Team Resilience Tracker
Experimental observation only • Zero weight in picks
Competitive response is measurable; “heart” is not. This first version tracks recovery when trailing by 1–16 at halftime and fourth-quarter scoring when the game enters the quarter within 8 points. No qualifying adversity means insufficient evidence—not poor resilience. Teams without enough qualifying situations are labeled “INSUFFICIENT SAMPLE — MORE QUALIFYING GAMES NEEDED” rather than “Collecting data.”
Loading verified quarter scores…
| Team | Rating /100 | Games with quarter data | Halftime opportunities | Close Q4 opportunities | Status |
|---|
Method and limitations
A score is withheld until a team has four games with valid quarter data, three distinct qualifying games, and at least one opportunity in each component. These are conservative display gates, not proof of statistical reliability.
Second-half net scoring and fourth-quarter net scoring are adjusted using a pregame strength proxy: half the difference between teams’ prior average scoring margins, each shrunk with four neutral prior games. Expected scoring is allocated by period length. Halftime residuals are divided by 16 and fourth-quarter residuals by 8, then capped from −1 to +1. Each component is 50 + 50 × sum of capped residuals / (opportunities + 4); the overall score averages the two components equally. These scales are exploratory, not fitted probabilities. Overtime is excluded from quarter components.
The tracker uses current-season regular-season finals only. Missing or inconsistent quarter scores are excluded; games are deduplicated by event ID. Opponent strength, coaching, quarterback quality and luck remain confounders. Turnover-response and play-by-play resilience are not yet measured. Predictive value must be tested on future games before any model weight is considered.
Data: ESPN final quarter scores. Refreshes at most every six hours through the existing scheduled sync; daily snapshots are retained when storage is available. Historical observations are descriptive, not historical pregame forecasts.