How the dynasty values work.
Displayed values come from Dynasty Cheat Code's own market-aware model. We do not sell or republish consensus ranks.
Market-aware means format scarcity, age curves, and resisting short-term hype — not pasting KeepTradeCut or FantasyCalc. Inputs are public facts: Sleeper rosters and depth charts, plus nflverse production and draft records. Buy/sell/hold is our outlook on where a player's dynasty value is heading, not a gap versus someone else's price.
Values are our own: talent (production weighted against draft pedigree), age curve, role security, upside, contract, and risk, scaled by format scarcity. We do not sell or republish consensus ranks. Scale: 0–100 base, open-ended after format adjustment.
Values are built from the player’s own profile and your league format. Where a player has not proved anything yet, the model is intentionally conservative, so unproven names do not leap established assets on draft pedigree alone.
The model reads multi-season production, a position-specific age curve, role stability, draft pedigree, next-season opportunity, upside, and risk together. The exact blend, thresholds, and calibration are proprietary.
Calls are a directional outlook: bullish means we expect a player’s dynasty value to rise, bearish that we expect it to fall. That is not a gap versus a consensus price. Elite cornerstone assets get extra context so “bearish” does not become “dump him.”
Inputs, not rankings feeds
Modeling lives in the Python engine under dynasty_engine/. The Next.js app consumes precomputed values; it does not re-rank players in the browser. Inputs are public facts: Sleeper identity, age, team, and depth chart; nflverse production, usage, and draft records. There is no market component in the score — market_score is always null — and we do not ingest paid consensus boards or ADP as the board.
Talent blends what a player has done with draft pedigree as a fading prior. Early in a career, pedigree carries more weight; as real games accumulate, production takes over. Production above the draft slot can supersede a low pick faster than a weak season can bury a high pick — one quiet rookie year, especially at receiver, is poor evidence either way. Players with no NFL snaps take an unproven discount so capital alone cannot float them over proven elites.
Age curves and position baselines
Peak windows are position-specific: roughly quarterback 27, running back 24, wide receiver 25, tight end 26. Decay after peak uses a convex curve — mild near peak, severe far past it — so a thirty-year-old still in his prime is not priced like a thirty-eight-year-old. Running backs decay fastest; quarterbacks hold longest. Pre-peak players get a youth premium that also varies by position.
On the base 1QB scale, equal evidence does not mean equal price across positions. Quarterbacks and tight ends sit lower than running backs and receivers because replacement is cheaper in 1QB and standard TE scoring. Running backs are not hit with a flat positional tax on top of the age curve — short careers are priced through peak age and steep decay. Superflex and TE-premium restore scarcity in the format layer instead of baking every format into the base board.
Role, opportunity, risk
Role security rewards franchise quarterbacks, bellcow backs, alpha receivers, and primary tight ends, then refines with snap share normalized to a full workload for that position. Unproven roles regress toward a prior so a synthetic depth-chart constant is not treated as measured evidence. Opportunity is the forward-looking tilt: when volume left a team, the engine hands a share of freed targets or carries to the players positioned to claim it. That term is deliberately modest — a vacancy is real, but it is still a projection. Risk combines injury risk, age decay, and role uncertainty.
Formats
Format settings change positional scarcity. Superflex and 2QB push quarterbacks up the board on a continuous ramp — elites rise most; streamers do not leap skill-position starters. TE-premium lifts elite tight ends more than replacement tight ends. Deeper lineups raise WR/RB depth value; shallow lineups concentrate value in elite difference-makers. Elite Superflex quarterbacks can exceed 100 by design and are normalized at display time rather than clamped into a flat ceiling.
Buy / sell / hold
The engine labels BUY when a non-elite asset is still short of peak with real role and upside, or when an elite asset is still ascending. SELL requires age past the position’s late window plus a real reason — weak role or elevated risk — so age alone never dumps a locked-in veteran. Everything else is HOLD. Public badges are Bullish / Bearish / Neutral. Plain-language companions: how to use buy/sell and how to read the DCC verdict.
Refresh & freshness SLA
Values refresh on a published cadence (Tuesday, Wednesday, and Saturday in-season). Every tool that uses the board shows an as-of stamp. If that stamp ages past 72 hours in-season or 7 days in the offseason, a banner appears so you never rank silently on stale data. The model is built to resist short-term hype rather than chase a spike.
League analysis
Importing a Sleeper league grades every roster with league-relative percentiles: starter value, franchise value (players plus picks), young core, and draft capital. Directions range from Elite Contender through Deep Rebuild, including Win-Now Fragile Contender, Retool, and Productive Struggle. Contender score and rebuild score are separate blends; confidence is distance to the nearest boundary. Future picks are valued from round and original-owner projected finish, discounted for distance.
Trade finder partners prefer window mismatch and positional fit, with rails so rebuilders are not told to buy aging backs. Guides: how to use league import and rebuild vs contend.
What we can prove
The validation numbers live only on /track-record. The honest headline from the published walk-forward: last-season PPG correlates at 0.660 and DCC value at 0.601. We do not rewrite that comparison here or invent other win rates. Dynasty value is still the right object for multi-year trade decisions; it is not a claim that we beat “just use last year’s points.”
FAQ
Are these copied market rankings?
Displayed values come from Dynasty Cheat Code's own market-aware model. We do not sell or republish consensus ranks. Market-aware means format scarcity, age curves, and resisting short-term hype — not pasting KeepTradeCut or FantasyCalc. Inputs are public facts: Sleeper rosters and depth charts, plus nflverse production and draft records. Buy/sell/hold is our outlook on where a player's dynasty value is heading, not a gap versus someone else's price.
How is a displayed player value calculated?
The model combines multi-season production, a position-specific age curve, role stability, draft pedigree, next-season opportunity, upside, contract, and risk. The exact weighting and calibration are proprietary.
Why can a hot player still be a sell?
Our calls are a forward-looking outlook, not a price comparison. A productive player can still read bearish if his age curve, role security, or opportunity points down. Elite assets are protected from blunt sell labels.
How do Superflex and TE premium change values?
The format layer accounts for positional scarcity. Quarterbacks rise in Superflex and 2QB, while premium tight ends receive a format-specific lift.
How often are values refreshed?
In-season the values job runs Tuesday, Wednesday, and Saturday. Freshness is stamped as-of on tool pages. If values age past 72 hours in-season or 7 days in the offseason, a site-wide banner warns you — we do not silently rank on a stale board.
Do private Sleeper leagues import?
Yes. Paste a username, league ID, or the league URL from the Sleeper app. Private leagues import the same way as public ones — Sleeper’s read API returns them. A missing league is a wrong or deleted ID, not a privacy wall.
Short hub: /faq · Data policy: privacy · rankings · buy / sell / hold