{"post_stream":{"posts":[{"id":62924,"name":"Dan Guo","username":"DanGuo","avatar_template":"/letter_avatar_proxy/v4/letter/d/48db29/{size}.png","created_at":"2026-10-31T21:25:55.533Z","cooked":"\u003ch2\u003e\u003ca name=\"p-62924-abstract-1\" class=\"anchor\" href=\"#p-62924-abstract-1\" aria-label=\"Heading link\"\u003e\u003c/a\u003eAbstract\u003c/h2\u003e\n\u003cp\u003ePropAMMs are on-chain venues whose prices are set by a single maker. The maker\u003cbr\u003e\npushes quote state on-chain, and fills read that state. We decoded 92,627\u003cbr\u003e\nfills on nine Ethereum PropAMMs over 30 days (10 September to 9 October 2026)\u003cbr\u003e\nfrom execution data. We priced the 69,827 WETH/stablecoin fills against\u003cbr\u003e\nBinance one-second trades.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eThree of the registry venues accept maker state only if it is stamped with\u003cbr\u003e\nthe current block’s timestamp. Their fills always execute against an update\u003cbr\u003e\nin the same block.\u003c/li\u003e\n\u003cli\u003eTakers pay about 0.45 bp at the fill, on every one of the 30 days. Over the\u003cbr\u003e\nfollowing five minutes the price does not move in their favour on average,\u003cbr\u003e\nin any taker class: there is no sign of adverse selection.\u003c/li\u003e\n\u003cli\u003eOf fills after a CEX move of more than 0.5 bp in the preceding 12 seconds,\u003cbr\u003e\n78.5% trade against it. The PropAMM quote follows the CEX in the same\u003cbr\u003e\nblock, while Uniswap lags.\u003cbr\u003e\nArbitrage contracts and routers use the PropAMM as the fresh side of a\u003cbr\u003e\ntrade against a stale AMM.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eFills are priced no later than their block’s sealing, which relays received\u003cbr\u003e\n0.1 to 3.1 s after the slot start on the two days measured. Markouts change\u003cbr\u003e\nby under 0.1 bp across that range.\u003c/p\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-propamms-and-the-venues-measured-2\" class=\"anchor\" href=\"#p-62924-propamms-and-the-venues-measured-2\" aria-label=\"Heading link\"\u003e\u003c/a\u003ePropAMMs and the venues measured\u003c/h2\u003e\n\u003cp\u003e[One paragraph: maker-controlled pricing; a quote-state update is a transaction\u003cbr\u003e\n(or an internal call) placed before fills in the block; builders and\u003cbr\u003e\nBuilderNet’s role. Cite RN-40 and prior public descriptions.]\u003c/p\u003e\n\u003cp\u003eAll nine venues emit the same \u003ccode\u003eSwapped(sender, tokenIn, tokenOut, amountIn, amountOut, recipient)\u003c/code\u003e event. For each, the event’s amounts match token\u003cbr\u003e\ntransfers in a sample of about 16,000 fills.\u003c/p\u003e\n\u003cdiv class=\"md-table\"\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eVenue\u003c/th\u003e\n\u003cth\u003eWETH/stable fills\u003c/th\u003e\n\u003cth\u003eIdentity\u003c/th\u003e\n\u003cth\u003eUpdate source\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eFermiSwap\u003c/td\u003e\n\u003ctd\u003e43,170\u003c/td\u003e\n\u003ctd\u003eEtherscan-verified ABI; listings\u003c/td\u003e\n\u003ctd\u003epricing-engine event\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eEl Zorro\u003c/td\u003e\n\u003ctd\u003e9,316\u003c/td\u003e\n\u003ctd\u003eSourcify exact match\u003c/td\u003e\n\u003ctd\u003e\u003ccode\u003ePrioUpdateRegistry\u003c/code\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eTempest\u003c/td\u003e\n\u003ctd\u003e7,626\u003c/td\u003e\n\u003ctd\u003elistings; implementation unverified\u003c/td\u003e\n\u003ctd\u003e\u003ccode\u003ePrioUpdateRegistry\u003c/code\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMetric\u003c/td\u003e\n\u003ctd\u003e2,811\u003c/td\u003e\n\u003ctd\u003eSourcify \u003ccode\u003eMetricRouterAdapter\u003c/code\u003e\u003c/td\u003e\n\u003ctd\u003eoracle (not decoded)\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ebopAMM (Bebop)\u003c/td\u003e\n\u003ctd\u003e1,772\u003c/td\u003e\n\u003ctd\u003eSourcify exact match\u003c/td\u003e\n\u003ctd\u003e\u003ccode\u003ePrioUpdateRegistry\u003c/code\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePark\u003c/td\u003e\n\u003ctd\u003e1,537\u003c/td\u003e\n\u003ctd\u003eSourcify proxy name only\u003c/td\u003e\n\u003ctd\u003enot decoded\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eGoliath\u003c/td\u003e\n\u003ctd\u003e1,426\u003c/td\u003e\n\u003ctd\u003eSourcify exact match\u003c/td\u003e\n\u003ctd\u003enot decoded\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eKipseli\u003c/td\u003e\n\u003ctd\u003e1,326\u003c/td\u003e\n\u003ctd\u003eSourcify exact match\u003c/td\u003e\n\u003ctd\u003e\u003ccode\u003ePrioUpdateRegistry\u003c/code\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ebopAMM, earlier deployment\u003c/td\u003e\n\u003ctd\u003e843\u003c/td\u003e\n\u003ctd\u003eSourcify exact match\u003c/td\u003e\n\u003ctd\u003enot decoded\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003eThe nine venues account for about 96% of \u003ccode\u003eSwapped\u003c/code\u003e logs in the window. The\u003cbr\u003e\nrest come from contracts we could not identify.\u003c/p\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-data-and-method-3\" class=\"anchor\" href=\"#p-62924-data-and-method-3\" aria-label=\"Heading link\"\u003e\u003c/a\u003eData and method\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eExecution data comes from Xatu (ethPandaOps): blocks, transactions, logs, and\u003cbr\u003e\ntraces where needed.\u003c/li\u003e\n\u003cli\u003eContract identities come from Sourcify.\u003c/li\u003e\n\u003cli\u003ePrices are Binance spot one-second klines: ETH/USDT, with USDC fills\u003cbr\u003e\nconverted through USDC/USDT, which ranged from 0.9999 to 1.0004 USDT on the\u003cbr\u003e\ndays checked. Binance does not archive spot quotes, so the price is the\u003cbr\u003e\nlast trade. At an ETH/USDT tick of 0.01, the difference from the mid is\u003cbr\u003e\nabout 0.04 bp.\u003c/li\u003e\n\u003cli\u003eEvery decoding rule is an exact address and event or selector pair, reviewed\u003cbr\u003e\nagainst verified source or token transfers. There are no heuristics.\u003c/li\u003e\n\u003cli\u003eThe pipeline is \u003ccode\u003etembri\u003c/code\u003e ( \u003ca href=\"https://github.com/TLM-Research/tembri\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\"\u003eGitHub - TLM-Research/tembri: Rust toolbox for reproducible Ethereum execution intelligence — mempool, builder, and reviewed PropAMM update-to-fill measurements. · GitHub\u003c/a\u003e ). Each\u003cbr\u003e\nstep has an ADR stating its evidence.\u003cbr\u003e\n\u003ccode\u003etembri propamm-fetch\u003c/code\u003e and \u003ccode\u003etembri propamm-report\u003c/code\u003e reproduce any day’s\u003cbr\u003e\ntables from public data. Binance files are checked against Binance’s\u003cbr\u003e\nSHA-256, and every input file is pinned by hash.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eA markout is signed from the taker’s side, in bp:\u003cbr\u003e\n\u003ccode\u003es * (P(t) - p_fill) / p_fill\u003c/code\u003e, with \u003ccode\u003es = +1\u003c/code\u003e for a WETH buy. A positive value\u003cbr\u003e\nmeans the taker gained. Horizon 0 is the block timestamp. We report 12 s\u003cbr\u003e\nbefore it, and 12, 60 and 300 s after it. Standard errors are those of the\u003cbr\u003e\n30 daily means.\u003c/p\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-freshness-is-enforced-by-contract-4\" class=\"anchor\" href=\"#p-62924-freshness-is-enforced-by-contract-4\" aria-label=\"Heading link\"\u003e\u003c/a\u003eFreshness is enforced by contract\u003c/h2\u003e\n\u003cp\u003e\u003ccode\u003ePrioUpdateRegistry\u003c/code\u003e stores each maker’s state with a maker-supplied\u003cbr\u003e\ntimestamp. Readers call \u003ccode\u003egetState(lane, minTimestamp, maxTimestamp)\u003c/code\u003e, which\u003cbr\u003e\nreverts outside the window the reader supplies.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTraces show El Zorro, bopAMM and Kipseli reading with\u003cbr\u003e\n\u003ccode\u003eminTimestamp = maxTimestamp = block.timestamp\u003c/code\u003e.\u003c/strong\u003e Tempest reads with\u003cbr\u003e\n\u003ccode\u003e[block.timestamp - 3, block.timestamp + 12]\u003c/code\u003e. Every registry write on\u003cbr\u003e\n2026-10-05 (1,148) carries its own block’s timestamp.\u003c/p\u003e\n\u003cp\u003eFor these venues, a fill can execute only against an update in the same\u003cbr\u003e\nblock. Same-block pairing is therefore a contract rule, not evidence that\u003cbr\u003e\nmakers are fast. FermiSwap’s engine is unverified. All of its 1,934 fills on\u003cbr\u003e\n2026-10-05 had an earlier update in the same block, but we cannot show that\u003cbr\u003e\nthis is enforced.\u003c/p\u003e\n\u003cp\u003eBecause every quote is at most one block old, quote age in maker time is zero\u003cbr\u003e\nby construction. Staleness that matters has to be measured against an outside\u003cbr\u003e\nprice.\u003c/p\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-takers-pay-about-045-bp-and-the-price-does-not-follow-them-5\" class=\"anchor\" href=\"#p-62924-takers-pay-about-045-bp-and-the-price-does-not-follow-them-5\" aria-label=\"Heading link\"\u003e\u003c/a\u003eTakers pay about 0.45 bp, and the price does not follow them\u003c/h2\u003e\n\u003cdiv class=\"md-table\"\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eHorizon\u003c/th\u003e\n\u003cth\u003e−12 s\u003c/th\u003e\n\u003cth\u003e0 s\u003c/th\u003e\n\u003cth\u003e+12 s\u003c/th\u003e\n\u003cth\u003e+60 s\u003c/th\u003e\n\u003cth\u003e+300 s\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eMean markout (SE)\u003c/td\u003e\n\u003ctd\u003e+1.40 (0.16)\u003c/td\u003e\n\u003ctd\u003e−0.45 (0.05)\u003c/td\u003e\n\u003ctd\u003e−0.65 (0.05)\u003c/td\u003e\n\u003ctd\u003e−0.57 (0.08)\u003c/td\u003e\n\u003ctd\u003e−0.74 (0.23)\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003e\u003cstrong\u003eTakers pay at the fill on every day, and the markout does not rise\u003cbr\u003e\nafterwards.\u003c/strong\u003e The horizon-0 daily mean ranges from −0.99 to −0.08 bp. It is\u003cbr\u003e\n−0.37 bp for WETH buys and −0.53 bp for sells.\u003c/p\u003e\n\u003cp\u003eBy venue, the horizon-0 mean is between −0.49 and −0.80 bp at the venues\u003cbr\u003e\nwith fills on 20 or more days (FermiSwap −0.51, El Zorro −0.79, Tempest\u003cbr\u003e\n−0.49, bopAMM −0.80, Kipseli −0.72). Two venues differ, on less data. Metric’s\u003cbr\u003e\nWETH/USDC fills fall on ten days in September, and takers paid nothing at\u003cbr\u003e\nthe fill there (0.00 bp, SE 0.03), then lost 0.19 bp by 12 s. Goliath, live\u003cbr\u003e\nfor the last four days, is the only venue with a positive horizon-0 mean\u003cbr\u003e\n(+0.40 bp), negative by 12 s.\u003c/p\u003e\n\u003cp\u003eHorizon 0 is the slot start, but a fill is priced when its block is built.\u003cbr\u003e\nWe asked five relays’ data APIs when they first received each winning block\u003cbr\u003e\nwith a fill, on 2026-10-05 and 2026-10-09. They had received 98% of these\u003cbr\u003e\nblocks, a median 1.0 s after the slot start on both days (10th to 90th\u003cbr\u003e\npercentile 0.9 to 2.6 s; range 0.1 to 3.1 s). The times fall in two groups:\u003cbr\u003e\n70% of blocks arrive within 1.5 s and 22 to 23% after 2 s. Priced at the\u003cbr\u003e\nfirst receipt instead, the mean markout is −0.68 bp against −0.64 bp at the\u003cbr\u003e\nslot start on 2026-10-05, and −0.63 against −0.56 bp on 2026-10-09. ETH/USDT’s return standard deviation was 0.58 bp over\u003cbr\u003e\n1 s, so an offset of a few seconds adds about 1 bp of noise per fill, which\u003cbr\u003e\nwidens the distribution without biasing the mean.\u003c/p\u003e\n\u003cp\u003eThe cost does not grow with size or volatility. It is 0.50 bp under $1k,\u003cbr\u003e\n0.38 bp at $1k to $10k and 0.48 bp at $10k to $100k, and 0.50, 0.44 and\u003cbr\u003e\n0.40 bp across volatility terciles. On Uniswap V3’s 0.05% pools for the same\u003cbr\u003e\npairs, the cost against Binance, fee included, is 5.3 bp under $1k and 1.6 bp\u003cbr\u003e\nat $1k to $10k.\u003c/p\u003e\n\u003cp\u003eCorrection (2 November 2026). An earlier version said FermiSwap re-quotes inside a block. It does not. Each FermiSwap update names one of its eight markets, and the 21% of its update blocks with more than one update are updates to different markets. Over the 30 days, no market on any venue was updated twice in one block, which matches the one update per block that Neuder and Bahrani describe (t/26129). 14% of fills are the second or later fill on the same venue and pair in their block. They trade against the same update: at the first fill’s price at FermiSwap and Tempest, and at a price that depends on size at El Zorro, whose contract adds an impact of 1 bp × fill / vault balance.\u003c/p\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-an-on-chain-reference-gives-the-wrong-sign-6\" class=\"anchor\" href=\"#p-62924-an-on-chain-reference-gives-the-wrong-sign-6\" aria-label=\"Heading link\"\u003e\u003c/a\u003eAn on-chain reference gives the wrong sign\u003c/h2\u003e\n\u003cp\u003eAgainst the Uniswap V3 USDC/WETH 0.05% pool’s last price before the fill, the\u003cbr\u003e\nmean horizon-0 markout is +2.28 bp, and both buyers and sellers appear to\u003cbr\u003e\ngain. A true mid cannot give that result.\u003c/p\u003e\n\u003cp\u003eThe pool’s price sits anywhere in its ±5 bp fee band: against Binance, its\u003cbr\u003e\n10th to 90th percentile on 2026-10-05 is −5.5 to +4.8 bp. In addition,\u003cbr\u003e\nrouters send a trade to a PropAMM only when it beats the pool. Markouts\u003cbr\u003e\nagainst a pool that competes for the same flow are biased towards the taker.\u003c/p\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-takers-trade-against-the-move-the-amm-has-not-yet-followed-7\" class=\"anchor\" href=\"#p-62924-takers-trade-against-the-move-the-amm-has-not-yet-followed-7\" aria-label=\"Heading link\"\u003e\u003c/a\u003eTakers trade against the move the AMM has not yet followed\u003c/h2\u003e\n\u003cp\u003eOver the 12 s before the block, the taker-signed CEX move averages −1.84 bp\u003cbr\u003e\n(SE 0.14). Of fills that followed a move of more than 0.5 bp, 78.5% traded\u003cbr\u003e\nagainst it: buying after a fall, selling after a rise. No day was below 68%.\u003c/p\u003e\n\u003cp\u003eGrouped by that move, the Uniswap price’s gap from Binance at the fill grows\u003cbr\u003e\nwith it, while the PropAMM’s markout does not change:\u003c/p\u003e\n\u003cdiv class=\"md-table\"\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003e12 s move\u003c/th\u003e\n\u003cth\u003e\u0026lt; −4 bp\u003c/th\u003e\n\u003cth\u003e−4 to −1\u003c/th\u003e\n\u003cth\u003e−1 to +1\u003c/th\u003e\n\u003cth\u003e\u0026gt; +1\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eFills\u003c/td\u003e\n\u003ctd\u003e13,243\u003c/td\u003e\n\u003ctd\u003e21,577\u003c/td\u003e\n\u003ctd\u003e26,882\u003c/td\u003e\n\u003ctd\u003e8,125\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eUniswap worse for the taker than Binance\u003c/td\u003e\n\u003ctd\u003e4.42\u003c/td\u003e\n\u003ctd\u003e3.60\u003c/td\u003e\n\u003ctd\u003e2.12\u003c/td\u003e\n\u003ctd\u003e0.09\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePropAMM markout against Binance\u003c/td\u003e\n\u003ctd\u003e−0.41\u003c/td\u003e\n\u003ctd\u003e−0.41\u003c/td\u003e\n\u003ctd\u003e−0.48\u003c/td\u003e\n\u003ctd\u003e−0.51\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003eWhen ETH moves on the CEX, the PropAMM quote moves in the same block and the\u003cbr\u003e\nAMM does not. The PropAMM is then the better price in the direction opposite\u003cbr\u003e\nthe move.\u003c/p\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-who-the-takers-are-8\" class=\"anchor\" href=\"#p-62924-who-the-takers-are-8\" aria-label=\"Heading link\"\u003e\u003c/a\u003eWho the takers are\u003c/h2\u003e\n\u003cp\u003eClasses come from transaction structure plus a reviewed address table\u003cbr\u003e\n(Sourcify-named contracts only). Unverified contracts are classed by structure,\u003cbr\u003e\nnever by guess.\u003c/p\u003e\n\u003cdiv class=\"md-table\"\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eClass\u003c/th\u003e\n\u003cth\u003eFills\u003c/th\u003e\n\u003cth\u003e12 s move before\u003c/th\u003e\n\u003cth\u003eAgainst the move\u003c/th\u003e\n\u003cth\u003e0 s\u003c/th\u003e\n\u003cth\u003e60 s\u003c/th\u003e\n\u003cth\u003e300 s\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eUnverified contract called directly\u003c/td\u003e\n\u003ctd\u003e37,098\u003c/td\u003e\n\u003ctd\u003e−2.67\u003c/td\u003e\n\u003ctd\u003e86%\u003c/td\u003e\n\u003ctd\u003e−0.33\u003c/td\u003e\n\u003ctd\u003e−0.45\u003c/td\u003e\n\u003ctd\u003e−0.59\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eAggregators (Tycho, KyberSwap, Relay, the PropAMM router, …)\u003c/td\u003e\n\u003ctd\u003e17,241\u003c/td\u003e\n\u003ctd\u003e−1.02\u003c/td\u003e\n\u003ctd\u003e69%\u003c/td\u003e\n\u003ctd\u003e−0.43\u003c/td\u003e\n\u003ctd\u003e−0.51\u003c/td\u003e\n\u003ctd\u003e−0.66\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCoW settlement\u003c/td\u003e\n\u003ctd\u003e5,301\u003c/td\u003e\n\u003ctd\u003e−0.18\u003c/td\u003e\n\u003ctd\u003e52%\u003c/td\u003e\n\u003ctd\u003e−0.78\u003c/td\u003e\n\u003ctd\u003e−0.83\u003c/td\u003e\n\u003ctd\u003e−0.86\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eUnverified contract reached another way\u003c/td\u003e\n\u003ctd\u003e4,914\u003c/td\u003e\n\u003ctd\u003e−1.51\u003c/td\u003e\n\u003ctd\u003e72%\u003c/td\u003e\n\u003ctd\u003e−0.59\u003c/td\u003e\n\u003ctd\u003e−0.55\u003c/td\u003e\n\u003ctd\u003e−1.13\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eERC-4337 EntryPoint\u003c/td\u003e\n\u003ctd\u003e2,891\u003c/td\u003e\n\u003ctd\u003e−0.37\u003c/td\u003e\n\u003ctd\u003e77%\u003c/td\u003e\n\u003ctd\u003e−0.95\u003c/td\u003e\n\u003ctd\u003e−1.25\u003c/td\u003e\n\u003ctd\u003e−0.37\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eEIP-7702 delegated account\u003c/td\u003e\n\u003ctd\u003e2,242\u003c/td\u003e\n\u003ctd\u003e−1.96\u003c/td\u003e\n\u003ctd\u003e89%\u003c/td\u003e\n\u003ctd\u003e−0.44\u003c/td\u003e\n\u003ctd\u003e−0.64\u003c/td\u003e\n\u003ctd\u003e−0.78\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003e\u003cstrong\u003eNo class has a positive mean markout at any horizon.\u003c/strong\u003e Of 50 senders with\u003cbr\u003e\nat least 100 fills, one has a 60 s mean more than two standard errors above\u003cbr\u003e\nzero (+1.45 bp), and it is positive on only 13 of its 29 days. With 50\u003cbr\u003e\nsenders, about one such result is expected by chance.\u003c/p\u003e\n\u003cp\u003eThe two classes that react most to the CEX move are unverified contracts\u003cbr\u003e\ncalled directly and delegated accounts. In the week of 3 to 9 October, their\u003cbr\u003e\ntransactions look like atomic arbitrage:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e99% of the direct-contract transactions trade the contract’s own inventory\u003cbr\u003e\n(the signer moves no tokens).\u003c/li\u003e\n\u003cli\u003e74% end with no token lost and a gain (87% for delegated accounts).\u003c/li\u003e\n\u003cli\u003e89% (97%) also swap on a Uniswap, Curve or Balancer pool.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eTheir profit comes from the lagging AMM. The PropAMM is the fresh side of the\u003cbr\u003e\ntrade, and it earns its spread on it.\u003c/p\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-fills-arrive-privately-and-the-propamm-leg-is-not-sandwiched-9\" class=\"anchor\" href=\"#p-62924-fills-arrive-privately-and-the-propamm-leg-is-not-sandwiched-9\" aria-label=\"Heading link\"\u003e\u003c/a\u003eFills arrive privately, and the PropAMM leg is not sandwiched\u003c/h2\u003e\n\u003cp\u003eWe joined every PropAMM transaction to the Flashbots Mempool Dumpster\u003cbr\u003e\narchive, which saw 47% of all transactions included on 2026-10-05.\u003cbr\u003e\n\u003cstrong\u003eIt saw 2.9% of PropAMM fill transactions.\u003c/strong\u003e Almost all of those are\u003cbr\u003e\naggregator flow: 10% of aggregator fills were public before their block,\u003cbr\u003e\nagainst under 0.1% for arbitrage contracts, delegated accounts, CoW and\u003cbr\u003e\nERC-4337.\u003c/p\u003e\n\u003cp\u003eNo fill was sandwiched on its PropAMM pool in 30 days. The AMM legs of public\u003cbr\u003e\nrouted transactions are another matter. Of 2,195 public fill transactions,\u003cbr\u003e\n95 (4.3%) were bracketed on a Uniswap V2 or V3 pool that they also traded,\u003cbr\u003e\nmostly by transactions directly adjacent to them. Of 76,349 unseen ones, 2\u003cbr\u003e\nwere.\u003c/p\u003e\n\u003cp\u003eAcross all Uniswap V2 and V3 swaps in the week of 3 to 9 October, public\u003cbr\u003e\nswaps were sandwiched more often the larger they were: 2.5% under $100, 8%\u003cbr\u003e\nat $1k to $10k and 27% at $10k to $100k, against at most 0.14% for private\u003cbr\u003e\nswaps. The public share of swaps falls with size, from 12 to 15% below $300\u003cbr\u003e\nto about 2% between $3k and $30k. Takers of mid-sized trades mostly stay out\u003cbr\u003e\nof the public mempool, and the ones that do not are often sandwiched.\u003c/p\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-limits-10\" class=\"anchor\" href=\"#p-62924-limits-10\" aria-label=\"Heading link\"\u003e\u003c/a\u003eLimits\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eThe pricing time is bounded by sealing, measured on two days at five\u003cbr\u003e\nrelays. Four other relays did not answer. Maker signing times are not\u003cbr\u003e\nobservable: updates carry only the slot timestamp.\u003c/li\u003e\n\u003cli\u003eThe Binance price is the last trade, and Binance is one venue.\u003c/li\u003e\n\u003cli\u003eFermiSwap, the largest venue, is decoded from BuilderNet’s description of\u003cbr\u003e\nits engine, not from verified source. Tempest’s and Park’s implementations\u003cbr\u003e\nare unverified too.\u003c/li\u003e\n\u003cli\u003eAbout 4% of \u003ccode\u003eSwapped\u003c/code\u003e logs come from contracts we could not identify.\u003c/li\u003e\n\u003cli\u003eThe atomic-arbitrage test covers one week, ignores native ETH and payments\u003cbr\u003e\nto the builder, and is not yet part of the pipeline.\u003c/li\u003e\n\u003cli\u003eBuilder names come from block extra-data, not a verified relay join.\u003c/li\u003e\n\u003cli\u003e30 days, with a largest daily ETH move of 6.7%. A sharper market may\u003cbr\u003e\nbehave differently.\u003c/li\u003e\n\u003cli\u003e“Not seen” means not seen by one public archive. The AMM-leg sandwich\u003cbr\u003e\ntest covers only Uniswap V2 and V3 and is not yet part of the pipeline.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-next-11\" class=\"anchor\" href=\"#p-62924-next-11\" aria-label=\"Heading link\"\u003e\u003c/a\u003eNext\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eMetric fills execute against oracle state a median 132 s old, with no\u003cbr\u003e\nadverse selection visible; its pools are unverified, so why is open.\u003c/li\u003e\n\u003cli\u003eHow much of the lagging AMMs’ loss to arbitrage now passes through PropAMM\u003cbr\u003e\nspreads.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003e\u003ca name=\"p-62924-acknowledgements-and-data-12\" class=\"anchor\" href=\"#p-62924-acknowledgements-and-data-12\" aria-label=\"Heading link\"\u003e\u003c/a\u003eAcknowledgements and data\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eExecution data: Xatu by ethPandaOps,\u003cbr\u003e\n\u003ca href=\"https://ethpandaops.io/data/xatu/\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\"\u003eXatu Data | ethPandaOps\u003c/a\u003e ( \u003ca href=\"https://github.com/ethpandaops/xatu\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\"\u003eGitHub - ethpandaops/xatu: Ethereum network monitoring with collection clients and a centralized server for data pipelining. · GitHub\u003c/a\u003e ).\u003c/li\u003e\n\u003cli\u003eContract identities: Sourcify, \u003ca href=\"https://sourcify.dev\" rel=\"noopener nofollow ugc\"\u003ehttps://sourcify.dev\u003c/a\u003e.\u003c/li\u003e\n\u003cli\u003eCEX prices: Binance public data, \u003ca href=\"https://data.binance.vision\" rel=\"noopener nofollow ugc\"\u003ehttps://data.binance.vision\u003c/a\u003e\u003cbr\u003e\n( \u003ca href=\"https://github.com/binance/binance-public-data\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\"\u003eGitHub - binance/binance-public-data: Details on how to get Binance public data · GitHub\u003c/a\u003e ).\u003c/li\u003e\n\u003cli\u003ePublic-mempool sightings: Flashbots Mempool Dumpster,\u003cbr\u003e\n\u003ca href=\"https://mempool-dumpster.flashbots.net\" rel=\"noopener nofollow ugc\"\u003ehttps://mempool-dumpster.flashbots.net\u003c/a\u003e\u003cbr\u003e\n( \u003ca href=\"https://github.com/flashbots/mempool-dumpster\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\"\u003eGitHub - flashbots/mempool-dumpster: Dump all the mempool transactions 🗑️ ♻️ (in Parquet + CSV) · GitHub\u003c/a\u003e ).\u003c/li\u003e\n\u003cli\u003eSealing times: the data APIs of the Ultrasound, Titan, Aestus, Flashbots\u003cbr\u003e\nand bloXroute (regulated) relays,\u003cbr\u003e\n\u003ca href=\"https://github.com/flashbots/relay-specs\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\"\u003eGitHub - flashbots/relay-specs: MEV-Boost Relay API Specs · GitHub\u003c/a\u003e .\u003c/li\u003e\n\u003cli\u003ePropAMM registry and oracle addresses: BuilderNet API reference,\u003cbr\u003e\n\u003ca href=\"https://buildernet.org/docs/api\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\"\u003eAPI Reference | BuilderNet\u003c/a\u003e .\u003c/li\u003e\n\u003cli\u003eVenue entrypoints: the Lambda Class PropAMM router deployment table,\u003cbr\u003e\n\u003ca href=\"https://github.com/lambdaclass/propamm-router-contracts\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\"\u003eGitHub - lambdaclass/propamm-router-contracts: Single-hop router that quotes and executes swaps against a proprietary AMM (FermiSwap, Kipseli, or Bebop) or directly against Uniswap V3, and falls back to Uniswap V3 when the chosen proprietary venue cannot fill the swap. · GitHub\u003c/a\u003e .\u003c/li\u003e\n\u003cli\u003eTitan Builder’s PropAMM documentation,\u003cbr\u003e\n\u003ca href=\"https://docs.titanbuilder.xyz/propamms\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\"\u003ePropAMMs | Titan Builder Docs\u003c/a\u003e .\u003c/li\u003e\n\u003c/ul\u003e","post_number":1,"post_type":1,"posts_count":3,"updated_at":"2026-11-02T21:49:43.205Z","reply_count":0,"reply_to_post_number":null,"quote_count":0,"incoming_link_count":58,"reads":54,"readers_count":53,"score":275.4,"yours":false,"topic_id":26201,"topic_slug":"propamm-fills-on-ethereum-mainnet-and-propamm-support-for-tembri-toolbox-same-block-quotes-a-small-spread-and-flow-taken-from-lagging-amms","display_username":"Dan Guo","primary_group_name":null,"flair_name":null,"flair_url":null,"flair_bg_color":null,"flair_color":null,"flair_group_id":null,"badges_granted":[],"version":3,"can_edit":false,"can_delete":false,"can_recover":false,"can_see_hidden_post":false,"can_wiki":false,"link_counts":[{"url":"https://github.com/TLM-Research/tembri","internal":false,"reflection":false,"title":"GitHub - TLM-Research/tembri: Rust toolbox for reproducible Ethereum execution intelligence — mempool, builder, and reviewed PropAMM update-to-fill measurements. · GitHub","clicks":1},{"url":"https://sourcify.dev","internal":false,"reflection":false,"title":"sourcify.eth","clicks":0},{"url":"https://data.binance.vision","internal":false,"reflection":false,"title":"Binance Data Collection","clicks":0},{"url":"https://github.com/binance/binance-public-data","internal":false,"reflection":false,"title":"GitHub - binance/binance-public-data: Details on how to get Binance public data · GitHub","clicks":0},{"url":"https://mempool-dumpster.flashbots.net","internal":false,"reflection":false,"title":"Mempool Dumpster ♻️","clicks":0},{"url":"https://github.com/flashbots/mempool-dumpster","internal":false,"reflection":false,"title":"GitHub - flashbots/mempool-dumpster: Dump all the mempool transactions 🗑️ ♻️ (in Parquet + CSV) · GitHub","clicks":0},{"url":"https://github.com/flashbots/relay-specs","internal":false,"reflection":false,"title":"GitHub - flashbots/relay-specs: MEV-Boost Relay API Specs · GitHub","clicks":0},{"url":"https://buildernet.org/docs/api","internal":false,"reflection":false,"title":"API Reference | BuilderNet","clicks":0},{"url":"https://github.com/lambdaclass/propamm-router-contracts","internal":false,"reflection":false,"title":"GitHub - lambdaclass/propamm-router-contracts: Single-hop router that quotes and executes swaps against a proprietary AMM (FermiSwap, Kipseli, or Bebop) or directly against Uniswap V3, and falls back to Uniswap V3 when the chosen proprietary venue cannot ","clicks":0},{"url":"https://docs.titanbuilder.xyz/propamms","internal":false,"reflection":false,"title":"PropAMMs | Titan Builder Docs","clicks":0},{"url":"https://github.com/ethpandaops/xatu","internal":false,"reflection":false,"title":"GitHub - ethpandaops/xatu: Ethereum network monitoring with collection clients and a centralized server for data pipelining. · GitHub","clicks":0},{"url":"https://ethpandaops.io/data/xatu/","internal":false,"reflection":false,"title":"Xatu Data | ethPandaOps","clicks":0},{"url":"https://ethresear.ch/t/ethereum-propamms-today-and-tomorrow/26129/3","internal":true,"reflection":true,"title":"Ethereum PropAMMs: today and tomorrow","clicks":0}],"read":true,"user_title":null,"bookmarked":false,"actions_summary":[{"id":2,"count":1}],"moderator":false,"admin":false,"staff":false,"user_id":15523,"hidden":false,"trust_level":1,"deleted_at":null,"user_deleted":false,"edit_reason":null,"can_view_edit_history":true,"wiki":false,"post_url":"/t/propamm-fills-on-ethereum-mainnet-and-propamm-support-for-tembri-toolbox-same-block-quotes-a-small-spread-and-flow-taken-from-lagging-amms/26201/1","reactions":[{"id":"heart","type":"emoji","count":1}],"current_user_reaction":null,"reaction_users_count":1,"current_user_used_main_reaction":false,"can_accept_answer":false,"can_unaccept_answer":false,"accepted_answer":false,"topic_accepted_answer":null,"can_vote":false},{"id":62949,"name":"Dan Guo","username":"DanGuo","avatar_template":"/letter_avatar_proxy/v4/letter/d/48db29/{size}.png","created_at":"2026-11-02T21:46:28.555Z","cooked":"\u003cp\u003e▎ Correction (2 November 2026). My post earlier above said \" FermiSwap often re-quotes inside a block. \" It does not. Each FermiSwap update names one of its eight markets, and the 21% of its update blocks with more than one update are updates to different markets. Over the 30 days, no market on any venue was updated twice in one block, which matches the one update per block that Neuder and Bahrani describe (t/26129). 14% of fills are the second or later fill on the same venue and pair in their block. They trade against the same update: at the first fill’s price at FermiSwap and Tempest, and at a price that depends on size at El Zorro, whose contract adds an impact of 1 bp × fill / vault balance.\u003c/p\u003e","post_number":2,"post_type":1,"posts_count":3,"updated_at":"2026-11-02T21:46:28.555Z","reply_count":1,"reply_to_post_number":null,"quote_count":0,"incoming_link_count":7,"reads":49,"readers_count":48,"score":49.6,"yours":false,"topic_id":26201,"topic_slug":"propamm-fills-on-ethereum-mainnet-and-propamm-support-for-tembri-toolbox-same-block-quotes-a-small-spread-and-flow-taken-from-lagging-amms","display_username":"Dan Guo","primary_group_name":null,"flair_name":null,"flair_url":null,"flair_bg_color":null,"flair_color":null,"flair_group_id":null,"badges_granted":[],"version":1,"can_edit":false,"can_delete":false,"can_recover":false,"can_see_hidden_post":false,"can_wiki":false,"read":true,"user_title":null,"bookmarked":false,"actions_summary":[],"moderator":false,"admin":false,"staff":false,"user_id":15523,"hidden":false,"trust_level":1,"deleted_at":null,"user_deleted":false,"edit_reason":null,"can_view_edit_history":true,"wiki":false,"post_url":"/t/propamm-fills-on-ethereum-mainnet-and-propamm-support-for-tembri-toolbox-same-block-quotes-a-small-spread-and-flow-taken-from-lagging-amms/26201/2","reactions":[],"current_user_reaction":null,"reaction_users_count":0,"current_user_used_main_reaction":false,"can_accept_answer":false,"can_unaccept_answer":false,"accepted_answer":false,"topic_accepted_answer":null},{"id":63015,"name":"Dan Guo","username":"DanGuo","avatar_template":"/letter_avatar_proxy/v4/letter/d/48db29/{size}.png","created_at":"2026-11-14T23:24:26.070Z","cooked":"\u003cp\u003e\u003cstrong\u003eCorrection: 1 July 2025 figures\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTwo column-order bugs in TEMBRI affected only 1 July 2025, the one day in an older Xatu block layout. The ba\u003cbr\u003e\nse fee was read from the wrong column (fixed in 598bf28), and so was the builder label (fixed in 96e616f), w\u003cbr\u003e\nhich left 177 locally built blocks unlabelled. The 2026 figures do not change.\u003c/p\u003e\n\u003cdiv class=\"md-table\"\u003e\n\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eFigure, 1 July 2025\u003c/th\u003e\n\u003cth style=\"text-align:right\"\u003eWas\u003c/th\u003e\n\u003cth style=\"text-align:right\"\u003eNow\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eSwap fee, first 10% / 30% to 60% of gas (gwei)\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e15.63 / 1.64\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e15.42 / 1.40\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eRatio, first 10% / 30% to 60%\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e9.5\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e11.0\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eSwap zero tip, all / first 5%\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e4.5% / 4.3%\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e5.3% / 7.2%\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMulti-pool swap zero tip, all / first 5%\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e5.1% / 10.4%\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e9.6% / 28.0%\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eZero-fee swaps at the front (ratio)\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e3.0\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e4.2\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eTitan / Quasar / BuilderNet correlation\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e-0.35 / -0.41 / -0.47\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e-0.37 / -0.44 / -0.49\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eLocally built\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e3.0%, -0.63\u003c/td\u003e\n\u003ctd style=\"text-align:right\"\u003e5.5%, -0.71\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003eThe conclusions hold: the front-to-middle fee ratio did not fall from 2025 to 2026, and locally built blocks have the most negative correlation. One row reverses: multi-pool zero-tip swaps at the front were more common on 1 July 2025 (28.0%) than on 1 July 2026 (21.0%).\u003c/p\u003e\n\u003cp\u003eThanks to akatsukiwork0320 for finding both bugs.\u003c/p\u003e","post_number":3,"post_type":1,"posts_count":3,"updated_at":"2026-11-14T23:24:26.070Z","reply_count":0,"reply_to_post_number":2,"quote_count":0,"incoming_link_count":1,"reads":5,"readers_count":4,"score":0.6,"yours":false,"topic_id":26201,"topic_slug":"propamm-fills-on-ethereum-mainnet-and-propamm-support-for-tembri-toolbox-same-block-quotes-a-small-spread-and-flow-taken-from-lagging-amms","display_username":"Dan Guo","primary_group_name":null,"flair_name":null,"flair_url":null,"flair_bg_color":null,"flair_color":null,"flair_group_id":null,"badges_granted":[],"version":1,"can_edit":false,"can_delete":false,"can_recover":false,"can_see_hidden_post":false,"can_wiki":false,"read":true,"user_title":null,"reply_to_user":{"id":15523,"username":"DanGuo","name":"Dan Guo","avatar_template":"/letter_avatar_proxy/v4/letter/d/48db29/{size}.png"},"bookmarked":false,"actions_summary":[],"moderator":false,"admin":false,"staff":false,"user_id":15523,"hidden":false,"trust_level":1,"deleted_at":null,"user_deleted":false,"edit_reason":null,"can_view_edit_history":true,"wiki":false,"post_url":"/t/propamm-fills-on-ethereum-mainnet-and-propamm-support-for-tembri-toolbox-same-block-quotes-a-small-spread-and-flow-taken-from-lagging-amms/26201/3","reactions":[],"current_user_reaction":null,"reaction_users_count":0,"current_user_used_main_reaction":false,"can_accept_answer":false,"can_unaccept_answer":false,"accepted_answer":false,"topic_accepted_answer":null}],"stream":[62924,62949,63015]},"timeline_lookup":[[1,21],[2,19],[3,7]],"suggested_topics":[],"tags":[{"id":90,"name":"mev","slug":"mev"},{"id":93,"name":"proposer-builder-separation","slug":"proposer-builder-separation"},{"id":86,"name":"eip-1559","slug":"eip-1559"}],"tags_descriptions":{},"fancy_title":"PropAMM fills on Ethereum mainnet and PropAMM support for TEMBRI (toolbox): same-block quotes, a small spread, and flow taken from lagging AMMs","id":26201,"title":"PropAMM fills on Ethereum mainnet and PropAMM support for TEMBRI (toolbox): same-block quotes, a small spread, and flow taken from lagging AMMs","posts_count":3,"created_at":"2026-10-31T21:25:55.357Z","views":253,"reply_count":1,"like_count":1,"last_posted_at":"2026-11-14T23:24:26.070Z","visible":true,"closed":false,"archived":false,"has_summary":false,"archetype":"regular","slug":"propamm-fills-on-ethereum-mainnet-and-propamm-support-for-tembri-toolbox-same-block-quotes-a-small-spread-and-flow-taken-from-lagging-amms","category_id":37,"word_count":2771,"deleted_at":null,"user_id":15523,"featured_link":null,"pinned_globally":false,"pinned_at":null,"pinned_until":null,"image_url":null,"slow_mode_seconds":0,"draft":null,"draft_key":"topic_26201","draft_sequence":null,"unpinned":null,"pinned":false,"current_post_number":1,"highest_post_number":3,"deleted_by":null,"actions_summary":[{"id":4,"count":0,"hidden":false,"can_act":false},{"id":8,"count":0,"hidden":false,"can_act":false},{"id":10,"count":0,"hidden":false,"can_act":false},{"id":7,"count":0,"hidden":false,"can_act":false}],"chunk_size":20,"bookmarked":false,"topic_timer":null,"message_bus_last_id":0,"participant_count":1,"show_read_indicator":false,"thumbnails":null,"slow_mode_enabled_until":null,"valid_reactions":["heart","+1","laughing","open_mouth","clap","confetti_ball","hugs"],"has_accepted_answer":false,"can_create_shared_issue":false,"shared_issue_visible":false,"can_vote":false,"vote_count":0,"user_voted":false,"discourse_zendesk_plugin_zendesk_id":null,"discourse_zendesk_plugin_zendesk_url":"https://your-url.zendesk.com/agent/tickets/","details":{"can_edit":false,"notification_level":1,"participants":[{"id":15523,"username":"DanGuo","name":"Dan Guo","avatar_template":"/letter_avatar_proxy/v4/letter/d/48db29/{size}.png","post_count":3,"primary_group_name":null,"flair_name":null,"flair_url":null,"flair_color":null,"flair_bg_color":null,"flair_group_id":null,"trust_level":1}],"created_by":{"id":15523,"username":"DanGuo","name":"Dan Guo","avatar_template":"/letter_avatar_proxy/v4/letter/d/48db29/{size}.png"},"last_poster":{"id":15523,"username":"DanGuo","name":"Dan Guo","avatar_template":"/letter_avatar_proxy/v4/letter/d/48db29/{size}.png"},"links":[{"url":"https://github.com/TLM-Research/tembri","title":"GitHub - TLM-Research/tembri: Rust toolbox for reproducible Ethereum execution intelligence — mempool, builder, and reviewed PropAMM update-to-fill measurements. · GitHub","internal":false,"attachment":false,"reflection":false,"clicks":1,"user_id":15523,"domain":"github.com","root_domain":"github.com"}]},"bookmarks":[]}