I fully agree the idea that ensuring solutions outperform the EBBO price will enhance user experience. Assuming we can currently define what EBBO should be—which may change in the future—and generate solutions within the auction timeframe, I still believe it’s unreasonable to expect solvers to self-censor their proposals if these solutions can compete effectively.
Solvers’ primary role is to present all possible solutions to match orders in the auction. This includes utilizing external liquidity and CoW orders. Given the significant risk of penalties for reverting solutions (42.86% result in a loss of 0.01 ETH), compared to the relatively small profits (67.97% earn less than 0.001 ETH, and 37.09% earn less than 0.0001 ETH), solvers should be cautious in employing public AMMs to ensure executability of their solutions.
After proposing a solution, it is not the solver’s role to determine the winning batch. If EBBO is prioritized, it should be considered before the current surplus-based scoring system. Thus, the solution aligning with EBBO should win the batch. Not sure if this is the idea of EBBO solver, since it seems that it is just to provide a ground-truth for a monitoring system. However, requiring solvers to meet the EBBO at all costs could be challenging and potentially conflict with their interests under the existing scoring rules if other competitive solutions are available. And IMHO, the fairness of penalizing solver with a solution that is not executed on-chain is questionable, as discussed in Move Solver Rewards Mechanism from Theory to Practice - #2 by AndreaC .