New to these reports? Start here
- Dotted-underlined words have a plain-English definition — hover or tap them. Every term is also on the glossary page.
- "Null" means we found nothing, not that something broke. Most reports here are negative results, on purpose — knowing an idea doesn't work is the point.
- Two questions get asked separately. First, is the effect real? Second, is it already priced into the betting odds? An effect can be completely real and still useless to bet on.
- A "calibration" row is a self-check. It runs the same method on something already known to be true. If that fails, the whole report is unreliable — so it's reported alongside the findings.
- If a confidence interval includes zero, the real effect might be nothing at all, so no claim gets made.
Defense-first fantasy: is the scoring skill or luck? (2026-11-02)
The question from the group chat: a format where you draft defensive players (DBs, edge rushers) and pick a whole offense scored on points and chunk plays. Garrett's worry: defensive stats are discrete, so scores would be mostly luck.
Test: split-halfsplit-half reliabilityA repeatability test: split the data in two (here, odd vs even seasons) and check whether the same units score similarly in both halves. If a trait does not repeat, it was noise, not a trait. reliability. Each player's mean score in odd weeks vs. even weeks, players
with ≥ 5 games in each half, 2023–25 regular season. r near 1 = the scoring rewards something
repeatable; near 0 = luck. Scripts: extract_charged.py, reliability.py → reliability.json.
| Scoring | DL | LB | DB |
|---|---|---|---|
| Standard IDP (solo 1, ast .5, sack 4, TFL 1, QB hit 1, INT 6, PD 2, FF 3, TD 6) | .59 | .72 | .54 |
| Big plays only (sack, INT, FF, TD) | .44 | .53 | .21 |
| Disruption (sack, TFL, hit, PD, INT, FF, TD; no tackles) | .58 | .64 | .29 |
| Yards charged to the tackler (completions and runs) | .38 | .88 | .61 |
Offense PPRpoints per receptionA fantasy scoring format that awards a point for every catch, which raises the value of high-volume receivers. for comparison: RB .86, WR .84, TE .81, QB .61. A whole-team-offense unit (points + 2 × 20-yard plays + 0.5 × first downs − 4 × turnovers): .54 (points alone .54, chunk plays .41, turnovers .18).
Reading: - Garrett is right about big-plays-only. For DBs it's .21, mostly luck. - Standard IDP is as repeatable as QB scoring (.54–.72). Tackles are what make it work. - "Yards charged to the tackler" is repeatable, but it measures being near the ball, not allowing yards. LBs score .88 because they tackle the most. As a penalty it would cancel out tackle points and punish good tacklers. Don't use it. - True coverage responsibility isn't free. nflverse on-field participation stops at 2025, and coverage assignments are paid data (PFF / Next Gen). Passes defended and INTs are the free stand-ins. - A team-offense unit is a fine slot (.54, like a QB), mostly driven by points scored.