Tells you who to draft, pick by pick. Set your league once, then tick players off as they
come off the board — it re-ranks after every pick and shows you the odds each player is
still there when your turn comes back around.
Rankings come from 89 fantasy analysts, not from average draft position, so the board
can actually disagree with the room. Injury news is swept six times a day.
Set up your league. Get a board that actually knows your draft.
Three questions. Everything after this — who to take, who won't last
until your next turn, which positions you can afford to wait on — is built from your answers.
1
How many teams?
2
Which pick is yours?
First round. Snake drafts reverse each round, so this decides every
turn you get.
3
How does it score?
Extras — tick only what applies:
Roster spots and rounds — only if yours are unusual
Starters
Format
Scoring detail
Rankings from 89 analysts, not the crowd. Most tools rank players by average draft
position, which just tells you what everyone else already thinks. This one values players on
expert consensus and uses draft position only to work out when someone will actually be gone.
It watches the news for you. Injury and depth-chart reporting is swept six times a
day from the outlets that carry Schefter and Rapoport, and anything that changes a player's
outlook is priced in before you draft.
Who to draft next
VORP points above a replacement-level starter — the higher the betterCX insurance value if the starter ahead of him goes downSURV chance he lasts until your next pickSCORE board rank — best available first, per the experts
Your starting lineup
Projected0
Bench
Connect a live draft (optional)
Follow a real draft pick by pick. Paste the draft URL or ID; your
username is optional and only used to work out which seat is yours. Connecting reads the
league's real size, roster slots and scoring, and configures the board to match.
Not connected
ESPN, Yahoo or NFL.com? No public API, so the draft tab streams
its own text to this one. Copy the snippet below, paste it into the console on the draft
tab you already have open, and picks start appearing — nothing to reopen. Or just copy the
pick list and paste it in the box; any format works.
The bridge button is for pushing picks
from a script or the command line instead. It needs
python3 scripts/draft_bridge.py running, and only works from pages served
over plain http — a draft site cannot reach it, which is why the snippet uses the tab link
above rather than the network.
What analysts are talking about
Injury news
Who’s moving
All your picks
01
What to expect at each of your picks
01
Rankings by position
01
Compare players
Up to four side by side. Every line comes from a number in the
data — expert rank, how much analysts disagree, the offence's implied points, role,
bye, news, survival. Nothing here is invented scouting prose.
01
How to use this during a draft
Built for a 30-second clock. Every action is one tap or one key —
if something takes longer than that, it is the wrong way to do it.
1
Before the draft — set your league
On Home: how many teams, which pick is yours, and the scoring. Thirty seconds,
once. Your draft slot decides every survival percentage on the site, so getting it wrong
quietly makes everything else wrong.
2
When the draft starts — jump to your pick
Hit space once per pick that goes by, or tap + on the
pick stepper. The board assumes the highest-ranked players went first, which is exactly what
happens when other teams are on auto-pick.
Faster: if you are pick 10, tap + nine times before the draft even
starts. Takes three seconds and you are ready before the clock hits you.
3
On the clock — read the gold box, draft that name
The big name at the top is your pick. Underneath it: the position rank, and the odds he
survives until your next turn — that number is why he is first rather than someone ranked
higher. A player who will still be there later is not the pick now.
A second name is always listed under "if he's gone", so you are
never left thinking when the clock is running.
4
After you pick — one tap
Hit enter or tap I drafted [name]. That records him to
your roster and moves the board to your next turn. Then keep tapping
space as other teams pick.
5
If someone unexpected goes
Search the name and hit Taken. Only worth doing when a pick genuinely surprises
you — the board already assumes the obvious ones are gone. If a star slides to you, the
board will spot it and put him top on its own.
Zero-effort mode. On the Draft tab, open
Connect a live draft and copy the auto-sync snippet into your draft tab's console
(⌥⌘J on Mac). Picks then stream in on their own and you never touch
the stepper. Works on ESPN, Yahoo and NFL.com; Sleeper connects directly by draft link.
If anything goes wrong mid-draft, do not debug. The stepper always
works. Tap it to the pick you are on and keep drafting — you can sort the rest out afterwards.
01
Set up your league
Do this first — it takes about thirty seconds. Every number on the
site is built from these settings: who counts as replaceable, when your picks land, which
positions you still need. Change one and the whole board reprices. Saved to this browser, so
you only do it once.
Presets
Format
Your starting lineup
Scoring
What that implies
How far the ADP stretches. The market data behind the survival
model comes from 12-team, one-QB, full-PPR mocks. Two corrections are applied automatically:
ADP is scaled by team count to preserve round position, and positions whose demand
changes — superflex, 2QB, three-WR lineups — have their ADP pulled forward in proportion, so
a superflex board stops telling you to wait on quarterbacks. What is not corrected is
pure scoring change: in a TE-premium or non-PPR league the values reprice correctly, but real
drafters will take those positions somewhat earlier than this ADP shows. Read the survival
percentages there as a slightly optimistic ceiling.
The board refreshes itself. Scheduled jobs re-pull ADP each morning
and sweep the news wire six times a day until kickoff on 9 September, adjust the affected players
and republish this page. The timestamp under the title is the real fetch time; Market movers and
The wire show what changed.
An open tab checks for a new publish every ten minutes and whenever you return to it. If you have
not recorded a pick yet it just loads the new version. Once you are drafting it will never
reload on its own — it offers a Refresh button instead, because losing your scroll position
mid-round costs more than slightly old ADP. If the sandbox blocks the check entirely, the page
falls back to warning you once the data passes eight hours old rather than reloading blindly.
01
Draft-day cheat sheet
A worked example: eight turns, sixteen picks, hand-written for a 12-team
full-PPR draft from slot 10. Each pick lists the primary target, the fallback if it's sniped,
the pivot if a run breaks out, and what to refuse at that price. Names already off the board
are struck through as you draft.
01
The doctrine
Five things this rule set rewards that a generic ranking list will not tell you.
1
Buy the touchdown environment, not the depth chart
Four of the top sixteen players by ADP play for bottom-five scoring offenses — Achane
(MIA, 19.0 implied), Jeanty (LV, 19.2), Love (ARI, 18.5), Judkins (CLE, 18.7). Meanwhile
Kyren Williams on the league's highest-scoring offense goes at 32 and Tee Higgins on
the fourth-highest goes at 37.
2
The rushing-QB premium is real; the price is not
Rushing yards outscore passing yards 2.5-to-1 and rushing touchdowns outscore passing
ones 1.5-to-1 under standard rules, so a quarterback's legs are worth more than his arm.
But the QB curve is flat through the middle — QB1→QB6 is 72 points while QB6→QB12 is only
41 — so take a rusher, and take him late. The exact break-even for your scoring is on the
next tab.
3
No TE premium makes tight end the cheapest punt on the board
TE has the flattest curve of the three skill positions, so a round 1–4 tight end carries
the highest opportunity cost anywhere. Offense-adjusted, Loveland at 57.9 beats both McBride
(31.7) and Bowers (38.0) — who play on the 31st and 28th offenses.
4
Your defense will actively lose you games
Points-allowed and yards-allowed net to roughly +0.36 points per game —
statistically zero. So all D/ST scoring rests on sacks, turnovers and defensive TDs, and
sacks are halved here. Median game: +4. A WR3 is +11.
5
Bench spots are insurance, and the market closes at the draft
192 players get rostered on draft night. Waivers cannot repair an injury. RB injuries hand
~100% of vacated carries to one identifiable player; WR injuries split three ways — so
contingency concentrates at running back. Target 5 RB / 2 WR / 1 flex, and zero backup
QB, K or DEF.
02
Where the rules break
The derivations behind every recommendation on this site.
Rushing-QB arbitrage — break-even
Rushing production
Equivalent passing
D/ST — what a week actually pays
Game
Score
Shutout, 180 yds, 5 sk, 3 TO, 1 TD
+25
10 pts, 280 yds, 4 sk, 2 TO
+13
22 pts, 330 yds, 2 sk, 1 TO — median
+4
31 pts, 420 yds, 2 sk, 1 TO
0
35 pts, 420 yds, 1 sk, 0 TO
−5
38 pts, 470 yds, 1 sk, 0 TO
−7
Streaming targets. Any defense facing these offenses is a top-5 weekly
play: NYJ 18.4 · ARI 18.5 · CLE 18.7 · MIA 19.0 · LV 19.2 · TEN 20.6 · CAR 20.8. Seven offenses,
roughly 119 team-games — far more streaming spots than a 17-week season needs.
03
Which archetype actually wins
200 Monte-Carlo drafts per opening from slot 10. Opponents draft on ADP with
realistic noise; the engine plays out every remaining pick.
Archetype
Opening
Median
p10
p90
Spread
Hero-RB
RB / WR / WR / WR
2001
1970
2016
46
Engine free-choice
RB / WR / RB / WR
1986
1940
2010
70
Zero-RB
WR / WR / WR / RB
1981
1949
1999
49
Dual-anchor
RB / RB / WR / WR
1976
1947
1999
52
Balanced
WR / RB / RB / WR
1975
1945
2004
59
Robust-RB
RB / RB / RB / WR
1919
1894
1976
81
Elite TE
RB / WR / TE / RB
1896
1869
1942
73
Early QB
RB / WR / QB / RB
1894
1873
1942
68
The board ranks best available, not scarcest. Within a position it follows
expert consensus exactly — the WR ranked 5th is always offered ahead of the WR ranked 11th.
A scarcity adjustment is then applied per position, never per player, so it can push
a whole position down when waiting costs little but can never invert two players at the same
position.
That choice costs about 35 projected points against ranking purely by scarcity
(1936 vs 1971 across 40 controlled drafts). The scarcity-first version scored higher and
produced picks no one would make — it preferred a WR11 over a WR5 sitting on the board purely
because the WR5 would not survive another turn. The trade is deliberate.
Positional scarcity is worth ~45 of those points. Pure best-available, with no
scarcity term at all, scores 1891 and spends round three on the TE1 whose round-eight
replacement is nearly as good.
Elite-TE and early-QB remain the expensive mistakes in the archetype table below,
which was measured under the scarcity-first ranking. Read it as evidence about
positional timing, not as a description of what the board now does.
04
How this board ranks players
Four steps, all of them running live in your browser as you mark picks off.
Value from experts, availability from the market. A player's rank for valuation
purposes is 65% expert consensus (~90 analysts, refreshed daily) and 35% ADP. Ranking talent
by ADP alone is circular — it prices everyone at whatever the crowd already thinks, so the
board can only ever re-sort the crowd and never disagree with it. ADP still decides survival,
which is the question it genuinely answers.
Projection. Positional value curves calibrated to full-PPR outcomes under this exact
rule set, scaled by the team's Vegas implied total, and a manual multiplier reserved strictly
for player-specific news. Offence quality is priced once, by the Vegas term — an
earlier build also hand-adjusted for it, double-counting on 86 players.
VORP. Surplus over positional replacement — RB32 / WR30 / TE14 / QB13. Multipliers
scale the surplus, not raw points; scaling raw points manufactures phantom edge at the
top of steep curves.
Contingency (cx on each row). P(role opens) ×
VORP if promoted, gated behind a complete starting lineup. This is what makes handcuffs
legible at all; without it a bench is invisible, since every bench player sits below replacement
by construction. It does not subtract his current value — you never start a
sub-replacement bench player, so that shortfall never actually happens, and subtracting it
rewarded being worse.
Survival. Each player's odds of lasting to your next pick, from a logistic on his
measured ADP standard deviation in the mock-draft sample — not a fitted guess. Real
dispersion is far tighter than it looks: under a point at the top of the board, where an earlier
version of this page assumed three and a half, and was correspondingly too gloomy about who
would still be there.
VONA (the score column). Value over the next available
player at that position on the round trip, using survival odds. Raw VORP says take Josh Allen in
round 3 — true in isolation, useless in practice, because the QB curve is flat through the middle.
VONA prices what waiting a turn actually costs.
What it can't see. Projections are ADP-derived, not independently
modelled — relative ordering is the deliverable, absolute totals are scaffolding. Survival treats
players as independent, which understates the odds a whole tier vanishes at once — when a run
starts, trust your eyes over the percentage. And the ranking is
greedy: in rounds 1–4 the written Script beats it (1931 vs 1901, and with a far uglier tail — a 221-point spread against dual-anchor’s 77), because it optimises one turn ahead and can't see
that the second WR is structurally required later.