Snapcount

How we predict · the methodology, in plain terms
Methodology · July 2026

Fantasy projections are unaccountable point guesses.
Ours are measured, ranged, and graded in public.

This page explains how the Snapcount projection engine actually works — the season-regime architecture, the volume-times-efficiency decomposition, the walk-forward training that never sees the future, and the discipline that keeps the numbers honest. It uses real terms (Spearman, calibration, conformal) without the formula walls. When you want the exact math, the white paper has every definition and every number.

§ 1 · The problem

Nobody grades the graders.

Every projection site hands you a single number — "17.4 points" — with no range, no confidence, no track record you can audit, and no way to know whether that number came from a model or a hunch. When the projection misses, it's quietly replaced next week. The incentive is to look precise, not to be accurate.

Our answer is not "we're smarter." It's a measurement discipline: one un-game-able yardstick, training that structurally cannot peek at the future, and a public record of hits and misses — including the head-to-heads we currently lose.

§ 2 · The architecture

A season is not one problem. We model it as regimes.

Most projection systems run one model from September to January. But the information available in Week 1 (nothing current — only last season plus an offseason of churn) is a different world from Week 10 (nine weeks of role and usage data). Our serving layer switches estimators by week-of-season, and each regime's estimator is the one that actually won its validation — even when the winner is embarrassingly simple.

W1 blind prior-season book no current data exists W2–4 data-starved strict prior-season model trained only on past seasons W5–17 normal regime opportunity model × efficiency, walk-forward the core engine W15–18 motivation-distorted resting starters, clinched seeds roadmap: stakes-aware model Playoffs separate regime roadmap SERVED TODAY (solid) week-aware estimator switching, promoted 2026-07-12 ROADMAP (dashed) recognized regimes, models not yet validated
The season-regime timeline. Solid regimes are served by validated estimators today; dashed regimes are on the roadmap — we treat late-season motivation distortion and the playoffs as unsolved, and say so.

Week 1: the humbling regime

At Week 1 every within-season signal our models feed on is null — there is no current season yet. We spent five research iterations trying to beat a "model-free book" (each player's prior-season average, playoffs included, plus recent form for QBs) with real models fed offseason information: betting-market priors, roster churn, coaching changes, contract news. Every model candidate failed the gate; the simple book won. So the book is what we serve at Week 1 — it scores 0.571 weekly rank correlation on 2021–2024 backtests vs 0.550 for the recency family we'd otherwise use (+0.021). Serving the humble winner instead of the fancy loser is the whole philosophy in one decision.

Volume × efficiency: predict the job, not the highlight

The core engine doesn't predict yards directly. It splits every player-week into opportunity (how many targets, carries, or attempts will this player get?) and efficiency (what does he do with each one?). Opportunity is driven by role — snap share, target share, depth chart — and is far more predictable than efficiency, which mean-reverts hard. Our headroom audits show most of the achievable ranking edge lives in projecting volume.

PROJECTED VOLUME targets · carries · attempts walk-forward model on lagged role signals: snap share, target/carry share, depth-chart order, air yards × SHRUNKEN EFFICIENCY yards per opportunity · catch rate trailing efficiency pulled toward the position average — hardest for players with few opportunities (noisy ratios) = PROJECTION weekly ranking score + calibrated floor–ceiling range + boom / bust probability + expected-TD layer (in validation)
The decomposition. Volume is the predictable half; efficiency is deliberately shrunk toward position priors because hot streaks mean-revert. (Exception: shrinkage measurably hurts QB ranking, so QBs don't get it.)

Walk-forward: the model never sees the future

Every fit is walk-forward: to predict season S, week W, the model trains only on rows from before that point — earlier seasons, or earlier weeks of the same season. Random k-fold cross-validation is structurally invalid for this domain (it leaks future weeks into training), so we never use it. For Weeks 1–4 the rule gets stricter still: the model may train only on fully prior seasons, because the normal in-season window would overlap the weeks being predicted.

The estimators, per regime

RB / WR / TE volume projection SERVED

Regularized walk-forward models project each player's this-week targets and carries from strictly lagged role signals; a gradient-boosted median (quantile loss) produces the displayed yards. Beats the recency baseline in 4/4 backtest seasons at WR and TE, 3/4 at RB.

QB: honest recency NO MODEL CLAIMED

Every QB model we've built failed validation (0/4 seasons vs the baseline), so QB serves a tuned recency average — and we say so instead of shipping a model that grades worse than the naive answer.

Calibrated TD models SERVED

Touchdowns are predicted as probability distributions (P of 0, 1, 2+ TDs), not counts — boosted-tree classifiers with per-position probability calibration. Measured any-TD calibration error: 0.031.

Conformal ranges SERVED

Every projection ships as a floor–ceiling band targeting 70% coverage, width scaled per player by trailing volatility. Measured coverage: 0.699 RB / 0.712 WR / 0.708 TE — the target, hit, not asserted.

W1–4 early-season book SERVED

Week 1 serves the validated prior-season book; Weeks 2–4 serve the strict prior-seasons-only model. Promoted 2026-07-12 after five research iterations failed to beat the book at W1.

Boom / bust confidence SERVED

A calibrated probability that a player beats or misses his own expectation — a confidence layer, not an accuracy moat, and we label it that way.

§ 3 · The discipline

Promote or revert — with numbers, in a public ledger.

The architecture is table stakes. What actually separates the system is the loop every change must survive before it can touch a served projection.

HYPOTHESIS one idea per iteration LEAKAGE-CHECKED FEATURE pregame-legal by proof WALK-FORWARD EVAL 4 seasons, paired tests ADVERSARIAL VERIFICATION independent, refute-first PROMOTE into serving REVERT with numbers, logged every verdict — including NULL — is appended to the research ledger; no silent drops
The loop. Most iterations end in REVERT — and those verdicts are recorded with the same rigor as the wins, because a documented dead end is how you avoid re-digging it.

The ledger in action — myths we tested and refuted

The "slow-start team" myth: refuted

We built the prior — do some teams reliably start seasons slow under the same coaching staff? Across 127 same-staff consecutive-season pairs, early-vs-late form anti-persists (correlation −0.217): last year's slow start predicts, if anything, the opposite. Verdict: NULL — the feature is refuted and retired, on the record.

Coaching changes don't move Week 1 rankings — we tested

We hand-curated every coordinator change 2015–2026 (with citations) and fed new-coach and scheme-import features to the Week 1 model. It failed the gate in 3 of 4 seasons and lost to the simple book. The offseason's coaching digest is team-level news; Week 1 player ranking stays owned by last season's volume.

Betting-market priors don't crack Week 1 either

Pregame spreads, totals, and season win totals — the market's own digest of the offseason — failed the Week 1 gate both inside the model and as a re-ranking tilt. (They do show a real Week 4 hint, now cross-confirmed by four independent feature families — that's a roadmap item.)

Even our own Week 1 models lost — so we don't serve them

Five iterations of Week 1 candidates, every one graded against a dumb prior-season book. The book won every time. What we serve at Week 1 is the book. Discipline means the incumbent keeps its seat until something honestly beats it.

§ 4 · The receipts

Graded in public — wins and losses.

Every claim above traces to a published artifact you can open right now.

We froze the model on 2016–2024 data, then replayed the entire held-out 2025 season through it — 5,227 player-week projections, generated after the season ended, scored against what actually happened. A holdout scorecard, not a live track record — and labeled as such.

The board, graded week by week for 2024 — including the biggest misses, because a replay that only shows hits is marketing, not measurement.

Per-week graded scorecards in machine-readable form. Same fixed grading rule every week.

Head-to-head vs RotoWire and ESPN on identical players, weeks, and yardstick. It currently shows us losing — see below — and it ships anyway.

⚑ The honest standing, today
On the yards-ranking yardstick we are not the most accurate provider. On 2024 W9–18 head-to-heads (identical player populations), RotoWire leads us at QB and RB, ESPN leads us at WR and TE; we are #1 at zero positions. Our model beats the naive recency baseline at RB/WR/TE — the market leaders beat us. The gaps are small outside QB (−0.03 to −0.04 rank correlation) and largest at QB (−0.13, where we claim no model). We publish this because a scoreboard you only ship when you're winning isn't a scoreboard. The plan to close the gaps — and what each lever is worth — is in the white paper.
PositionSnapcountBest providerGap
QB.416.546 (RotoWire)−.130
RB.730.771 (RotoWire)−.042
WR.715.745 (ESPN)−.030
TE.643.682 (ESPN)−.038

Mean weekly rank correlation (Spearman), TD-free score, 2024 W9–18, all-common population — every row graded on players covered by truth, our board, and every provider. Source: market_scoreboard.json.

5,227
player-week projections in the 2025 frozen replay — every one published
~70%
measured range coverage (0.699 / 0.712 / 0.708 RB / WR / TE) — the target, verified
0
positions where we currently beat every provider — published anyway
§ 5 · Go deeper

Read the exact how — or use the board.

The white paper documents every formula, mask, hyperparameter, and validation gate on this page, with the same numbers — nothing rounded up for marketing.

walk-forward only anti-leakage gated adversarially verified 2025 season sealed, replayed frozen losses published