A base rate built from the data, an adjusted rate from the tape, a composite if you want both. Read this once and you'll know exactly what the board is and isn't telling you — and which number to draft against based on how you evaluate.
Rookie projections are an industry of confident guesses. Most of them aren't wrong because the guesser is dumb — they're wrong because nobody bothered to ask what the base rate looks like, and nobody bothered to watch the tape. We do both. This document explains how.
01 · The problem
Walk into any rookie draft and you'll hear analysts confidently project a college kid as a future WR1 — based on what, exactly? A highlight reel and a vibe. We wanted the prior. Across twelve years of SuperFlex ADP, how often does a rookie at any given draft slot actually pay off?
But the prior alone isn't enough. The prior tells you what the average prospect in this slot did historically. It doesn't tell you what this prospect on tape can do. So we layered in the All-22 study — five-plus games per prospect, by hand, every class. The base anchors the projection in twelve years of evidence. The tape edits it for what the data can't see.
02 · The base rate
We pulled SuperFlex ADP from every season 2014 through 2025. Bucketed each rookie by position and ADP range — 1.01–1.06, 1.07–1.12, 2.01–2.06, 2.07–2.12, third round, fourth round. The base rate for any given cell is "of the N rookies who landed in this bucket, what percentage hit?" Intentionally simple. The simplicity is the point.
The full matrix lives in the hit rate tool — position × ADP bucket, every cell, with sample sizes on hover.
03 · Hit definitions and sample inclusion
Hit definition is where every analytics shop differs. Ours is the one that maps to lineup decisions: full-PPR top-12 QB, top-15 RB, top-24 WR, top-12 TE in any season of the rookie's first contract. Position-specific because positions score differently. Top-of-positional-tier because that's the threshold above which a rookie is actually starting on your team.
The board exposes three hit thresholds — 1+, 2+, 3+ — for the readers who want to ask what it takes to be a sustained league-winner versus a one-year wonder. Each threshold uses a different sample window because each one needs a different amount of NFL runway to evaluate fairly:
The 2025 class is excluded from all three thresholds — one season is too small a sample to evaluate against any of them. The graduated rule is what keeps each rate honest: we're not punishing recent prospects for having less time to accumulate hit seasons, and we're not crediting older prospects for runway recent classes haven't had yet.
04 · The film eval
A base rate cell is a population-level statistic. It says: "In the last twelve classes, X% of the rookies in this bucket hit." That's useful — it's the prior any sane projection has to anchor against. But it can't see route tree. It can't see contact balance. It can't see what happens on third-and-six when the quarterback hangs the ball over the middle.
That's what the All-22 is for. Every prospect on the board gets five-plus games of charting — every snap, by hand, with the tape paused, rewound, paused again. We're looking for the things the data can't show: separation patterns, ball-tracking, processing speed, finishing through contact, whether the route stem actually breaks where the chart says it does.
The tape's verdict either nudges the base rate up or down. Adjusted rates can't move more than 12 points in either direction.1 The data still leads; the tape edits.
The base rate is honest about the past. The tape is honest about this player. The composite is what you draft with — if you want both.§ 04 · The film eval
05 · Three lenses
Not every dynasty manager evaluates the same way. We didn't want to force one. The board exposes all three numbers — BASE, ADJ, COMP — and the right answer is whichever one matches how you actually think about prospects.
BASE — the analytics lens. If you trust the historical prior over any single shop's tape, draft against BASE. Twelve years of evidence is a hell of an anchor and we won't fault you for sticking to it. The base alone is a defensible model — better than most published rankings.
ADJ — the film lens. If you trust the tape over any historical aggregate, draft against ADJ. You're saying: "I want HIVE's All-22 verdict, full weight, applied to this prospect's prior." We watch tape because we believe it adds signal the prior misses. If you do too, lean here.
COMP — both, equally. A locked 50/50 blend of BASE and ADJ. If you think both inputs deserve equal weight, COMP is the one number that respects both. This is where we land internally — and it's why HIVE has been asked half a dozen times to expose the weighting as a slider. The answer is no. A slider invites a user to overweight the layer they like and call the result "their model" — which is exactly the failure mode this whole exercise was built to avoid. If the weighting ever changes, the methodology changes with it, and the reason will be published alongside.
06 · How we'll keep it honest
Each summer we publish how the previous rookie class actually performed against the model and update the methodology if the gap is structural rather than noise. The 2025 scorecard is already live on the hit rate tool. The 2026 scorecard will land in summer 2027.2 We expect to be wrong sometimes; what we won't do is rewrite history when we are.
