ATROPOSWhat a survivor pool optimizer should optimize
Every survivor optimizer solves a constrained problem: one team per entry per season, doubles where the contest demands them, eighteen-plus weeks of picks. The difference between tools is the objective function — and most of the category maximizes the wrong one.
Survival probability is the wrong objective
Maximizing the chance you are alive in January sounds right and prices wrong. The pool pays a shared pot: surviving with the herd means splitting with the herd. In 1,000,000 simulated seasons of the $21M Splash World Championship, the pure survival-maximizing strategy returned 0.36x its entry fee — it lived the longest and finished alongside ~47 co-survivors. Routes chosen on split-adjusted EV returned 1.76x, five times more, by trading a little survival for a lot of separation.
The optimization is only as good as the field model
Half of every pick's value is who else is on it. An optimizer working from win probabilities alone is blind to the herding that makes an 80% favorite worth very different amounts at 20% versus 80% ownership, and to the late-season convergence where collapsed inventories force the surviving field onto the same teams. Atropos simulates the contest jointly — every entrant, every week, both endings — across millions of Monte Carlo season simulations, so ownership, availability, and the wipeout ending are priced into every route, not bolted on after.
Optimize the season, not the week
Greedy weekly optimization spends the best teams early and arrives at the double-pick gauntlet holding leftovers. Full-season route planning nearly triples survival odds versus week-at-a-time picking — the planning horizon is itself an optimizer decision. Multi-entry portfolios add a second layer: routes striped across distinct paths instead of near-duplicate entries that survive together and split together.
FAQ
What is a survivor pool optimizer?
A tool that chooses your survivor picks — this week's and usually the whole season's route — by maximizing an objective under the pool's constraints (each team once, doubles where required). The critical question is the objective: an optimizer that maximizes survival probability produces a different, and measurably worse, set of picks than one that maximizes expected prize money.
What should a survivor pool optimizer maximize?
Expected share of the prize pool — split-adjusted EV — not survival probability. In 1,000,000 simulated seasons of a $21M contest, the survival-maximizing strategy returned 0.36x the entry fee because it survived alongside the herd and split with it. Leverage strategies chosen on split-adjusted EV returned 1.76x with only slightly lower survival odds.
What is the difference between a survivor pool optimizer and a solver?
An optimizer picks the best route for you against a fixed picture of the world. A solver simulates the whole contest jointly — every opponent's entry, week by week, through both endings — and prices your choices against how the field actually behaves: herding, hoarding, and forced late-season convergence. The optimization is only as good as the field model underneath it.
Do survivor pool optimizers work?
Optimization over devigged win probabilities beats manual picking, but the gains concentrate in two places most tools skip: modeling the field (ownership is half of every pick's value) and planning the full season (full-season planning nearly triples survival versus greedy weekly picking, because late-season supply crunches are decided months early).
Can an optimizer handle multiple survivor entries?
It has to treat them as a portfolio, not as N independent answers. Identical entries survive together and split together, so a multi-entry optimizer diversifies routes deliberately — striping entries across distinct paths measurably outperforms rebalancing picks inside a single week.
The survivor pool calculator: what to compute → · Split-adjusted EV · Glossary