Priority Gas Auctions and Block Building
When many bots spot the same profitable trade at once, they used to bid the transaction fee up against each other in public until nearly all the profit was gone — modern block-building infrastructure exists largely to fix the waste that created.
Prerequisites: MEV Searcher Strategies, Blockchains and Consensus
Early Ethereum bots competing for the same arbitrage or liquidation opportunity had only one lever to win: pay a higher gas price (transaction fee) than everyone else, since miners simply picked the highest-paying transactions first. When several bots detect the same opportunity, they end up in a priority gas auction — a bidding war fought in public, transaction by transaction, each one rebidding slightly higher than the last, visible to everyone in the mempool before it confirms. The winner gets the opportunity; every loser has still paid gas for a transaction that failed or was never included, and the winner has typically bid away most of the profit just to secure the win.
A priority gas auction is what happens when multiple bots chase the same fixed opportunity using only public fee-bidding to compete — it burns value in wasted gas and drives the winner's bid up toward the full size of the opportunity, leaving the original profit source, not the bots, as the one who captured almost nothing better.
From public bidding wars to private order flow
The fix that emerged was private order flow: instead of broadcasting a transaction to the public mempool, a searcher (a bot operator) submits a bundle privately to a specialized block builder, along with a bid for how much of the profit it is willing to pay to be included. Builders compete to assemble the most valuable block and pass it to whoever proposes the next block (the validator), a structure known as proposer-builder separation. This turns a repeated public rebidding war into a single sealed-bid auction, which wastes far less gas and lets searchers keep strategies private from competitors who would otherwise see every bid in the mempool.
Worked example
Two bots spot the same $500 arbitrage opportunity at the same time. In the old public-mempool world: bot A bids a gas fee equivalent to $50, bot B rebids $80, bot A rebids $150, bot B rebids $250, and A finally wins at $300 — netting only $200 of the original $500, with B having also burned perhaps $20 in gas on failed transactions that never got included.
Under a private builder auction, both bots submit sealed bids for inclusion directly to the block builder: A bids $260, B bids $220. The builder simply includes whichever bundle pays more, so A wins by paying $260 once, with no repeated failed transactions and no telegraphed rebidding — netting $240, and B loses nothing beyond the trivial cost of the sealed bid itself (rather than gas spent on failed on-chain attempts).
Private order flow reduces wasted gas, but it does not mean searchers keep more of the underlying value — competition among builders and searchers still tends to push bids up toward the full size of the opportunity. It shifts where the competition happens (off-chain, sealed-bid) rather than eliminating the competition itself.
Practice in interviews
Further reading
- Daian et al., Flash Boys 2.0