Scorecard
Week ending Fri 13 Nov 2026Published Fri 18:00 ETCan I trust it?
j k move↵ detailA ask1–4 tabs
92 of 97
Russell Dec adds on model list before 13 Nov
15 sessions
Median lead
5
Adds missed by model
6
Model adds not on prelim list
23 of 23
Provider notices caught
2
Calculation errors found and fixed
41 / 3
Alerts sent / marked wrong
214 / 61
Broker notes read / event facts extracted
61 of 61
Extracted facts with a source quote
2 min
News and notices to screen, median
Broker accuracy by index family
Illustrative · 8-quarter target window| Broker | S&P adds | Lead (sessions) | Russell | Morningstar |
|---|---|---|---|---|
| Broker A | 41/48 · 85% | 6 | 85/97 · 88% | 44/64 · 69% |
| Broker B | 38/48 · 79% | 4 | 81/97 · 84% | No coverage |
| Broker C | 43/48 · 90% | 7 | 77/97 · 79% | No coverage |
| Broker D | 34/48 · 71% | 3 | 70/97 · 72% | No coverage |
| Broker E | 36/48 · 75% | 4 | 74/97 · 76% | 39/64 · 61% |
| Broker F | 39/48 · 81% | 5 | 79/97 · 81% | No coverage |
| Broker G | 42/48 · 88% | 6 | 83/97 · 86% | No coverage |
Real accuracy requires each broker’s dated call matched against provider announcements over the past eight quarters.
Quote coverage
12/12Seed research records quoted. Original links pending.
News speed
2 minSeed median · illustrative. 9 source-timed notices; production ingestion measurement pending.
Pilot gates · pending
Licensing · AI-summarization sign-off · original quote links · broker/provider matching. Track false alerts, usefulness and Blog retyping.
Sources on Russell Dec prelim
STQ model v1.86 false15 lead58/61 del
92/97
Broker A9 false9 lead55/61 del
85/97
Broker B7 false7 lead52/61 del
81/97
Broker C11 false5 lead50/61 del
77/97
Provider notices caught
| FTSE Russell | 14 of 14 | median 4 min |
| S&P DJI | 6 of 6 | median 7 min |
| Morningstar (ex-CRSP) | 3 of 3 | median 11 min |
Model misses · why · fix
STUApril IPO; model treated it as ineligible on seasoning→ Fix eligibility rule for recent IPOs
VWXFloat data 6 weeks stale→ Refresh float vendor file weekly
YZARank-day price moved it across the cutoff→ Show near-cutoff names as a watch list
BCXCorporate action changed share count after model run→ Re-run model after corporate-action feed
CDYShare-class handling error→ Fix share-class mapping
Flow estimate accuracy
Provider share changes vs our estimate (20 confirmed events)Within 5% on 18 of 20
Largest missA tender-offer share assumption, 12% off
Flow estimates revised this week11
Calculation errors found and fixed
ABC close multiple used dollar ADV against share close
Fixed 12 Nov; Test 1 added to release checks
GHI split pre-loaded a day early in days-ADV
Fixed 13 Nov; Test 10 covers it
Earnings: priced vs actual (3 of 8 moved more than priced)
AAQ
BBR
CCS
DDT
EEU
FFV
GGW
HHX
No P&L or cost-savings claims until measured against actual trading records.
Illustrative data, fictional tickers. STQ is read-only and not connected to an OMS/EMS: it doesn't see working orders or fills, and makes no routing, algorithm or price recommendation.