Build status

Built today, at launch, Year 2.

What the current build does, what the production engine does at commercial launch, and what is planned for Year 2.

CapabilityDemo build today (September 2026)At commercial launch (Q1 2027)Year 2Evidence
Coverage universe482 stocks (503 current S&P 500 constituents less 21) + 28 ETFs; no capitalization or volume screenapproximately 2,200 screened U.S. equities plus liquid ETFsLiquid ADRs addedMethodology · Universe
Index membershipCurrent constituents; survivorship bias disclosed on every pagePoint-in-time membership; delisted and demoted names retainedDemo-build note on every PDF page
Share countsCurrent shares outstanding applied to historical pricesAs-reported share counts on each dateValidation record
FundamentalsNone; library uses price and volume onlyNonePoint-in-time fundamentals licensed; Value and Quality familiesFactor table
Corporate actionsSplit and dividend adjustment embedded in the adjusted close; no event tableSplits, dividends, mergers, delistings, symbol changesData table
Macro dataVIX and VIX3M; Treasury, credit and gold ETFs as proxiesRates, term structure, credit spreads, volatility indicesCore review, section 3
Factor families4: momentum, low volatility, size, short-horizon reversal46: Value and Quality addedValidation record
Construction variants1: sector-neutral momentum, documented with its own validationPlanned variants per familyValidation record
Sector neutralizationVariant only; not applied to production scoresApplied where the client question warrants it, stated per deliverableConstruction standards
Validation statisticsFama-MacBeth t (uncorrected), IR full and holdout, hit rate, maximum drawdown, turnover, net-of-cost spreadSame, reissued quarterlyValidation page
Newey-West standard errorsNot appliedProposed for library v0.2; effective only when recorded in the changelogChangelog
SimulationMonthly equal-weight deciles held one monthSameDaily-frequency simulation with position, turnover and liquidity constraintsSimulation
Transaction costsLinear haircut at 0, 10, 20, 40 bps per sideSameSpread-based estimation scaled by volume participationValidation record
Core Quantitative Market ReviewSpecimen issue, 9 pp, public12–16 pp monthly to subscribers; two-business-day email support from commercial launchEvidence
Monthly Quantitative NoteSeptember 2026 issue, 4 pp, public, with a governance record on its last pageMonthly, with a governance record on its last pageNote
Portfolio Evidence ReportSpecimen, 4 pp, on a hypothetical portfolio8–12 pp, quarterly per portfolio (monthly on Enterprise)Evidence
Manager Attribution ReportSpecification published; no specimen, no code pathFirst delivery, 6–10 pp, annual, up to three managersProducts
Substantiation FileSection plan published; no specimenFirst delivery, 8–12 pp plus computation CSV, delivered in 10 business daysProducts
Factor data feedDecile membership and score CSVs exported by export.py each run (Marnello_Factor_Deciles_2026-09.csv, Marnello_Factor_Scores_2026-09.csv); not yet delivered on the siteMonthly, at each score publicationProducts
Read-only REST APINoneNone; arrives 2027Read-only v0 (refresh at each score publication in v0; daily target at production)Products
Custom model licensingNone deliveredFirst engagements; five-phase process publishedProducts
Historical factor archive licenceNo archive yet; it begins at the first production month and cannot be back-filledArchive accumulates monthlyLicensing from Year 3Products
Data licence (point-in-time prices, membership, share counts)Not yet contracted; vendor scoping Q4 2026; target signature before commercial launchIn forceFundamentals licence added
Why we publish this

The gap is a roadmap, not a claim.

A prospect who opens a PDF after reading a marketing page should find the same numbers. Capabilities marked for launch depend on the data licence, which is not yet contracted; vendor scoping Q4 2026; target signature before commercial launch.

Related