Every dynasty ranking you have ever argued about was built for one format. Most of them do not say which.
That matters most at quarterback. The same player is a cornerstone in a superflex league and a mid-round luxury in a 1QB league, and the gap between those two prices is the largest format effect in dynasty. This post measures it from the Edge table published September 8, then does the same for everyone else, for tight end premium, and for league size.
Every number comes from repricing the same table for different settings, so the players are held constant and only the format changes. No names appear; only the size of the move matters here.
How the model prices format
Each player in the Edge table has a base price and five slopes: team count, superflex, points per reception, tight end premium and points per passing touchdown. His value in your league is the base plus each slope times your setting; the Edge explainer covers where the base comes from.
The superflex slope is fit per player, from that player’s trades in superflex leagues against his trades in 1QB leagues. It is only fit individually when he has at least 30 observations including at least 8 in each kind of league; below that he inherits a position-level fallback. In the September 8 table, 107 of the 112 quarterbacks on the board carry their own fitted superflex slope, and the twelve most valuable are each priced on between 23,842 and 71,670 trade observations.
The quarterback haircut
Take the quarterbacks in a 12-team superflex PPR league, rank them by superflex value, and reprice the same league as 1QB:
| Quarterbacks, ranked by superflex value | Median value lost in 1QB | Range |
|---|---|---|
| QB1 to QB6 | 40.5% | 31.5% to 42.0% |
| QB7 to QB12 | 41.1% | 40.3% to 44.6% |
| QB13 to QB24 | 46.1% | 13.0% to 49.9% |
| QB25 to QB36 | 27.0% | 7.9% to 46.6% |
The twelve most valuable quarterbacks lose a median 40.6% of their value going to 1QB. A superflex quarterback price is worth about three-fifths of itself in a 1QB league.
Run the same table the other way and it reads as the premium superflex pays. The top twelve quarterbacks gain a median 68% to 70% over their 1QB price. Quarterbacks 13 through 24 gain a median 85%, the largest lift in the table, which fits the football: a QB2 is a weekly starter in superflex and a bench body in 1QB, so the format changes his job more than it changes a QB1’s.
The bottom bucket is where to stop trusting the exact number. Quarterbacks 25 through 36 range from 8% to 47% lost, because cheap quarterbacks trade less often in 1QB leagues and the contrast the slope is fit on is thinner. Read that row as “somewhere between a quarter and a half.”
Our own diagnosis of the model found that, where its superflex premium is wrong, it tends to be too small rather than too large, because the fit shrinks a weakly identified slope toward zero. If the table errs, the true gap is larger.
What it does to the board
The haircut changes who is at the top of the board:
| Overall board, 12-team PPR | Superflex | 1QB |
|---|---|---|
| Quarterbacks in the top 12 | 3 | 1 |
| Quarterbacks in the top 24 | 7 | 3 |
| Quarterbacks in the top 50 | 15 | 7 |
| Quarterbacks in the top 100 | 28 | 21 |
| First quarterback’s overall rank | 4th | 10th |
| Twelfth quarterback’s overall rank | 33rd | 68th |
The best quarterback on the board is the fourth asset overall in superflex and the tenth in 1QB. The twelfth-best quarterback, a starter in every league, is the 33rd asset in superflex and the 68th in 1QB, with 28 receivers and 23 running backs above him. The market prices him like a WR3 because only twelve quarterbacks start, and that line is the one that catches 1QB managers in trades.
Everyone else barely moves
The same repricing, for the 24 most valuable players at each other position:
| Position (top 24 by superflex value) | Median change in 1QB | Range |
|---|---|---|
| Running back | +4.1% | −1.6% to +13.0% |
| Wide receiver | −1.4% | −6.3% to +5.7% |
| Tight end | −4.9% | −21.8% to +4.3% |
Running backs tick up in 1QB, which is what you would expect when the flex spot that could have gone to a quarterback goes back to a runner. Receivers are flat. Tight ends drift down a little, with a wider spread that comes from a thinner position. If an offer’s value changes with format, the quarterback in it is doing the changing.
Tight end premium
Tight end premium runs the other way from superflex: it lifts one position and leaves the rest exactly where they were.
| Tight ends, ranked by no-premium value | +0.5 per reception | +1.0 per reception |
|---|---|---|
| TE1 to TE6 | +11.9% | +23.8% |
| TE7 to TE12 | +13.2% | +26.4% |
| TE13 to TE24 | +15.0% | +29.9% |
| Top 12 overall | +12.3% | +24.5% |
The lift grows as you move down the position, for the same reason the superflex lift peaks at QB2: the premium turns a marginal starter into a real one. Running backs, receivers and quarterbacks move by exactly zero, because the model pins the tight end premium slope to zero for anyone who is not a tight end. Before that rule, a thin-data running back could pick up a nonsense premium slope and swing between formats for no football reason.
League size
Team count is the third slope worth checking. Comparing a 12-team superflex PPR league with 10-team and 14-team versions of the same league:
| Position | 10 teams | 14 teams |
|---|---|---|
| Quarterback (top 12) | −9.8% | +9.8% |
| Running back (top 24) | −10.8% | +10.8% |
| Wide receiver (top 24) | −11.0% | +11.0% |
| Tight end (top 12) | −14.0% | +14.0% |
The symmetry is by construction: the model’s team-count slope is linear, so two teams up and two teams down move a player by the same amount. And 80.6% of the corpus is 12-team leagues, so a player gets his own team-count slope only when he has enough trades in other sizes; the rest lean on a position-level fallback. Read the size effect as roughly ten percent per two teams rather than a precise per-player number.
What the format does not reprice
Draft picks carry one price per pick in the Edge table, regardless of format. A 2027 first is the same number in superflex and 1QB. The players you draft with it will not be, so weigh your format yourself; the table does not.
Points per reception, the fourth slope, barely moves the top of this table: the 24 most valuable receivers lose a median 1.9% at half-PPR and 3.9% at zero, and the top running backs do not move. We have separately documented that the table’s PPR slopes read conservative against the market, so treat those as a floor rather than the whole effect.
Checking your own league
The Dynasty Rankings board has the superflex, PPR and tight end premium toggles built in and reprices live, so you can watch the quarterback column move. The Dynasty Trade Calculator does the same for a specific offer with your league’s settings.
Format also changes how much a package has to carry to buy a stud. In the July calibration study, the consolidation premium was 17.4% in 1QB leagues against 14.1% in superflex, because depth is easier to replace when only one quarterback starts. The Week 1 trade-market post covers that curve.
Then ask the co-manager. Once your Sleeper league is connected, it reads your lineup slots and scoring and knows you are in superflex before you say a word. Paste an offer with a quarterback in it and it prices him for your league.
Start free and see your quarterbacks priced for your format
Data note: every percentage in this post is computed from the published Edge table values/value_tables.json on sharpersunday-data@val-160 (schema 160-continuous-v1, generated 2026-09-08T04:52Z) by evaluating each served player’s coefficients with the scaledFeatureVector formula in api/values.js at a 12-team, PPR, 6-point passing touchdown reference and toggling one setting at a time (superflex on/off; tight end premium 0, 0.5, 1.0; team count 10, 12, 14; PPR 1, 0.5, 0). Positions are from the Sleeper player database. Buckets are ranked by superflex value (quarterbacks and non-quarterbacks) or no-premium value (tight ends); medians are of per-player percentage changes. The top-12 quarterback (median −40.6%, range 31.5% to 44.6%) and top-12 tight end (+12.3% / +24.5%) figures match those verified for the Week 1 post. The slope-fit rules (MIN_FORMAT_SPREAD_OBS 30, FORMAT_SPREAD_MINOR_OBS 8, position fallback), the pinned non-TE premium slope and the linear team-count slope are in scripts/trainValueModel.mjs; the 107-of-112 and 23,842-to-71,670 figures are the table’s per-player fitDims and observations; the 80.6% 12-team share is its coverageCaveats block. The “too small rather than too large” statement summarizes docs/SF_QB_PREMIUM_DIAGNOSIS_2026-06-24.md. The conservative-PPR-slope statement is in .claude/skills/prior-decisions/SKILL.md (Issue 670 entry) and docs/audit/TE_PPR_SLOPE_INSTABILITY_2026-07-24.md. The 17.4% / 14.1% consolidation premiums are from docs/BAND_CALIBRATION_STUDY_2026-07.md (run 2026-07-10). api/values.js serves stage1.picks without format terms. All figures describe the model at one moment, not a forecast.