# FOOTO PREDICTO 3.26.2 — 2026 defensive scheme and dated injury evidence

Release date: October 30, 2026. Base: 3.26.1 scorecard deployment.

## Verified defensive scheme update

The 2026 Minnesota Vikings have retained Brian Flores' general disguised-pressure/zone defensive architecture while substantially increasing the frequency of pressure calls. The Vikings' official October 29 write-up quotes 69.9% blitz rate in 2026, compared with a first-three-season combined 44.9% under Flores. The prior model's Minnesota blitz value was 57.521%. The new unconditioned **simulation prior is 67.672%**: 82% of current published 69.9% plus 18% of the prior model estimate. That shrinkage is a conservative modeling choice, **not** an estimated confidence interval.

A current Sharp Football Analysis coverage report (updated October 27) lists **86.5% zone**, replacing the stale 75% projection in the shipped 3.26.1 input. The coverage group total is now 86.5%. Individual Cover 0/1/2/3/4/6/9 distributions are still **2025 evidence, reweighted to the 2026 overall zone rate**, and must not be described as exact 2026 shell charting.

The code now treats evidence as a team-season record (`defensiveScheme` on each supported team), without hard-coding Minnesota inside the rules. Other teams continue using their measured or prior-regularized tendencies until supported current-season evidence is added. A modest blitz/zone association is applied and calibrated so the overall reported zone rate is preserved. Individual blitz decisions also respond to down, distance, field position, opponent QB pressure history, and unit pass-blocking versus pass-rush. This is situational behavior, not a prediction that 69.9% will occur against every opponent.

The calibrated game mode applies an **opponent-sensitive, bounded ±1.0 expected-point contextual correction** only when there is sufficient QB pressure evidence and relevant unit blocking data. This replaces a universal Minnesota defensive bump, which would double count observed team results. The adjustment is heuristic and unbacktested; no predictive accuracy improvement is claimed. Player target allocation, simulated pressure, sacks, coverage and play descriptions also receive the updated playcalls.

Sources:
- Official Minnesota team report, Oct. 29: https://www.vikings.com/news/lunchbreak-defense-built-on-blitzing-nfl-trade-activity-2026
- Current coverage report, Oct. 27: https://www.sharpfootballanalysis.com/stats-nfl/nfl-coverage-schemes/amp/
- October 29 scheme analysis: https://www.prideofdetroit.com/detroit-lions-analysis/178389/lions-week-8-preview-breaking-down-vikings-offensive-defensive-schemes

## Injury and roster eligibility

Philadelphia's November 1 game against Washington now includes its official **Thursday October 29 practice report**, including Saquon Barkley, DeVonta Smith, Dontayvion Wicks, Jihaad Campbell, Jonathan Jones and Britain Covey as nonparticipants; limited/full practice statuses for other named players are also captured. These are **not final game statuses**. Each is projected available by default unless a subsequent confirmed Out/inactive/reserve restriction qualifies. No automatic workload reduction was inferred from the report.

Previously recorded IR/suspensions continue to exclude roster-ineligible players. A dated reserve restriction continues to exclude a player beyond the seven-day injury-report scoring window until an explicit activation is verified. PHI practice report coverage is through Oct. 29, while the other teams retain the original Oct. 26 archive. The UI now displays individual team injury evidence dates instead of implying the entire league was refreshed.

The Cloudflare worker's embedded availability snapshot was refreshed alongside the static `availability.json`; its daily cache key was versioned to prevent serving an old release snapshot after deployment. The on-request injury refresh remains opportunistic: **it does not run automatically in the background**. New confirmed official designations or inactives after this release must be imported via refresh or future maintenance.

Sources:
- Official Philadelphia practice report, Oct. 29: https://www.philadelphiaeagles.com/news/eagles-at-commanders-injury-report-2026-nfl-week-8-sunday-night-football-saquon-barkley-devonta-smith
- Official injury table: https://www.philadelphiaeagles.com/team/injury-report/
- Reserve-status cross-check: https://www.cbssports.com/nfl/teams/PHI/philadelphia-eagles/injuries/

## Forecast ledger and grading

The 3.26.1 Week 8 pregame records and all finals remain frozen and marked as historical. **They are not retroactively revised or regraded using the new October 29 evidence.** New single-game simulations use 3.26.2. Refreshing future archived forecasts would require a separate before-kickoff snapshot and new revision label.

## Verification

- JavaScript syntax checks for patched modules, worker and UI passed.
- Current MIN blitz prior 67.672%; overall coverage group exactly 86.5%.
- Simulated coverage and blitz probabilities match bounded expected rates within Monte Carlo uncertainty.
- PHI Thursday DNP/LP statuses do not produce automatic absences; current IR cases do.
- Five full MIN at DET games completed with calibration and player/accounting outputs; all smoke assertions passed.
- This is a focused release smoke test, **not** the full previous 242-test regression suite or a historical forward backtest, because the full 3.26.1 source archive was unavailable in a valid extractable format. Do not infer model accuracy from software tests.

## Installation

Upload the **flat deployment ZIP** directly to the existing Cloudflare Pages Direct Upload project. A second source-overlay ZIP contains the changed runtime files and a test script, suitable for applying over a verified 3.26.1 source tree. No Cloudflare publish has been performed in this session.