CLOB-first fill edge analytics — read-only, fair-in-market benchmarks
Data-quality warnings
Headline read
Net edge timeline — 15m buckets
Green/red bars = net edge per 15-minute bucket. Amber dashed markers = config changes inside the window.
Maker / Taker split (options)
Trends — daily sessions from the rolling metrics history
Sharpens as the 35-day metrics history fills; thresholds recalibrate from these distributions (~14 days).
Net edge vs mark (USD/day)
Maker spread capture (0.50 = mid)
Theo-vs-mark bias (median bps)
Perp taker share (%)
Report history
This window — traded volume
Options in underlying BTC (contracts) and premium; perps in USD notional. Maker % is volume-weighted.
Volume inside this window
Cyan = options (BTC contracts), amber = perp hedges (BTC base qty). Hourly reports bucket at 15m; dailies at 1h.
Flow by hour — rolling history
From the 21-day flow history; backfilled from 2026-07-05 onward.
Hour-of-day profile — when does flow concentrate?
Average volume per UTC hour across the rolling history. Sharpens as days accumulate.
Flow by day — last 7 daily reports
Three rulers, none of them an opinion: the ledger (what the balance actually did), the tape (where edge goes after the fill), and the venue experiment (the same engine under two fee regimes).
Ruler 1 — the ledger (this window)
Transaction-log balance decomposition, plus the winners/losers split hiding under the net edge number.
Ruler 2 — where edge goes after the fill
Constant-cohort options markout. At-fill edge vs where the market marked the SAME fills moments later.
Ruler 3 — the venue experiment
Identical quoter, two fee regimes. Last 7 daily sessions per venue from each account's metrics history.
Market-fair analysis is live. Quoter-intended edge will appear after #24 quote/theo logs are wired.
Maker spread capture — where inside the spread do we get filled?
0.50 = filled at mid; below = adverse selection (counterparty picked us off before mid).
Intended edge — quoter theo at fill (data lake)
Filters
DTE x Delta leak matrix
DTE / Delta summary heatmap
Net edge by DTE bucket
Net edge by delta bucket
Net edge by expiry
Worst net option fills
Sub-min perp prints are usually partial fills / escalation-ladder fragments, not parent min-size breaches. Roll up by order_id before alerting.
Delta alignment
Delta band discipline — book delta vs band (data lake, 5s samples)
Perp execution quality — arrival vs fill (data lake)
Options vs Perp — point-in-time bridge
Window transaction-log decomposition
Coverage & freshness
Manifest
Pipeline health
Run warnings
Data lineage
Intended edge is sourced from the quoter's live theo captured at each fill (data lake execution capture, market-maker#24 Wave A/B). Full quote-lifecycle replay (Wave C) remains pending.