Intentional Grounding

About & Methodology

Intentional Grounding is built around transparent, evidence-based rankings — not unexplained AI answers.

Why “Intentional Grounding”

The name is a play on the NFL penalty of the same name — and on the idea that every recommendation here should be intentionally grounded in evidence, statistics, news, and context, rather than generated without support. If Intentional Grounding can't show its work, it doesn't publish an opinion.

How scores are built

Every overall score starts from a base of 32 and adds or subtracts a weighted contribution from each category below. The score is always computed by code from these components — a language model may eventually explain the score in prose, but it never invents or edits the number itself. See docs/SCORING_MODEL.md for the full write-up.

Season Performance

±18

Fantasy production so far this season relative to position peers.

Recent Performance

±10

Trailing 3-week performance trend versus season baseline.

Expected Opportunity

±14

Projected target/carry/snap share looking ahead.

Matchup

±8

Strength of the upcoming opponent against this position.

Team Context

±8

Offensive line, supporting cast, and overall team quality.

Coaching / Scheme Context

±5

Scheme fit, play-calling tendencies, coordinator history.

Depth Chart Position

±6

Clarity and security of the player's role on the depth chart.

Historical Consistency

±6

Week-to-week scoring variance over the sample period.

Injury / Availability

±12

Current injury designation and games-missed risk.

News Impact

±6

Net effect of recent news (roster moves, coaching comments, etc).

The recommendation standard

Every future recommendation — start/sit, waiver, draft — is held to the same six-part structure. Uncertainty is never hidden behind authoritative-sounding language.

1. Recommendation

A specific, actionable call.

2. Why

Reasoning in plain language, tied to the score.

3. Supporting evidence

Sourced and timestamped.

4. Risks or conflicting evidence

Stated plainly, never hidden.

5. Confidence level

How sure the system is, separate from the score itself.

6. Source timestamps

So freshness is always visible.

Milestone 1: this is a foundation, not the finished product

The rankings, scores, and player data shown across this site are seeded demo values used to prove out the architecture. No live statistics, news, or injury feeds are connected yet. See the project README and docs/ROADMAP.md for what's coming next.