Year one of a keeper league is great. Everybody keeps a guy, everybody feels smart, nobody complains.
Year three is when you find out what your rules actually do.
That's when one team has kept the same receiver for three seasons at a price that stopped making sense two Augusts ago, half the league has quietly checked out by Week 6, and somebody in the group chat types the sentence every commissioner dreads: "we should probably just redraft next year."
The rules didn't break. They worked exactly as written. The problem is almost nobody reads a keeper rule and asks the only question that matters: what does this reward, and what happens when it rewards it four years in a row?
The real problem is compounding, not fairness
Every keeper format is fair on day one. Everyone gets the same option.
What separates the good ones from the league-killers is whether the advantage stacks. If a team can keep a stud cheap, they enter the draft with more resources than everyone else. More resources means a better draft. A better draft means more studs to keep cheap next year. Do that three times and the gap isn't a gap anymore — it's the standings.
Meanwhile the guy who finished 11th gets to keep his best player too, which sounds like help but usually isn't, because his best player is a guy he drafted in the 4th and now has to keep in the 3rd.
So when you're picking a format, don't ask "is this fair." Ask "does this compound."
Round-cost keepers
How it works: you keep a player at the round you drafted him, usually one round earlier each year you keep him. Drafted a guy in the 10th? Keep him in the 9th, then the 8th.
What it actually rewards: hitting on late-round picks. That's it. Which is a genuinely good thing to reward — it's the closest thing fantasy has to a skill.
Where it breaks: undrafted players and waiver pickups. If a guy went undrafted and blew up, what round is he? Whatever you decide, somebody's mad about it. Most leagues slot free-agent adds in the last round, which is the single most abusable rule in this format — the best breakout of the season becomes a last-round keeper for three straight years.
Verdict: the most popular format for a reason. Fix the waiver-pickup hole before your first season and it stays healthy for a long time. A pickup should cost a mid-round pick, not the last one.
Escalating-cost (auction / salary) keepers
How it works: you keep a player at what you paid, plus a raise every year — a flat bump or a percentage.
What it actually rewards: knowing when to let go. The raise is the whole design. A great value doesn't stay a great value forever; it prices itself out, and you have to decide which season is the last one worth paying for.
Where it breaks: if the raise is too small. A $5 player going to $10 to $15 is still a steal in year three. That's not a keeper rule, that's a subscription. Bump it hard enough that the decision is actually hard.
Verdict: the best format for stopping the compounding problem, because the advantage has an expiration date built into it. Costs you a bit of simplicity — you need a real record of what everybody paid, which is exactly the thing leagues lose between seasons.
Years-limit keepers
How it works: keep whoever you want, but only for two or three years total, then he goes back in the pool.
What it actually rewards: nothing in particular. It's a cap, not an incentive. Which is fine — sometimes a cap is all you need.
Where it breaks: it's blunt. A guy who's a bargain and a guy who's barely worth keeping both hit the same wall. And there's no cost to keeping, so everybody keeps their best player every year regardless of value, which makes the whole thing feel less like a decision.
Verdict: great as a second rule stacked on top of round-cost or escalating-cost. Weak on its own.
The one nobody thinks about: how many
Format gets all the arguing. The number does more damage.
One keeper is a fun wrinkle. Two or three is a real strategic layer. Four-plus and you've quietly become a dynasty league without anybody voting on it — the draft stops being the main event, new managers can't join without inheriting somebody's ruins, and last place in year three is genuinely unrecoverable.
If your league has any turnover at all — and every league does — keep the number low. One or two. You can always raise it; you cannot lower it without somebody losing a player they'd already planned their season around.
Decide these five things before you draft, not during
Every keeper argument I've ever seen traces back to one of these being undecided:
- How many keepers, and can a team keep fewer if they want?
- What a waiver pickup costs to keep. Name the round or the dollar amount.
- How long a player can be kept in a row.
- The deadline to declare keepers — a real date, before the draft order is even known, so nobody's declaring based on where they pick.
- What happens on a trade. Does the keeper cost follow the player, or reset for the new team? This one is the sneaky-worst offender. If cost resets, every contender dumps picks for cheap keepers at the deadline and the trade market goes sideways.
Write them down. Not in the group chat, where they scroll away by Thursday — somewhere the league can actually look them up in eleven months, when somebody swears the rule was different.
The part that actually kills keeper leagues
It's not the format. It's that nobody remembers what happened.
Keeper rules only work if the history is intact. What round he went in. What you paid. How many years you've had him. Who traded whom in Week 9 of a season two years ago. Lose that, and every August turns into forty messages of "I'm pretty sure I got him in the 11th" versus "no way, that was the 7th" — and whatever the commissioner decides, somebody's starting the season annoyed.
That's what Big 4 HQ is for. Your league's history stays in one place season over season, so keeper decisions come with receipts instead of vibes.
Pick the format that fits how competitive your league actually wants to be. Then write it down, and keep the paper trail — that's the part that gets you to year five with the same twelve guys still talking to each other.