FairValue pre-open

Methodology

How each fair value is produced, with the standing defaults from WATCHLIST.md.

Pipeline (each trading day, ~7:30 AM ET / 5:30 AM MST Phoenix)

  1. Financial Data Collector pulls prices, shares, cash/debt, beta, TTM financials, consensus FY1/FY2, multiples, earnings dates and news from public sources (Yahoo Finance, stockanalysis.com / S&P GMI, U.S. Treasury) into data/DATE/.
  2. DCF Analyst runs a 5-year DCF (FCF after SBC) with bear/base/bull and a WACC × g sensitivity grid into dcf/DATE/.
  3. Relative Valuation Analyst values each name on peer NTM P/E, NTM EV/Sales, TTM EV/EBITDA and FY1 P/FCF (25th / median / 75th) into relative/DATE/.
  4. Fair Value Coordinator blends, applies margin-of-safety rules, flags issues, writes reports/DATE.md, emails it, and rebuilds this site.

Fair value and call rules

  1. Blended FV = 0.5 × DCF base + 0.5 × relative mid. If the DCF is not meaningful (negative FCF), FV = relative mid, low confidence.
  2. Divergence = |DCF base − relative mid| ÷ relative mid. > 25% → indicative; > 50% → low confidence.
  3. Margin of safety = (FV − price) ÷ FV. Cheap if ≥ 20%; Rich if price ≥ 15% above FV; otherwise Fair. SPCX (short history, wide ranges) uses 30% both ways.

Comps basis change (from the next run)

New default from 2026-11-11 (Kevin approved): comps are SBC-consistent with the DCF. P/E and P/FCF use GAAP EPS and FCF minus SBC (not adjusted / non-GAAP consensus); EV/EBITDA uses EBITDA minus SBC where available. If only adjusted consensus exists, it is converted using the latest SBC and noted.

Today's comps (2026-11-11) still predate that change, so they are labelled adjusted basis on every ticker page. That mismatch (DCF deducts SBC, comps don't) is a main reason DCF and relative differ by more than 25% on every name with a DCF. Runs from 2026-11-12 onward are labelled by what the analyst file states.

Default assumptions (from WATCHLIST.md)

Methods

Ticker-specific notes

Rules

Delivery

Key assumptions as applied in the 2026-11-11 report

Known data gaps (2026-11-11)

Change log