How we simulate drafts

• Players and projections. Every active player with position, team, age and status, plus their projected stat lines.
• ADP. Nine lists in all. PPR, half PPR and standard, the same three again for dynasty startups, then one each for superflex, dynasty superflex, and rookie drafts.

• Consensus rankings. By scoring format and scope, IDP included. Browse the rankings
• Tie-breaks only. Separates players our own numbers call too close to call. Never overrules them.

In an ESPN redraft league, this is the list we use to predict the rest of the room when you flip the board to ADP. Dynasty, rookie and superflex drafts read Sleeper's instead, because ESPN doesn't measure those rooms.

• Trade values. For players and draft picks, dynasty and redraft.
• Roster strength. Lets us look at how good a team's receivers already are, not just how many they have. Three good ones and they don't need a fourth just because the slot is open.

• Your settings. League size, scoring rules, starting lineup, who already owns whom, keepers and draft order, straight from Sleeper or ESPN.
• Why your board differs. Everything above is the same for every user. This is the part that makes your league's board different from every other league's.

Start with every active player, then remove three groups: anyone already on a roster, anyone at a position your league never starts, and anyone with no projection.

• Sleeper. Kept players leave the pool and join their owner’s roster before pick one. A team that kept two receivers starts strong there and never gets told to take a third.
• ESPN. Tells us how many players a team kept but not which ones, so kept players stay in the pool. The board can forecast someone who is already spoken for.

Joe Bryant's value based drafting is the linchpin of our model. Not how many points a player scores, but how many more than a replacement on the waiver wire.

• Get projected points. Sleeper gives us stat lines; your league's scoring turns them into points.
• Find the baseline. The player you measure everyone else at that position against. Ours is the last one worth rostering, so a 12-team 1QB league lands around QB16.
• Subtract. Positions become comparable. It's why an elite tight end is worth more than his point total suggests, and why the highest scoring QB is worth less.

We group players of similar value into tiers so they get compared against each other rather than against the whole board. Four percent of the pool's value spread sets a tier, around 25 in all.

Inside a tier, expert consensus rank decides.

• Dynasty leagues. Dynasty rankings carry most of the weight in a player's value: we re-deal each position's values in dynasty rank order, then blend in this season's projection. That's what separates a 24-year-old from a 31-year-old projected the same.
• Rookie drafts. Projections undersell rookies, so we push them up the board to account for it, while leaving room for a veteran who is clearly worth more.

The board works out whose pick each slot really is: snake or linear, keeper slots already gone, traded picks moved to whoever holds them now.

At every slot we score the next twelve available players against the roster that team has built so far, and only that team updates afterwards, so it answers differently every time it is on the clock. Four things adjust that score:

• Roster need. Judged on the team's positional rank against the rest of the league, not a headcount.
• Kickers and defenses. Deferred until the roster has to be legal.
• Surplus. Already have two tight ends and a third gets marked down, even when he's still the best value left.
• Bye crowding. Too many of your players already off in the same week.

Highest score takes the cell. The next few become the suggestions, and on your own picks those are what you see.

Your league mates are predicted by average draft position for your format, because that is what a real room drafts on.

That models your league mates, not your pick. ADP never reaches your advice, which is always ranked on value.