per-trade execution cost (bps), binned, last 24h. No RFQ/CEX venues exist in this data - one AMM-wide distribution, not faceted.
Cost by venue
leaderboard
Venue
Venue Type
Notional
VW Cost (bps)
Median
P90
Fee Component
Price-Impact Component
Spread Component
Not loaded yet
Cost by size bucket
matrix
cell = median AMM price impact (bps) vs pool's own pre-trade price. Cells below 10 trades in the 24h window render as "insufficient data".
Cost vs trade size
scatter
2,000-trade sample, last 24h, colored by venue. x = notional (log USD).
Cost decomposition
NOT BUILT THIS PASSstacked_area
bpsDATA NEEDNeeds a spread_bps term (basis between the pool's pre-trade price and the external reference at that moment) plus a sign reconciliation with price_impact_bps - not built this pass.time (hourly)
feespreadprice impact
Best-venue analysis
NOT BUILT THIS PASSmatrix
DATA NEEDNeeds same-timestamp cross-venue matching (same token, same size, same moment) - not built this pass
cell = % of trades where that venue was cheapest for that size
execution_cost_bps = side_sign × (effective_price − reference_price_at_trade) / reference_price_at_trade × 10000. Positive always = worse than benchmark.
P3 · all headline figures notional-weighted · median and p90 published alongside
AMM: simulate pool state at prior block → price impact + fee; residual = slippage / MEV.
RFQ: counterfactual replay vs deepest CLOB (pre-CEX: deepest AMM) at same timestamp → surfaced as rfq_vs_clob_bps, never blended into book-walked slippage. Reference price is an hourly Yahoo Finance close. Across nights/weekends/holidays when NYSE is closed but on-chain trading isn't, it's anchored to Hyperliquid's real, continuously-traded HIP-3 perpetual price for the ~40 tickers that have one, or log-interpolated between the surrounding real bars otherwise — a deliberate simplification short of real-time/NBBO, sufficient to demonstrate the execution-cost methodology for this MVP. The Ondo Tokenized Equities External Dashboard Spec goes into more detail on reference-price methodology (see also Page 6's Benchmark methodology section).