect's 2021 documentation claimed that its RPC would keep user transactions out of the public mempool, where sandwich bots could see them.
"Frontrunning protection: transactions will not be seen by hungry sandwich bots in the public mempool."What changed on docs.flashbots.net/flashbots-protect/overview between 2021-10-06 and 2026-08-19
[18] supported
Protected order flow services such as Flashbots, Eden, and OpenMEV do not eliminate attacks; they shift trust to a different set of parties.
"The drawback with thesesolutions is that attacks are noteliminated, buttrustisdelegatedtoadifferentsetofparties."Eating Sandwiches: Modular and Lightweight Elimination of Transaction Reordering Attacks (2023)
[19] documented
The 2021 documentation warned that protected transactions could still be included in uncled blocks, emitted to the public mempool, and then included on-chain, which means protection is not guaranteed.
"Note: transactions could be included in uncled blocks, emitted to the mempool, and then included on-chain."What changed on docs.flashbots.net/flashbots-protect/overview between 2021-10-06 and 2026-08-19
[20] supported
Of the 99,928 sandwich attacks observed between November 23rd, 2021 and March 23rd, 2022, about 81% were performed using Flashbots.
"Outofthese99,928sandwiches,81,089(81.15%)wereperformed"A Flash(bot) in the Pan: Measuring Maximal Extractable Value in Private Pools
[21] supported
Sandwich attacks caused about 174M USD in losses for Ethereum users over a 33-month span, according to the cited study.
"Sandwich attacks are one of the most common types of MEV [38] accounting for a loss of 174M USD over the span of 33 months [48] for users of Ethereum."Eating Sandwiches: Modular and Lightweight Elimination of Transaction Reordering Attacks (2023)
[22] supported
By 2026 the documentation described Flashbots Protect as sending transactions to a private Flashbots mempool where they are hidden from frontrunning and sandwich bots.
"Frontrunning protection: Transactions are sent to a private Flashbots mempool where they will be hidden from frontrunning and sandwich bots."What changed on docs.flashbots.net/flashbots-protect/overview between 2021-10-06 and 2026-08-19
[23] supported
The 2026 documentation states that Flashbots Protect only includes transactions that do not revert, so users do not pay fees for failed transactions.
"No failed transactions: Transactions are only included in the block if they do not revert."What changed on docs.flashbots.net/flashbots-protect/overview between 2021-10-06 and 2026-08-19
[24] supported
Flashbots Protect is a free private RPC that sends transactions into the Flashbots builder lane, and the source gives no measured sandwich-loss reduction for it.
"Drops your transaction into the Flashbots builder lane."MEV Bot Guide 2026: Types, Flashbots and Protection