PAPER DESK -- PAPER ONLY, simulated maker quotes on the recorded Kalshi tape; not a verdict

Execution (paper)

export 2026-10-22, generated 2026-10-23T19:43:59Z; 2 refusal(s)

Three dates frame this record. Capture history starts 2026-09-24, the first night in the fill-realism table. The first ledger night with paper fills is 2026-10-03. The prospective forward paper series, with schedules written before the games, runs from 2026-10-06 (evidence packet section 5). No calibration verdict yet, for or against the model.

Maker-only quoting

as of: not stated in the receipt

The desk only ever rests one-contract maker bids on each side; it never takes liquidity and sends no orders. Every fill below is a counterfactual credit from replaying the paper quote against the captured trade prints, net of the venue maker fee schedule. These are execution-quality diagnostics, not results.

Quote pacing (hours quoted per arm)

as of: not stated in the receipt

not available: ingame: docs/evidence/forward/live_shadow/2026-10-22.json missing or unreadable

Fill realism (night 2026-10-22)

as of 2026-10-22

TTL s	units	quotes placed	fills credited
30	1	20216	39
30	2	20216	38
30	4	20216	37
30	8	20216	37
5	1	20216	12
5	2	20216	11
5	4	20216	11
5	8	20216	11

by side {"no":27,"yes":12}; by fill basis {"through":36,"touch_queue":3}; quote-life expiries 13121

Markouts (fill realism; probability points x 100 per contract, fill-weighted; a diagnostic)

as of 2026-10-22

horizon s	n	mean
30	7	0.64
120	9	0.94
300	21	-0.69

not available: execution.markouts: docs/evidence/forward/markouts/ holds no receipt

Latency ledger (first mid change minus game event, s)

as of: not stated in the receipt

sport	feed	games	n	p10	p50	p90	p99	book moved first
nba	espn_scoreboard	2	0	-	-	-	-	-

Execution-quality receipt

as of: not stated in the receipt

days 2026-09-27, 2026-09-28, 2026-09-29; estimand S550.v1; census {"served":3,"status_by_sport":{"mlb":{"OK":2},"nfl":{"TAPE_MISSING":1}}}

not verified

mid-shard edits that keep size and both end megabytes (identity limit)
parser behaviour on Python 3.12 (receipt built and tested on 3.10)
one development week only: exploratory, no confirmatory claim
focus depth exists only in the first focus hour of each game
in-game prints arrive by backfill (replay-labelled tape)
join instants overlap; fill shares are not a policy throughput