[ethresear.ch] [Execution Layer Research] Ethereum PropAMMs: today and tomorrow

Thank for this great post. I learned a lot. Your description matches what we see in 30 days of mainnet PropAMM fills (nine venues; data in PropAMM fills on Ethereum mainnet and PropAMM support for TEMBRI (toolbox): same-block quotes, a small spread, and flow taken from lagging AMMs ): - On a day with traces, every fill at El Zorro, bopAMM, Kipseli and FermiSwap had its maker’s update earlier in the same block. No market was updated twice in one block. - 14% of fills are a later take against the same update. The fade differs by venue: none at FermiSwap and Tempest, by size at El Zorro, about 0.3 bp per earlier fill at Kipseli.

i have a few questions for your perspectives: 1. With 12 s slots against about 400 ms on Solana, do you see a need in the future for more than one update per maker per block on Ethereum? Inside a block every take has the same timestamp, so a second update could respond only to the earlier fills, which is a last look at the later takers. Is the fade the intended substitute?
2. Takes are ordered by fee. In Titan blocks over 30 days, transactions with PropAMM fills paid $28.8k in priority fees to the builder, while takers paid makers $20k in spread against the Binance mid (fees are for the whole transaction). So takers pay the builder more than they pay makers. Is that the split you expected?
3. Builders without maker deals built 14% of blocks but carried almost no PropAMM fills (4 of about 87,000 at venues that need a maker update from my measurement). Is access to PropAMM flow becoming a reason some builders win more blocks?