NFL · SURVIVOR POOL · 2026
The split is the game.
One team a week, no reuse, ties eliminate, and a modeled field of thousands chasing the same favorites. The win probabilities are the market's — we claim no edge over them. The product is the optimization: which favorite to spend, and when.
POOLa ~14,000-entry $100 survivor poolMODELED POT$1,260,000FIELD SIM10,000 worldsUPDATED3d ago
WEEK 5 · THE PICK
GB 95.6%
Green Bay Packers (vs Chicago Bears) — GB is the market's biggest Week 5 favorite (95.6% after the tie adjustment) and the pool-EV play — no trade-off this week.
PICK DEADLINE
Sun, Oct 11, 12:59 PM ET
WEEK 5
The pick, in numbers.
GREEN BAY PACKERS · vs CHI
95.6%
tie-adjusted win probability — the consensus moneyline, de-vigged, minus the 0.23% tie haircut
10.5%EST.
estimated share of the field on GB this week — modeled, not observed
WHY THIS SPEND
GB is the market's biggest Week 5 favorite (95.6% after the tie adjustment) and the pool-EV play — no trade-off this week.
Modeled EV of the full remaining path: $1,693.76 per $100 entryEST. · 0.75% to still be alive after Week 18.
THE ENDGAME
When the pool forces two picks.
After Week 13 the pool turns conditional: survive above a threshold and the next week demands TWO picks, both outright winners. These are the trigger rates from 10,000 simulated worlds of the field — estimated, and they move as the season reprices.
WEEK 14
77.7%EST.
chance of a double-pick week (triggers above 200 alive)
~340 alive enteringEST.
WEEK 15
65.0%EST.
chance of a double-pick week (triggers above 100 alive)
~142 alive enteringEST.
WEEK 16
43.1%EST.
chance of a double-pick week (triggers above 50 alive)
~53 alive enteringEST.
WEEK 17
33.6%EST.
chance of a double-pick week (triggers above 25 alive)
~23 alive enteringEST.
WEEK 18
46.9%EST.
chance of a double-pick week (triggers above 10 alive)
~11 alive enteringEST.
WEEKS REMAINING
The optimal path from here.
Weeks marked ×2 are planned as two-pick weeks in the worlds where the field sim says the trigger fires — the survival shown already weighs that probability. Current week is the consensus moneyline; future weeks are the season sim's priors and re-read as the market posts each week.
THE PORTFOLIO
5 entries, one sheet.
THE POT
$1,260,000
$100 × 14,000 entries × 0.9 after the rake — modeled
Last season 10 survivors split a $1.23M pot — ~$123K each from 13,652 entries. The split is the game.
Modeled winners: ~5.1 · expected split $205,085EST.
THE RECORD
Journaled, then graded.
THE METHOD
What is market, what is modeled.
Win probabilities are the market's, not ours: the current week is the de-vigged consensus moneyline (each market normalized proportionally, then averaged; source: our pre-game odds journal), and future weeks are the season sim's published priors. We claim no edge over these numbers — the product is the optimization built on top of them. A tie eliminates (picks must win outright), so every win probability is multiplied by (1 − 0.23%): 16 ties in 7,045 regular-season games since 1999. Field pick percentages and pool outcomes are MODELED ESTIMATES, not observed data: 10,000 seeded simulations of a ~14,000-entry pool where each modeled entry picks proportional to win probability^3 × a brand tier, with no team reuse. Double-pick weeks trigger per the pool's published thresholds (>200 alive entering Week 14, >100 alive entering Week 15, >50 alive entering Week 16, >25 alive entering Week 17, >10 alive entering Week 18), including the endgame split rules (no 2-0 in a double week → 1-1 entries split; all 0-2 → everyone who picked splits). Pool EV = survival probability × expected share of the $1,260,000 modeled pot ($100 × 14,000 entries × 90%), computed over the same simulated worlds so the correlation between our picks and the field's survival is priced in.
16 ties in 7,045 regular-season games since 1999 (0.23%). A tie eliminates, so every win probability below takes that haircut. Field popularity is estimated — win probability^3 × brand tier, no observed pick data. The pool is described generically — a ~14,000-entry $100 survivor pool — and no win probability on this page is an Arcline edge claim.
Survivor is a season-long allocation problem. The board is the map.
The win-total board carries the same sim by team — means, bands, division and playoff odds. Player props and the research desk live on the game-week stack.