For many Mariners fans, a ritual has been added to their daily perusal of the MLB standings. Now they are living and dying over a relatively new wrinkle that’s available on the internet: Playoff odds.
Such an analytical tool was not around, or at least not prominently, the last time the Mariners made the postseason in 2001. I’m not even totally sure electricity had been invented yet. But now, three well-established sites put forth daily (or in some cases, multiple times daily) calculations of each team’s chances to make the playoffs.
Considering the focus that has been put on these odds down the stretch, I decided to delve into the nuts and bolts of the calculations behind them. They painted a rosy picture of the Mariners’ chances even before many fans allowed themselves to believe it could really happen.
And happen it will. With 20 games to play for the Mariners and a five-game AL wild-card lead on Baltimore (plus the tiebreaker), all three projection models Thursday had the Mariners with at least a 97 percent chance to end their 21-year playoff drought. Baseball Prospectus had them at 97.5 percent, Five-Thirty-Eight at >99 percent, and FanGraphs at 99.8 percent. If those were presented as the chances of rain, you’d be gobsmacked if it turned out to be a sunny day.
I called up Ben Clemens, a FanGraphs writer and an expert on its playoff odds model, to get the lowdown. I’ll get into the mechanics of the system in a second, but I asked him, bottom line, if the Mariners are in the playoffs. He wasn’t quite ready to make such a definitive statement, but he made it clear that it would take a miracle – otherwise known as an epic collapse – for it not to happen.
“The circumstances where the Mariners don’t make the playoffs, you’d basically need to, I don’t know, roll the dice 20 times and get snake eyes every time,” he said. “It’s at the point where it’s less about the strength of the team, and more about everything needs to go wrong. The wheels need to completely fall off. And that almost never happens.”
The guiding principle behind calculating playoff odds comes, as so much of baseball analytics does, from the mind of Bill James. Specifically, from his epiphany that you can extrapolate a team’s future success in wins and losses from its runs scored and surrendered. That formula is known as the Pythagorean Winning Percentage.
Each of those three aforementioned sites (plus Baseball Reference and ESPN, who also post playoff odds) have their own nuances but similar architecture.
“It’s funny, because they’re very complicated, and I don’t actually understand a lot of the math that goes into making our projections,” Clemens said. “But the way they work is actually really simple.”
The complicated mathematical part is the projections systems at the root of the odds formulations. FanGraphs uses Steamer and ZiPS, which come up with projections for every player in baseball. Those, in turn, are used to estimate how many runs a team is going to score and allow per game. And as James figured out decades ago, from that you can predict how often a team will win.
“So that gives us a winning percentage for every team,” Clemens said. “And then we basically just roll dice.”
It’s more complex than that, of course. Roster changes are updated constantly. (Baseball Prospectus is particularly good at factoring in injuries, something FanGraphs doesn’t do yet.) Then after every game played each day, starting on opening day, a computer has been programmed to simulate the remainder of the schedule, played out 20,000 times (a computation that is finished in under 10 seconds). Once the last game is played each night, the final calculations are done, and that’s what you see on the FanGraphs site when you wake up in the morning.
For example, when FanGraphs crunched the numbers Wednesday, they projected the Mariners to play at approximately a .590 winning percentage the rest of the season. The computer uses that as the basis of what Clemens describes as “a weighted coin flip.”
“If they’re playing a .500 team, they’ll win 59 percent of the time,” he explained. “If they’re playing a .400 team, they’ll win a lot more than that. … All we do is come up with a single estimate of how good we think the team is based on our projections of who’s on the roster. And that gives us a winning percentage. And we have them play out their schedule with that winning percentage. So it doesn’t actually have very many moving parts.”
Clemens believes the playoff odds are most useful as a fun tool and not something to be regarded as the definitive word.
“I wouldn’t take our odds to Vegas and bet on them,” he said. “I think you’d do fine. But I don’t think they’re some secret elixir that tells you a solution you’d never have otherwise seen.”
The vicissitudes of the system play out virtually every year. Last season, for instance, the Cardinals entered September with about a 3 percent playoff chance – and then won 17 games in a row to, indeed, make the playoffs. Such examples – and opposite cases of teams squandering what appeared to be sure things – are sprinkled throughout history (as fans of the ’95 Mariners and ’64 Phillies, among others, know all too well).
“That’s what 3 percent means; it means if you do something crazy, you can still make it right,” Clemens said.
Take the 2022 Mariners. When they fell 10 games under .500 on June 19, FanGraphs had their playoff odds at 5.3 percent. And here we are. But it took a 14-game winning streak, and a 51-23 record since June 19, to turn those odds on their ear. In other words, the Mariners did something crazy and made it right.
“You could argue that we were a little low on them, and I probably would,” Clemens said. “But if you’re saying, ‘Were they 5 percent to make the playoffs, or were they 30 percent to make the playoffs?’ I think it’s five and, yeah, they’ve just been awesome since then. Credit to them.”
One thing that has struck me, and Clemens as well, is that the odds projections remained bullish on the Mariners even as the Orioles were making a furious charge up the wild-card standings in the second half. The M’s exceeded 90 percent odds for the first time Aug. 31, and the number has been rising ever since.
“The thing it [the system] does best is it thinks about the schedule. A lot,” Clemens said. “That’s the thing I am worst at when I’m looking at the odds myself. … The model just can’t help it. It actually runs the games. And so it cares a ton about who they play and who their opponents play. The thing that helped [the Mariners] a bunch two weeks ago is that it thinks they have a very easy schedule. It still does.”
Conversely, the Orioles have a tough schedule. And furthermore, the player projection models that fuel the FanGraphs odds system aren’t very high on the Orioles’ talent. The bottom line is in the bottom line: 99.6 percent, a number that should leave Mariners fans with joy (and a teensy weensy bit of trepidation) in their hearts.
“I think the one thing that I always tell people when they ask about our playoff odds is that odds don’t mean that something can or can’t happen,” Clemens said. “And odds don’t tell you that this won’t happen, or that ‘You’re 3 percent to make the playoffs, so we think you’re bad.’
“Nothing like that at all. I think they just enhance it if you’re at 3 percent to make the playoffs and then make it. A great example is when the Mariners were 5 percent on June 19. And they make it. That doesn’t mean, ‘Why do we even bother with these protection systems?’ It means something awesome happened. It means that if we played out the season 100 times, and we’re even close to right, this was pretty unlikely, so savor it.
“I think the odds are at their best when they’re a storytelling tool. When they’re telling you you’re seeing something really cool happening right now.”