Playoff odds are the most-quoted number in September and the least-explained. A site tells you your team has a 42% chance and shows you nothing about where that came from — which is a problem, because two simulators can hand the same team wildly different numbers depending on what they think a “good team” is.

Baseball NUT runs its own. Every team’s odds sit on the standings page, free, no account, updated through the day. Below are today’s numbers and — more usefully — the entire method, so you can decide how much to trust them.

As of August 30: 376 regular-season games remain, and each refresh replays all of them 5,000 times— roughly 1.9 million simulated games per update.

AL Playoff Odds

Kansas City, the Athletics, and the Angels are all below 0.1% and effectively out.

NL Playoff Odds

The NL race is essentially one open seat: Arizona and San Diego are splitting a single wild card at 58% and 43%, and nobody behind them is above 5%.

See the live odds on the standings page →

How the simulator works

Five steps, and none of them are a black box.

1. Rate each team by run differential, not record

Every team gets a “true” win percentage from the Pythagorean formula — runs scored raised to the 1.83 power, divided by that plus runs allowed to the same power. Run differential predicts future winning better than win-loss record does, because a team’s record includes how it happened to do in one-run games, and that is mostly noise.

2. Regress small samples toward .500

Early in a season, 12 games of run differential is not evidence. The rating is pulled toward .500 in proportion to how little has been played, with the training wheels fully off at 81 games. At this point in the calendar every team is past that, so today’s ratings are pure Pythagorean.

3. Play each remaining game with Log5

For a specific matchup, Bill James’ Log5 formula turns two win percentages into the probability that one beats the other, plus a 4-point bump for the home team. So a game is not a coin flip — it is weighted by who is actually playing, and a last-place team hosting a first-place team still wins its share.

4. Run the whole season 5,000 times

Each of the 376 remaining games is decided by a weighted coin flip, the final standings are assembled, and the six playoff teams in each league are picked — three division winners plus three wild cards. Ties are settled by coin flip, which is a fair stand-in for a Game 163. Do that 5,000 times and count how often each team is among the six.

5. Report the share, not a verdict

A 42% team made it in 2,100 of 5,000 seasons. That is the whole claim. It is not a prediction that they will miss — it is a statement that in a bit more than four seasons out of ten, they get in.

The simulation re-runs on a six-hour cycle, so the number moves as games finish rather than once a night.

Why the odds argue with the standings

The most useful thing about a run-differential-based sim is where it disagrees with the win column.

Houston is the clearest case. The Astros are 69-68 with a -43 run differential — they have been outscored by 43 runs and are still over .500. The simulator rates them as a below-average team, and yet gives them a 63.2% chance to win the AL West, because everyone else in that division is worse. Those two facts are not in conflict. A weak division is a real path to October, and it is entirely possible for a team to win one while being outscored on the year.

Detroit is the inverse. The Tigers have a +65 run differential — better than Houston, Cleveland, Texas, Toronto, and Baltimore, all of whom have higher odds. The sim likes Detroit’s underlying team and still gives them 5.7%, because at 63-73 they are out of calendar. Quality does not get you in if you have already spent the games.

That is the distinction worth internalizing: playoff odds measure the path, not the team. They fold together how good you are, how many games are left, and who you have to climb over. A low number is not an insult and a high one is not a compliment.

Where the odds and NUT disagree

Baseball NUT also grades every team by NUT Score, which sums the production of every hitter and pitcher on the roster — park- and league-adjusted, with nothing about schedule or standings in it. Sitting the two numbers next to each other is where it gets interesting, because they answer different questions: NUT asks how good has this roster been, the sim asks how likely are they to get in.

The Pittsburgh Pirates rank 4th in all of baseball at +8.3 NUT and have 1.2% playoff odds. Their players have produced like a contender; the wins landed elsewhere. The Miami Marlins are 10th at +2.5 NUT with 4.6% odds — same story, smaller gap.

The Arizona Diamondbacks are 27th at -7.3 NUT and hold the NL’s last wild card at 58.1%. The Toronto Blue Jays rank 30th — dead last at -8.9 NUT— and still carry 18.9% odds. Both are living on a soft remaining schedule and a division that has not put anyone away.

Neither number is wrong. If you want to know who is playing the best baseball, read NUT. If you want to know who will be playing in October, read the odds. The gap between them is usually the most honest thing on the page. For more on the roster side, see which contenders are real by NUT.

How to read a playoff odds number

Three rules that will save you a lot of arguing:

Big daily swings are normal in September. With 25 games left, one win is 4% of your remaining season. Odds that jump 8 points overnight are not a broken model; they are a small denominator.

Anything above 95% or below 5% is basically settled. The interesting range is the middle, and right now only six teams in baseball live there.

Division odds and playoff odds are different questions.Boston is 99.5% to make the playoffs and 0.8% to win the AL East. Those are consistent — they are getting in through the wild card, and Tampa Bay is not going anywhere.

Where to find it

Odds live on the standings page alongside each team’s record, streak, and NUT rank — free, no login, no ads. Tap any club for its team page, with the full hitting and pitching NUT split and the roster behind the number.

Following one race in particular? The trending board shows which teams are actually getting hotter over the last seven days — the leading indicator that moves these odds next.

The odds move every few hours from here to Game 162. They are at baseballnut.app/standings, and they will still be free in October.