Survivor pool: NFL rules, elimination and how strategy scales

A survivor pool is an office or online contest, not a sportsbook line. You pay an entry. Each NFL week you name one team to win its game. If that team wins, you survive and pick again next week from the teams you have not used. If that team loses, you are eliminated. The last living entry wins the pot.

That is the whole spine. Everything else is a house rule.

Straight up, not the spread

If Kansas City is −7 and wins 20–17, a spread bettor loses and a survivor picker survives. If they lose 17–20, both lose. The point spread is irrelevant except as a hint about how likely a win is. Buying a sportsbook ticket on the same game is NFL betting, a different product with a hold. Do not mix the two in your head: one is a pool against other pickers, the other is a price against a book.

One team, one use

You pick each franchise at most once in the regular season. Use Buffalo in week 2 and you cannot use Buffalo in week 16, even if that later spot is a home game against a weaker opponent. That scarcity is the game. An 18-week season and 32 teams sounds generous until upsets burn your "easy" list and you are staring at a Thursday night with three bad choices.

Deadlines

Some pools lock the entire slate at the Thursday kickoff. Some lock each pick at that game's kickoff. Late locks let you see Thursday and Sunday morning scratches; early locks punish indecision and reward people who write a card on Tuesday. Know which one you bought.

Ties and voids

NFL regular-season ties are rare but they happen. House options: a tie kills you, a tie is a win (you survived), a tie is a "push" that does not use the team. Postponements, abandoned games and cancelled games need a line in the rules too. If the commissioner shrugs in week 14, you do not have a pool, you have an argument.

Survivor pools live in the Games of chance topic because the weekly result is still a random football game. The skill is allocation, not covering −3. A 70% side still loses three weeks in ten. Over a long season those three weeks are why "I only pick locks" still dies. Plan the inventory around that, not around a fantasy of sixteen straight covers.

Last entry wins. If two or more entries survive the Super Bowl — or week 18, if you do not play the postseason — they usually split. Some pools reset the used-team list for the playoffs; some do not, which can eliminate you because both remaining teams are already on your card.

Everyone dies the same week. If the last four entries all lose on Sunday, common rules are: those four split, or those four are revived and play week+1, or a secondary tie-break (margin, leftover "strength") decides. Write it down in August.

Variants that are not the default

A two-strike pool is closer to a season-long budget. A 2,000-entry single-life pool is a survival problem that often needs 16–18 correct picks. An 12-person office pool may be over in November. Strategy that ignores field size is the wrong strategy.

Money: if the pot is cash, this is gambling. Adults only, 18+ or the local legal age. An office spreadsheet does not make it a raffle just because you know the commissioner.

The naive plan is "pick the biggest favourite every week." That plan has two leaks. First, you spend elite teams early, then sit through a brutal November with leftovers. Second, you sit on the same team as most of the field, so when that favourite finally loses you die in a crowd and nobody else dies.

Winning a survivor pool means you are alive and other people are dead. A pick that wins 80% of the time and is on 70% of cards is a good survival pick and a bad differentiation pick. A pick that wins 65% and is on 8% of cards is a worse coin-flip for your own life and a better bomb if it hits.

Worked comparison (illustrative probabilities)

Suppose 100 entries remain. Two reasonable sides:

• Team A: you judge 80% to win. 70 other entries will pick A (70% of the field).
• Team B: you judge 65% to win. 8 other entries will pick B.

If you take A:

• You survive with probability 0.80.
• If A loses (0.20), you die with most of the room; about 70 cards vanish with you.
• If A wins, the field only shrinks by the people who took worse sides.

If you take B:

• You survive with probability 0.65.
• If B wins and A loses, you wipe a huge cluster and the pot's expected share jumps.
• If B loses, you die while the A crowd sails on.

There is no universal best row. In a 12-person pool that is likely to end by week 10, taking A every time you still have A is often correct: you need fewer unique bombs. In a 1,000-person pool that will last into January, saving A for a week when the public is elsewhere, and taking B when the crowd is jammed on A, is how you avoid a 200-way split at the end.

Future value

Before you tap a 75% home favourite in week 3, look at weeks 8, 12 and 16. If that same team has a softer spot later, and this week has a second 70% side the public is ignoring, spend the second side. Future value is just inventory management. It is not a prophecy about injuries.

What strategy is not

It is not a published "lock" list. It is not fading a team because they "always lose in the rain" without a number. It is not treating last week's upset as a reason the next favourite is due to fail — that is the gambler's fallacy applied to a new game. Each Sunday is a new sample.

Let p be your win probability on a side, s the share of remaining entries on that side, and assume everyone else is on the obvious favourite or on noise. A full Kelly-style survivor model is a tree through 18 weeks; you do not need it to see the shape.

A one-week differentiation check: if the favourite loses, the fraction of the field that dies is about s_fav. Your expected share of a $1 pot, ignoring future weeks, moves with (your survival probability) × (1 / expected remaining field). Taking the crowded 80% side keeps your survival high and the remaining field high. Taking the thin 65% side lowers survival and, when the crowd's side dies, lowers the remaining field a lot.

Practical rules that do not require a spreadsheet

• Write a three-week plan, then rip it up when a quarterback is ruled out Saturday. The plan is a default, not a vow.
• Look at public pick percentages if the host publishes them. If 80% of a large pool is on one team, ask what you are buying by joining them.
• Do not save a team for week 18 in a 16-person pool. You will not be there, or the pool will already be over.
• Do not spend two elite sides in the first fortnight of a 500-person pool without a reason.
• Never pick a team you think is under 50% just to be different. Dead is dead. Contrarian only works when the side is still a real favourite.

If you want sportsbook technique — closing-line value, holds, sides versus totals — that is how to win at sports betting, not this page. Survivor is a combinatorial game against the other cards.

A survivor pool is already player-versus-player. The NFL games are the random number generator; the other entries are the opponents. Your edge, if you have one, is using inventory and crowd percentages better than they do. The games themselves still have huge variance. One blocked punt ends a perfect card.

PVPspinArena is the same structure without a football season. It runs three player-vs-player games in USDC or ETH on Base. Jackpot sets your win chance equal to your share of the pot, the clean version of "my entry versus their entries." Coinflip is a 50/50. Roulette is a 15-slot wheel, 7 Purple and 7 Silver at 2x, 1 Green at 14x, every bet returning 14/15 (6.67% edge). Seeds are committed; check a settled round on fairness.

If the next week's "must-use" pick is already a chase to get even from last Sunday, stop. Tools are on responsible gambling. Play is 18+ only.

In the same cluster, see also quick pick, pull tabs, and lucky number.