How Signals builds its dynasty rankings.

Dynasty is the deepest game in fantasy football — and one of the hardest valuation problems contained within it. A redraft ranking has to answer one question: how good will this player be this season? A dynasty ranking has to answer that one, plus all the others: How long will he be good? How does a 24-year-old's next five years compare to a 29-year-old's next two? What is a rookie pick worth before anyone knows where in the draft it will land and whose name gets called? And what will the manager across the table actually pay?

We built Signals because we wanted answers with receipts. We don't watch film, we don't have favorites, and we don't rank on vibes — every component of our dynasty value earned its place by predicting real future production in backtests, and the trade values you see were calibrated against thousands of real dynasty trades, not our opinion of what a fair trade looks like.

One thing to know up front: those answers lean forward. Our board deliberately tilts toward youth, upside, and outlook — it's built to sit where our models believe the market is heading, not where it stands today — so you can get ahead of age curves and breakouts at the trade table instead of reacting to them after everyone else. We'll make the full case for that tilt later in this guide.

And the same honest limit applies too, maybe even more in dynasty than anywhere else: models can't see everything. Dynasty is a game of conviction — the year-early breakout call, the aging star you move one season before the cliff. That's why your Signals board is yours: pin any player to the rank you believe in, and everyone else flows around your call. We give you the most rigorous starting point we know how to build. The conviction is yours to add.

Here's the short version, then the deep dive:

A Signals dynasty value is built in two layers. First, a global ranking — a league-independent positional ordering of every player, built from a production model (what he'll do), career-arc modeling (for how long), leading-indicator signals (which direction he's trending), and a market anchor (what the dynasty community actually pays). Then a league layer re-prices that order for your league — whether it's 1QB, superflex, or tight-end premium, all of your roster positions, roster depth, and scoring settings are analyzed to optimize the relative positional value for your league's exact setup. And the whole board deliberately leans young and forward-looking — less a mirror of today's market than a forecast of where it's heading — so you can be proactive about age curves and breakouts, not reactive.

One board, two layers — and why value follows the slot

The architecture is the first thing to understand, because everything else hangs off it.

Layer one is the global positional ranking, and it's league-independent on purpose. Whether you play superflex or 1QB, 10-team or 16-team, the order of our global dynasty board is computed from a single neutral baseline — so the rankings you study, the rankings your leaguemates see, and the rankings behind every tool in the app are all the same board. That consistency matters more than it sounds: a value system where the player order quietly changes depending on which page you're looking at is a system you can't build trust in. And even better, you only have to maintain one set of rankings across all of your leagues.

Layer two turns that order into your league's prices. Here's the simple version. The global board decides the order — who's the WR1, the WR2, the WR30. Your league decides what each of those slots is worth. A shallow 10-team league concentrates almost all the value in the top slots; a deep 14-teamer keeps real value flowing far down the board. We say it around here like a mantra: value follows the slot, not the player.

The cleanest illustration is superflex. Josh Allen is the QB1 on the global board whether you play 1QB or superflex — the order doesn't change. What changes is what QB slots are worth: start a second quarterback and you've roughly doubled the league's demand for startable QBs, so every QB slot's price rises, and a solid veteran starter in the QB12–15 range goes from an afterthought to a genuinely valuable asset. Same player, same rank, very different price — because the slot he occupies is worth more in your league.

And your overrides are part of layer one, not a cosmetic layer on top. When you pin a player to a rank, we call it a "rock in a river": your pinned player holds his slot, and the live value order flows around him. Pin Malik Nabers as your WR2 and he takes the WR2 slot's value; everyone he leapfrogged slides down one slot and takes that slot's value, automatically. Every derived number in the app — league analyzer values, trade calculator verdicts, roster grades — all respect your custom rankings, because they're baked in where the order is built, not painted on where it's displayed.

The rest of this guide is about how that global order gets built — and then how your league, the trade market, and rookie picks get priced on top of it.

What a dynasty value is made of

Dynasty value starts where our redraft model ends. The full team-first projection engine described in our redraft guide — team ecosystems, conserved opportunity, availability, the market check — produces the "what is he right now" input. Then the dynasty composite blends that present-day production with everything that describes the future. Six components, each measured for every player:

• Production — the player's current projected output, straight from the redraft engine. The present is still the best single predictor of the future, and no dynasty model should forget it.
• Trajectory — which direction he's trending, measured as real year-over-year production movement. A second-year riser and a fading veteran can post the same season total and deserve very different futures.
• Opportunity — the quality of his role: snap share, target share, and weighted usage. Role is the leading indicator of production, and it often moves a season before the box score does.
• Efficiency — what he does per opportunity, blended with positional norms in proportion to track record, exactly as in the redraft model.
• Insulation — how durable his situation and career window are: age, role security, the stuff that determines whether this year's production repeats for five more. A target monster is insulated; a touchdown-dependent role is not.
• Upside — the ceiling window. Youth carries option value in dynasty, and this component prices it explicitly instead of letting "he's only 23!" float around as a vibe.

Here's the part we're proud of: the components aren't weighted by intuition — they're weighted by outcome backtests, position by position. We tested which components actually predict future production at each position, and the weights follow the evidence. Quarterback value leans hardest on present production and upside, because QB careers are long and the good ones stay good. Running back value leans on opportunity — for a RB, the role very nearly is the value, and the data validates it. Tight end value leans on insulation, because an established TE role is one of the most persistent assets in fantasy while an unestablished one is a rumor. Wide receiver spreads its bets most evenly. When a component can't be measured for a player, its weight redistributes to the others instead of silently pretending.

The leading indicators

Underneath those components sits a layer of per-player signals, each expressed as a percentile within the player's position and refreshed throughout the year — the usage-driven ones weekly, the prospect grades as new information lands:

• Trajectory — the year-over-year production delta described above.
• Opportunity — the role-share measurement.
• Efficiency — per-opportunity performance, including the licensed charted signals explored below.
• Prospect grade — for young players, our backtested scouting score, which folds college production, athletic testing, and draft capital into a single number. It's important enough to get its own guide in this series.

These signals are what let the dynasty board move early — the ascending sophomore whose role grew all December, the veteran whose target share quietly eroded — instead of waiting for a full season of box scores to say what the usage already said. Jaxon Smith-Njigba's target share was climbing for months before the box scores made his breakout obvious; the usage told the story first.

The advanced signals, in plain English

The efficiency layer includes professionally charted data we license — the same dataset from the redraft guide, but tested here against a different question: not "who scores next season," but "whose talent holds up over multi-year horizons." Some metrics fail the one-season test and pass the multi-year one, and vice versa; each signal earns its seat per model. Here are a few examples of what's in the dynasty layer, what each metric means, and why it matters:

• Yards after catch over expected (receivers). Every catch comes with an expected YAC based on where and how the ball arrived; this metric measures what the receiver creates beyond that. It's the cleanest separator we have between production a scheme handed a player and production the player made himself — and self-made production is what travels across coordinator changes, quarterback changes, and years. It's the kind of signal that tells you a rookie tearing through tacklers, a Brock Bowers or a Puka Nacua, is announcing real talent rather than riding a friendly scheme.
• Rushing over expected (running backs). The same idea on the ground: given the blocking and the defensive front, what did the back create beyond what the play design gave him? It's how you spot a talented back buried in a bad situation — the profile Bucky Irving showed in 2024, when his per-touch creation screamed long before his workload caught up, and the workload duly followed.
• Scramble rate (quarterbacks). How often a QB turns dropbacks into runs. It sounds mundane, but it's one of the most persistent traits a quarterback has — far more stable year to year than accuracy stats — and in fantasy, rushing value is the difference between a good QB and a league-winner. A Jayden Daniels rushing floor compounds across every remaining year of his career, and this signal prices it.
• Role-quality context (everyone). Target share, snap share, and weighted usage measures that tell you what a player's role is before the season totals tell you what it produced. Usage is the earliest honest signal in football — Puka Nacua's league-winning breakout was visible in his target share from the first week, not just in the record-setting box scores that followed.

The discipline is the same as everywhere else in the model: every one of these was validated against real multi-year outcomes before it earned a weight, and charted metrics that sound just as smart — but didn't predict — were left out.

Young players run on prospect grades

For players in their first couple of NFL seasons, production history is too thin to carry a valuation — so the composite injects our prospect grade: the same backtested scouting model that powers our prospect rankings, distilled into a single score per player. The grade blends the full pre-NFL picture — college production profiles, measured athletic testing (which matters most exactly where NFL data is thinnest), and draft capital, where real teams spent real picks and which remains the single most durable prior about a young player's opportunity. Rather than juggling those threads separately, the dynasty model consumes them the way they were validated: together, as one grade with its own promotion gates and its own strict no-hindsight discipline (a prospect grade may never peek at how the player's NFL career actually went). The prospect model gets its own methodology guide in this series.

Rookies are also ranked inside the dynasty board, against everyone. Your rookie-draft view is just the dynasty board filtered to the incoming class — so a rookie's rank always means the same thing everywhere, and "he's my rookie 1.03" and "he's my WR28" can never silently disagree.

Age and the career arc — dynasty's defining variable

In redraft, age is an adjustment. In dynasty, it's half the game. Two players with identical seasons ahead of them can sit thirty slots apart on a dynasty board, and they should — because dynasty value is the whole remaining arc, not the next 17 games.

Derrick Henry is the cleanest illustration in football: for years he's been a top-shelf redraft pick and a far cheaper dynasty asset at the same time, and both prices were right. Redraft buys this season. Dynasty buys the window — and the window is what our career-arc modeling prices, using the same measured curves we detailed in the redraft guide, extended across years:

• Running backs depreciate fastest. The age-28 cliff and the doubled bust rate past it mean an RB's dynasty value front-loads hard into his early-to-mid 20s. This is why a 24-year-old workhorse like Jahmyr Gibbs is dynasty royalty while equally productive backs five years older trade at a fraction of the price — the market isn't being disrespectful, it's being an actuary.
• Wide receivers hold value longest. Stable through their late 20s with the gentlest decline of the skill positions, which is why the elite-young-WR archetype — the Ja'Marr Chase and Justin Jefferson class — is the gold standard of dynasty assets: top-of-position production and the longest remaining window in the game.
• Quarterbacks are dynasty's bonds. The position ages so gracefully that an established QB1 can anchor a roster for a decade — which is precisely why superflex changes everything (more below).
• Tight ends develop late and are honest-to-goodness weird. Rookie TEs almost never matter, established TE roles persist beautifully, and the aggregate aging data is murkier than any other position — we said so in the redraft guide and the hedge applies double across a multi-year horizon. When a Brock Bowers breaks the "rookie TEs don't matter" rule, that's the exception announcing a special player, not the rule changing.

And the same honesty from the redraft guide carries over: these curves are cohort averages built on survivors. They are priors, not verdicts — Henry has been making the RB curve look silly for half a decade. When you have conviction the curve is wrong about your guy, that's what the override is for.

Our board skews young — on purpose

Put the career arcs and the leading indicators together and you get a board with a personality, so let's name it before you notice it yourself: compared to dynasty ADP and most ranking sets, Signals tilts toward youth, upside, and outlook — and away from aging production. If you're used to boards that are basically redraft rankings with a light age discount, some of our placements will look aggressive. That's not an accident, and it's not recklessness — it's the point.

Here's the reasoning. A lot of dynasty rankings — and dynasty ADP itself — are anchored hard to current production. But current production is the one thing every manager in your league can already see, and the age curves that erode it are among the most predictable forces in football. The market knows all this and reprices anyway — late. It waits for the aging star's down year to actually happen, then crashes his price all at once. The value didn't disappear that November; it had been leaking for a year while the ADP stood still.

So we build the board to be early instead. Think of our rankings less as a snapshot of today's market and more as a forecast of where our models believe the market will be in the near future. The ascending sophomore whose role and underlying signals are surging sits higher on our board than his current price, because his price is heading there. The 28-year-old back coming off a fine season sits lower than his name value, because his price is heading there too, and no one sends you a calendar invite for the cliff.

The payoff is that you get to be proactive instead of reactive. Dynasty is won at the trade table, and the trade table rewards whoever moves first: selling the aging producer while his price is still at its peak, buying the riser while he still costs a second. A board that mirrors today's ADP can only ever tell you what everyone already knows — by the time it moves, the window moved with it. A board that's pointed a few months ahead is the one that hands you the trade before it's obvious.

And if you think we're too early on someone — that's a real risk of a forward-looking board, and we own it — that's exactly what the override is for. Rerank him where your conviction says, and the whole system follows you.

The market anchor — priced against what dynasty managers actually pay

A dynasty value that ignores the market is a philosophy, not a tool. You don't get to trade with a spreadsheet; you trade with the manager across the table, and what the community actually pays is real information. So the global composite is anchored to a dynasty market signal assembled from three kinds of evidence:

• Community dynasty value markets — established public consensus values, aggregated rather than trusted singly.
• Real startup ADP. We continuously aggregate startup drafts from real dynasty leagues — actual leagues with actual stakes, not mock drafts — to see where players are really being taken. And because a startup draft ages, the data is recency-weighted: a draft from a few months ago carries a fraction of the weight of one from last week, so the anchor tracks the market as it moves rather than where it stood in June.
• Trade-implied values. We continuously ingest real dynasty trades from across the ecosystem and infer market prices from what players actually clear for — the purest market signal there is, and the same trade corpus that powers the Recent Trades feed on every player card.

Two things make our anchor unusual, and both were measured rather than assumed:

1. The anchor's grip varies by position — because we backtested who's right. When we tested our composite against the market on actual future outcomes, the model out-predicted the market for running backs, receivers, and tight ends — so the anchor holds those positions more loosely and lets the model's opinion breathe. At quarterback the market proved as sharp as the model, so QBs stay tightly anchored. Trust is allocated by scoreboard, not by ego — in both directions.
2. Conviction is allowed, but it has to be earned. The model can pull a player off market consensus — that's the youth tilt you just read about — but how far it's allowed to pull is bounded and evidence-gated. A model that never disagrees with the market is redundant; one that disagrees freely is arrogant. The tuning lives in the measured middle.

Your league re-prices the board

Now layer two, in practice. Your league's actual roster template and scoring produce its value curve — and this is where dynasty leagues, which vary far more than redraft leagues do, get treated like it:

• Superflex is the big one. As in the Josh Allen example up top: starting a second QB roughly doubles the demand for startable quarterbacks, so the whole QB curve re-prices — not just the top. A startable veteran backup goes from roster clogger to real asset. In 1QB leagues the same players re-price back down. Same order, different world.
• Tight-end premium lifts the TE curve's ceiling — in leagues that reward the position, an elite TE like Brock Bowers or Trey McBride gets scarcer-than-scarce pricing, because he's producing WR2 numbers from a position where replacement level is a shrug.
• League depth changes the tail. In a 14-team league with deep starting lineups, the WR55 is somebody's flex and priced accordingly; in a 10-teamer, the same guy is a waiver churn body and the curve says so. Value concentrates at the top in shallow leagues and spreads downward in deep ones.
• Best ball is the one exception to "never re-rank." With no weekly lineup decisions, ceiling beats floor — so best-ball leagues get a deliberate upside tilt within positions, the single place league settings are allowed to reorder players. Everywhere else, order is sacred and only prices change.

Calibrated against nearly 8,000 real trades

Here's the section we'd put on a billboard if billboards were cheaper.

A value system's real test is the trade market: if your numbers call trades "unfair" that sharp dynasty managers make every day, the numbers are wrong, not the managers. So we replayed 7,965 real dynasty trades with Signals values through the exact math our Trade Analyzer uses and compared its verdicts to how the market actually behaves.

The raw model failed that test, and the failure taught us something. Textbook value-over-replacement math says value should fall off a cliff below the stars. The real market disagrees — dynasty managers pay genuinely real prices for mid-tier players and rookie picks, trade after trade, year after year. When theory and eight thousand receipts disagree, the receipts win. So we bent the value curve toward real clearing prices, position by position — with one hard rule: calibration changes prices, never order. No player moved a single rank.

The before-and-after tells the story best with rookie picks: before calibration, the analyzer graded first-round rookie picks dramatically below what the market pays for them — the kind of error that would have told you a fair pick trade was a fleecing. After, pick pricing sits within a whisker of the market. The verdict thresholds you see on every trade (Fair / Solid / Strong) were then recalibrated the same way — set from the actual distribution of real trades rather than round numbers we liked.

Two more trade-market realities the analyzer prices because the data demanded it:

• Packages get a quantity discount. Three decent players for one star is the oldest trap in dynasty — three WR3s do not add up to a Ja'Marr Chase, because roster spots have value and depth pieces don't sum linearly. The analyzer dampens multi-player packages, so a 3-for-1 is judged the way the market judges it, not by naive addition.
• Your roster context matters. A trade that's fair on raw value can still be wrong for a contender or right for a rebuild — which is why the analyzer reads your actual roster and evaluates the deal in context, not in a vacuum. (That's a whole feature of its own, but it stands on the values described here.)

Rookie picks: priced off the class, not off folklore

Rookie picks are dynasty's currency, and most pick charts are folklore — a decaying curve someone drew years ago. Ours are priced off the actual class:

• A pick's value starts with where the pick will actually land. We project every team's picks from its own projections, and in season we recalibrate toward actual results — so you can see which picks are likely to land early and which late. A first from a rebuilding team and a first from a contender are not the same asset, and a one-size-fits-all approach to pick valuation misses exactly that.
• Then the pick is priced as the player you should expect at that slot. We project the incoming class with our prospect model, order it, and value the 1.03 as "the third-best prospect in this class, priced in your league." A class with a Jeremiyah Love at the top is worth more than a thin one, and the picks know it.
• Current-year picks carry a small option premium. A pick is a player plus a choice — the right to react to landing spots, injuries, and draft-day slides. The market pays for that flexibility, and so do we.
• Future picks are discounted by distance, because a 2028 first is a promise about an unknown class from a team whose record you can't know yet.
• Startup picks work the same way in the other direction: during a startup draft, a pick is valued as the best player projected to be available at that slot, straight off the live board.

A value that lives and remembers

Dynasty is a long game, and the model is built to be watched over months, not glanced at once:

• Every player's value and rank are snapshotted daily, powering the trend sparklines on every player card — so "he's been quietly climbing for six weeks" is a chart, not a hunch.
• The leading-indicator signals refresh weekly, the market inputs continuously, and the production engine underneath rebuilds daily with injury news triggering immediate updates — the same living pipeline from the redraft guide, feeding the dynasty layer.
• The rankings you see are the rankings the system uses. The server that writes the daily history runs the same model code as the app you're looking at — not a reimplementation that can quietly drift. Boring engineering discipline, but it's the reason your sparkline and your board can never tell you two different stories.

How we keep ourselves honest

Same rules as the redraft model, with dynasty's own flavor:

• Outcome backtests, not taste tests. The composite's position weights, the anchor strengths, and the career-arc curves were all validated against what actually predicted future production — with strict no-hindsight discipline throughout. And a seat in the model isn't tenure — signals get re-tested as the foundation under them improves, and the ones an upgrade makes redundant are retired.
• We grade ourselves against independent expert boards — not to copy them, but to catch our own wiring errors. When we diverge from smart people, we want it to be on purpose.
• Every model change re-runs the trade-fairness harness against the real-trade corpus before it ships. If a tuning idea makes the trade verdicts worse, it dies, no matter how clever it sounded.

From the cutting-room floor, dynasty edition

Same policy as the redraft guide: we show you the rejects, because the graveyard is where the credibility is.

• The seasonal market-trust dial. Dynasty markets have a rhythm — startup drafts dominate the offseason, trades dominate the season — so we built a version of the anchor that shifted its trust with the calendar. Measured and rejected: the recency-weighting already inside the ADP was quietly solving the staleness problem on its own, and the calendar dial's only real effects landed on deep-roster players where no reliable benchmark exists to prove it helped. A second mechanism doing a solved job, verifiable only where verification is impossible, is not an improvement. Out.
• A dozen anchor formulations that couldn't beat the incumbent. We swept variation after variation of how tightly the model should hug the market. Nearly all of them looked fine — which is exactly the problem. "Looks fine" isn't the bar; provably better is, and only the shipped configuration cleared it.
• "Fix the picks by inflating the prospects." When pick values first graded out below market, the tempting lever was juicing rookie values until picks looked right. The data said no: the actual miss was in depth-player values, and fixing those fixed the picks. The lesson generalizes — when an output looks wrong, find the input that's actually broken instead of bending the nearest knob until the symptom goes away.

We keep this list for the same reason as the redraft one: anyone can tell you why their rankings are right. We'd rather show you the ideas we killed to get here.