Improving redistribution efficiency

[sketsdever]

Right now, the builder transfers kickbacks directly to each recipient. This transfer incurs a gas cost, so the user’s effective profit = kickback - cost of transfer. We would like to increase effective profit by reducing the cost for a user to receive their kickback.

One idea is to create a claims contract that stores a mapping of addresses → balances. Every time a user receives a kickback, the builder updates the balance that is mapped to their address. The user can claim() their balance from the contract at any point in time. Because the cost of storage is less than the cost of a transfer, the user can save fees by waiting for several kickbacks before claiming their balance. Further analysis is needed to evaluate precisely how long a user will need to wait for this to be cost effective. It partially depends on the average number of mev-share kickbacks that are processed in a given block.

The claims contract could be avoided by having builders custody funds themselves, though this would introduce additional trust assumptions and possible liabilities for builders.