How MLB playoff odds are calculated

When a site says a team has a 68% chance to make the playoffs, it means something precise: if you played out the remainder of this season many thousands of times, that team qualified in about 68% of them.

It is not a ranking, a confidence level, or an opinion. It is a frequency, generated by simulation. And it accounts for things a bare games-back figure cannot — how many games remain, who they’re against, how the other contenders are positioned, and the fact that a three-game lead in September is worth far more than the same lead in May.

That last point is the whole reason the number exists. “Two games back” means completely different things at different times, and a single percentage collapses all of it into one honest figure.

The technique is called Monte Carlo simulation, and the principle is simple: when a problem is too tangled to solve with algebra, simulate it many times and count the outcomes.

For playoff odds it runs roughly like this:

1. Estimate each team’s true strength — usually from run differential, projections, roster quality, and record, rather than record alone.
2. Assign a win probability to every remaining game, adjusting for opponent, home field, and sometimes the scheduled starting pitcher.
3. Simulate the rest of the season once, deciding each game by a weighted coin flip. Apply the tiebreakers. Record who made the playoffs.
4. Repeat tens of thousands of times.
5. Count. A team that qualified in 68,000 of 100,000 seasons has 68% odds.

Everything sophisticated in a playoff-odds model lives in step 1 — how well it estimates true strength. Steps 2 through 5 are bookkeeping.

You could imagine computing playoff odds directly, but the structure of a pennant race defeats a clean formula.

Teams are interdependent: a rival’s loss helps you, and rivals play each other, so outcomes are correlated rather than independent. Tiebreakers are conditional rules that depend on head-to-head records. Multiple wild card slots mean a team competes against several others at once. And an unbalanced schedule means the remaining games aren’t equally hard for every contender.

Simulation handles all of that naturally, because it just plays the season and looks at what happened. That’s why essentially every public playoff-odds model — and the one inside TwentySeven — uses it.

A few things worth knowing before you trust a playoff percentage too far.

• Different models disagree, sometimes by a lot, because they estimate true strength differently. One weighting projections and another weighting record can differ by 15 points on the same team.
• Odds move fast late. In September each game carries far more weight, so a single result can swing the number several points. Early on, a whole week barely moves it.
• A low number is not zero. A 5% team makes the playoffs one season in twenty. That is genuinely rare but entirely possible, and it is exactly what makes late-season baseball worth watching.
• Injuries and trades lag. Most models take time to reflect a roster change.

The shape of the line over a season is often more informative than today’s value.

TwentySeven runs Monte Carlo playoff odds as part of its season view, next to standings, power rankings with heat trends, strength of schedule, and Pythagorean over- and under-performers.

The connection to the rest of the app is the point: playoff odds are the season-scale version of the same idea that drives win probability inside a single game. One asks “who wins this game?”, the other asks “who plays in October?” — both answered by counting how often it happens.