Fantasy football draft strategy, by league type

Search "draft strategy" and you'll mostly get the same article: lock in RB and WR early, wait on quarterback, chase upside late. That advice is fine, but it's written for exactly one kind of league. Take it into a dynasty startup and you'll leave with a room full of 27-year-old running backs that's worthless by year two. Take it into a guillotine league and you might be the first team out.

We build the model behind our Draft Assistant, and it runs four separate modes because the formats genuinely disagree about what a player is: this season's points, a long-term asset, a lottery ticket, or insurance against one bad week. What follows is the strategy each mode actually plays. Use it with the tool or without it.

Redraft: value over replacement, tiers, and the deadline

Redraft is where the standard advice comes closest to being right, because the question is simple: who scores the most points for this roster, this season?

Even so, "the most points" is the wrong scoreboard. A 350-point quarterback is worth less than a 250-point running back if the QB you can grab in round 12 scores 300 and the last startable RB scores 150. You're not buying points, you're buying the gap between a player and the best free alternative at his position. In a 12-team league that starts two RBs and a flex, the free alternative at RB is roughly the 30th one off the board. At QB it's the 13th. That gap is the whole reason RBs and WRs dominate early rounds, and it's the number our engine ranks on.

Need is mostly overrated, until suddenly it isn't. Early in a draft, "I need a WR" is a bad reason to pass a better RB; our model caps the need bonus at 25%, enough to break a tie and nothing more. The exception is the deadline. Count your remaining picks and your empty starting slots. When those two numbers get close, need stops being a tiebreaker and becomes the assignment. Eight picks for three open slots, take value. Four picks for three slots, fill the roster.

Tiers answer a different question: not who, but when. Before each pick, ask what the position will look like when the draft snakes back around to you. If the answer is "roughly the same," wait a round and take the value elsewhere. If the answer is "a full tier worse," you've found the one legitimate reach in fantasy football. Be honest about it, though. A real cliff is a drop worth something like 15% of the player's own value in a single step. The gap between WR14 and WR15 is not a cliff, it's a rounding error.

One more thing redraft punishes hard: doubling up at a solved position. Once your tight end is set, a second good TE is worth about 40% of his sticker price to you, because he's now competing for your flex spot against every RB and WR on the board, and there's no next season to stash him for. Spend those picks at positions you start two or three of, or on the handcuff to your own injury-prone starter.

Dynasty startup: draft a portfolio, not a lineup

A startup looks like a redraft until about round three, when someone takes a 22-year-old with forty career touches over a proven veteran and both managers think they won the pick. They're playing different games. The veteran wins this season. The rookie might win 2027 through 2030.

Our engine prices startups on dynasty market value instead of season projections, and age is half the reason. Market value bakes in the whole career arc: age, situation, pedigree, the years of production still ahead. When a young breakout costs more than an older player who will outscore him this year, the market isn't being dumb. It's pricing years three through six, which a season projection can't see at all.

The other big change is that we cut the need weight in half, from 1.0 to 0.5. A hole at RB in a startup is a problem you can trade your way out of for the next several months. The elite asset you passed on to fill that hole is gone permanently. So let need break ties, and let it do nothing else.

This is also the one draft where "best player available at a position I'm already strong at" is a real strategy instead of a leak. Your second elite receiver isn't blocked, he's inventory. Somebody in your league is thin at WR, and by October they'll pay full price. The model still discounts a redundant pick here, but noticeably less than in redraft, for exactly this reason.

Last thing: startups run 25 or 30 rounds, so pace yourself positionally. Your next pick is a full lap of the draft away, nearly two if you're on the turn, and a position run hurts a lot more when you can't respond for 24 picks. Plan a round ahead, not a pick ahead.

Rookie drafts: take the best lottery ticket, every time

Rookie drafts are short, they come once a year, and managers ruin them the same way every time: drafting for the depth chart instead of for the asset.

Rookie outcomes are wildly spread. NFL front offices whiff on these players constantly, hit rates outside the top of a class are ugly, and "safe" prospects bust at rates that would crater a veteran's price. When the range is that wide, the tier gap between prospects is the only signal worth much, and roster fit is close to noise. Our model sets the need weight at 0.15 here, about a seventh of redraft, and honestly it could probably go lower.

Play the reach out. You're deep at WR, thin at RB, and the board hands you a receiver one tier above the best back. Take the RB "for need" and miss, and you got nothing while the better ticket went to a rival. Take the WR and hit, and you either start him or flip him to a WR-needy manager for the RB you wanted in the first place. The trade market converts surplus into need. Nothing converts a busted reach into anything.

Related: taking the same position three picks in a row is fine. A prospect sitting behind your stud isn't blocked, he's developing on his own clock while your stud ages on his. The only stacking the model really punishes is true hoarding, four deep at one spot while a starting slot sits open.

Treat the picks themselves as assets with a price chart, too. They're cheapest midseason, when contenders are shopping the future for a playoff push, and priciest in April when every selection feels like a future star. Rebuilders live in that spread. It also means selling a pick at peak hype is sometimes the best pick you can make.

Guillotine leagues: survive first, win later

Quick recap if your league hasn't tried this format yet: nobody has an opponent, every team's weekly score counts, and each week the lowest total is eliminated and its whole roster hits waivers. Last team standing wins.

That one rule breaks most redraft instincts, which is why guillotine gets its own mode in our engine with settings no other format uses.

Start with what you're actually scored on. There's no matchup to win, so ceiling is nearly worthless; a 180-point week and a 120-point week are the same result as long as neither was last. What kills you is the bottom. One 2-point dud from one starter in one bad week can be the whole season. So the guillotine mode scores weekly floor directly: consistent players get up to a 10% bump, boom-bust players lose up to 15%, and players without enough game history to measure take a flat penalty, because variance you can't price is still variance. The penalty side is bigger than the bump side on purpose. Dodging the zero matters more than maximizing the floor.

Two redraft rules flip outright:

• An empty starting slot goes from carryable weakness to emergency. A near-zero at one position, every week, is exactly how you post the low score. Guillotine runs our most aggressive need setting, above redraft, to get a real body into every slot early.
• A second startable player at a filled position goes from wasted pick to insurance. In redraft he keeps about 40% of his value; here we keep 55%, because a bye week with no replacement doesn't lose you a matchup, it can end your season. (A third and fourth body still compound down. This is insurance, not a collection.)

And don't pay a premium at tier cliffs. Every elimination dumps a full roster into free agency, so the pool restocks with starters all season long; the engine literally halves its tier-timing term in this mode. The draft's job is to survive September with your FAAB budget intact. The real roster gets built off dead teams in October.

The same engine, four different games

These are the actual settings, side by side:

One thing stays constant across all four: the rankings are built with zero ADP input. ADP tells you what the room will pay, which is useful for timing (will he last until your next pick?) and for spotting a faller. It tells you nothing about what a player is worth in your league's settings. Rank on value, and use the market to time your picks, not to make them. (There's a whole article on this: ADP is a price, not a value.)

Go deeper

Each of the ideas above has its own full guide:

• What is VOR? Value over replacement, explained
• Tier-based drafting: what a real cliff looks like
• ADP is a price, not a value
• Superflex draft strategy: the QB2 trap
• Floor vs ceiling: when boring wins
• How to value a fantasy trade
• Rookie pick trade value: buy in October, sell in April
• FAAB strategy: a budget that has to last
• Start/sit: a framework for 11:40am on Sunday

FAQ

When should I draft a QB in a 1-QB redraft league? Later than feels comfortable. The 10th-best QB costs almost nothing and scores within a few points a week of the fifth-best. Wait until the gap to replacement justifies the pick over the RBs and WRs still on the board, which in most 12-team leagues means well after the elite tiers are gone.

How do I know a tier cliff is real? Check two things: the drop should be unusually large for that position, and it should be a meaningful chunk of the player's own value, something like 15% in a single step. Most cliffs people panic about at the draft table are neither.

Should I punt this season in a dynasty startup? A full teardown is a bet on your own patience as much as on the market. We'd rather weight long-term value, take what falls, and let the roster tell you whether you're contending than force a rebuild in round one.

How different is a superflex startup? Bigger than it looks. A superflex slot adds about half a QB's worth of starting demand per team, and QB prices in the market values are already superflex-priced. Elite QBs become first-rounders. But that second QB is a half-weight need, so don't auto-draft your QB2 immediately after your QB1 while an elite RB or WR sits on the board. Full breakdown in the superflex draft strategy guide.

Should I draft for need in a rookie draft? Only as a tiebreaker. A real tier gap beats a roster hole every time, because you can trade surplus into need but you can't trade a bust into anything.

Is stars-and-scrubs viable in a guillotine league? It's how good managers die in this format. Two studs and a zero-floor bench will eventually hit the one week both studs go quiet, and there's no second chance. Ten decent floors beat two ceilings and eight prayers.

How much FAAB should I save for eliminated rosters? More than feels right on draft day. The best players you'll roster all season come off eliminated teams, and the managers who emptied their budget in week two just get to watch. Surviving the early weeks with money left is the whole strategy.