How We Set Future Win Odds in Season Planner & Optimal Path Tools

Learn how we estimate future win odds for our Season Planner & Optimal Path tools in NFL Survivor and the blended ratings behind them.

by Jason Lisk - Oct 13, 2025

Los Angeles Rams head coach Sean McVay looks on in the first half during the game between the Los Angeles Rams and Philadelphia Eagles on September 21, 2025 at Lincoln Financial Field in Philadelphia, PA.

(Photo by Kyle Ross/Icon Sportswire)

We’ve received questions about how the future win odds shown in the Season Planner Tool (as well as the Season View of the Data Grid) and Optimal Path tools are created. This article explains what those numbers represent and what they don’t.

What These Odds Are Not

The weekly “PG Odds” you see in the Data Grid are model outputs for current games only. They rely heavily on recent stats (rate and volume) and can shift week to week based on performance.

For future games, we can’t generate “PG Odds” because we don’t yet know the exact profiles of teams or opponents. Instead, the win odds you see in Season Planner and Optimal Path are projections of where we expect the betting market to set spreads and moneylines.

Think of it this way:

  • PG Odds = a model output for one game, at one point in time, and is only determined once the week of that game arrives.

  • Future Win Odds = an estimate of where the market is likely to be in upcoming weeks.

How We Build Future Win Odds

Our numbers are a blend of:

  • Our base predictive power ratings (TR), which you can see here

  • inPredict’s “Generic Points Favored” ratings.

  • Adjustments to keep projections close to current and next week’s betting lines.

The weighting varies by team and situation. Sometimes we lean on our ratings, sometimes on inPredict, and often both are already very close. We generally keep projections within a point of the market unless we believe injuries or unusual circumstances distort the line.

2024 Validation

We backtested this blended approach in 2024. On average, our method was slightly closer to eventual closing spreads from using inPredict alone:

  • Across all weeks, our numbers were about 0.1 points closer.

  • Excluding games where all ratings agreed, our edge grew to 0.23 points closer to the eventual actual spread on average.

By Week 6, the average error vs. eventual spreads was about 3 points—reflecting the reality that injuries and team swings happen—but the blended approach still improved accuracy compared to either input on its own.

  • TR ratings performed well early but lagged later as injuries and unsustainable performances (e.g., turnover luck) distorted raw point-based models.

  • The blend, with manual adjustments, kept us closer to market lines.

When We Diverge From One of the Other Ratings

Some examples from 2024:

  • Carolina: Early blowouts and Bryce Young’s benching crushed their rating to –13.4. We judged that too extreme and shifted toward the market.

  • Miami: When Tua was out with a concussion, we manually lowered ratings, then restored them in the weeks of a projected return to play.

  • Washington: With a new QB/coach and unexpected performance, we leaned more on TR than market priors, which we felt were too slow to adjust to the dynamic and likely real impact of a star quarterback.

Adjustments Beyond Ratings

We also apply situational adjustments:

  • Resting starters: In 2024, we lowered the win odds for teams locked into playoff spots in Week 18. Similar adjustments are already in place for top contenders in 2025 in Week 18.

  • Injuries/suspensions: For example, our Kansas City rating was lower in early 2025 without WR Rashee Rice (suspended) and while WR Xavier Worthy was injured, but the rating used is higher later in the season when both are projected to be available.

  • QB injuries: We’ve already created separate near-term and long-term ratings for Cincinnati (Burrow), San Francisco (Purdy), and Baltimore (Lamar Jackson), among several situations that required accounting for different probabilities of having the primary QB available in different weeks.

Putting It All Together

Once ratings are finalized, we generate projected win odds for every team and every week. These are the odds shown in:

  • Season Planner & Data Grid (future view) → lets you sort matchups by week.

  • Optimal Path → powers the Max Survival path and alternatives based on which teams you’ve already used.