NFL research guide

Survivor Pool EV Explained: Why the Safest Pick Is Not the Whole Answer

By Richard Colby

You are reading the full article.

In a Survivor pool, expected value—usually shortened to EV—is an estimate of how much prize money an entry would receive on average across many possible endings. It asks a different question from “How likely is my team to win this week?”

You still need to survive. But the number of opponents who survive with you, the teams you keep for later, and the way the prize is paid all affect what your entry may eventually earn.

The Cold Line brings those decisions together in its free regular NFL Survivor planner and Circa Survivor planner. You can compare a path aimed at surviving with a path that also considers the field and possible prize outcomes.

Expected value is an average, not a payment

Imagine an entry with a 20% chance of receiving $500 and an 80% chance of receiving nothing. For this simple example, assume those are the only two possible payments.

Its expected value is $100: 20% multiplied by $500. That does not mean the entry will receive $100. In this example, the actual result is either $500 or zero.

It also does not mean someone will buy the entry for $100. A calculated estimate, a sale price, and money the contest already owes are different things.

Real Survivor contests can have many possible prize splits. A model estimates how likely each payment is, then combines them into an average. The answer depends on its assumptions about games, future picks, and contest rules.

A higher chance of cashing can still have lower EV

Consider two fictional paths. Path A has a 25% chance to receive $200 and otherwise pays nothing. Path B has a 10% chance to receive $1,000 and otherwise pays nothing.

Example path	Chance of the stated payout	Payout if it occurs	Expected value
A	25%	$200	$50
B	10%	$1,000	$100

Path A is more likely to pay something. Path B has the higher average payment because its possible prize is much larger.

This example is not a recommendation to take the riskier route. It explains why “most likely to cash” and “highest expected value” can point in different directions. A useful tool makes that tradeoff visible instead of hiding it inside one score.

Why other players’ picks matter

Suppose eight of ten remaining entries choose Team A and two choose Team B. The teams are in separate games, and this example considers only outright wins and losses.

If both teams win, all ten entries continue. If Team A loses and Team B wins, only the two Team B entries remain. If Team A wins and Team B loses, eight entries remain. If both lose, the pool’s all-lose rule decides what happens next.

The point is not that Team B must be the better pick. We have not said how likely either team is to win. The example shows how the same winning selection can leave you in very different positions depending on the other game.

Pick popularity matters because it changes who shares your outcome. It does not change the final score of the game.

An unpopular pick is not automatically valuable

A team with a much weaker chance of winning may offer too little upside to justify the added risk. Choosing it only because few other entries have picked it can be an expensive way to be different.

Compare the full picture: estimated win chance, expected pick share, remaining teams, and possible payout. Then ask what happens if the estimates are wrong.

The same applies to a popular favorite. Sometimes the strongest team is still the sensible choice, even when many rivals are on it. EV is a way to examine the tradeoff, not a rule that says to oppose the crowd every week.

Availability changes the crowd you are comparing with

In Circa, an entry cannot select a team it has already used. That means a strong favorite may attract fewer picks later in the season simply because many surviving entries no longer have it available.

TCL’s Circa forecast uses the actual entry records and used-team information behind those restrictions. It separates projected choices from the picks Circa has already published.

A private pool is different. Public pick percentages may provide a rough reference, but they are not confirmed choices from your office or family pool. Use the information available for your contest and label estimates as estimates.

Saving a team changes later opportunities

A weekly choice can improve today’s survival chance while leaving a difficult gap later. Another choice might be slightly weaker now but preserve a valuable team for a round with few alternatives.

That is where team future value fits into EV. The path matters because an entry must still reach the rounds where a saved team could help.

Pool size matters too. A small pool may finish before a carefully protected December matchup arrives. A much larger contest may require a longer route. Treating both as the same problem can put too much weight on a team you may never need.

The proxy the TCL crew uses

For football contests that require picks to be submitted in Nevada, The Cold Line recommends Vegas Football Proxy. They have been the proxy for The Cold Line crew for 10 years running.

Vegas Football Proxy lists support for Circa Survivor, Circa Million, and Westgate contests. TCL helps you research and plan; the proxy handles official submissions under your service arrangement. Check their current contest coverage, fees, and submission deadlines when arranging service.

Use TCL to compare the reasoning

TCL connects the current choice with the remaining path. For Circa, that includes the actual field, team availability, and holiday rounds. For regular pools, it includes the pool’s own settings and the pick-share information supplied to the plan.

The Survivor tools are completely free. You can explore a plan without logging in, while a free account lets you save picks and unlimited named builds. Separate builds make it easier to compare a safer route with an alternative and revisit the difference as the week changes.

A good EV explanation should leave you able to say why one route looks better under the assumptions. It should also make clear that the assumptions can change. No estimate controls an upset, and a model value is not a prize already won.