Six Takeaways From Our Playoff Odds

Earlier this week, as is tradition, FanGraphs founder David Appelman went into his garage, turned off all the lights except for some candles, and performed a dark and arcane ritual. The words were carefully chosen and spoken precisely, with any variation promising disaster. Then he went back inside, pushed a few buttons on his computer, and now we have playoff odds for 2023!
Okay, fine, that isn’t exactly how it goes down, but it’s close. Our playoff odds bring together pieces of a lot of features you’ve already seen on the website. We start with a blended projection that incorporates ZiPS and Steamer’s rate statistic projections. We add in playing time projections from RosterResource, which incorporate health, skill, and team situation to create a unified guess for how each team will distribute their plate appearances and innings pitched.
With playing time and production in hand, we use BaseRuns to estimate how many runs each team will score and allow per game. That gives us a schedule-neutral win percentage for each team, because you can turn runs scored and runs allowed into a projection via the Pythagorean approximation. From there, we simulate the entire season 20,000 times, with an odds ratio and a random number generator determining the outcome of each game on the schedule, and voila! Our playoff odds.
Why am I telling you all of this? First, so you can look at them. They’re accessible from the main page, but you can also click here to dive in. Second, because I’m going to walk through some projections I found interesting, as well as a few places where the gap between common perception and our odds merit an explanation. I’ve come up with one takeaway per division, in order of how interested I am in my own takeaways. Let’s get started!
Oh My Goodness, the Dodgers
This is a code red. The Dodgers are projected for fewer than 90 wins. This is not a drill. Please proceed directly to the panic room. FanGraphs began calculating playoff odds and projected win totals in 2014. We’ve projected the Dodgers for 90-plus wins (or a pro-rated version thereof in 2020) every year. Until now, that is.
Why is that? In a word, pitching. We’re projecting the Dodgers to have average run prevention overall, which is a far cry from their recent form. In the past six years, they’ve allowed the fewest runs in baseball three times and the second-fewest three times. Our central projection has them allowing 150 more runs than they did in 2022. That’s, uh, a lot of runs.
We project them to come back to the pack for two reasons: they have worse pitchers and worse defenders. Cody Bellinger plays a mean center field, Trea Turner is a better shortstop than Gavin Lux, and the rotation has less margin for error (and an older Clayton Kershaw) than ever.
Subjectively, I think that projections are probably slightly too pessimistic about the Dodgers’ ability to keep runs off the board. They’ve run a preposterously low BABIP for years on end, likely thanks to an unparalleled ability to position their defense, and I think that will help them even with the new restrictions on the shift. They also have a good track record of working with pitchers, particularly relievers, to emphasize their strengths. Projections have a tough time nailing down some of the little things the organization does well.
Still, even if we’re a handful of wins low, the general point makes sense to me. These aren’t the 2020 Dodgers, who took one of the best teams in baseball and added Mookie Betts for fun. They aren’t the 2021-22 Dodgers, who poached Trea Turner and Max Scherzer to add to their embarrassment of riches. They’re starting Trayce Thompson and counting on a bounce back from Noah Syndergaard instead. They look like a very good team, but not quite the unstoppable force on offense and immovable object on defense that they’ve been for most of the last decade.
AL Central Grindfest Redux
The Guardians rode a hot finish to a comfortable division title in 2022, but the Twins and White Sox were neck-and-neck with them until each swooned in September while Cleveland prospered. This year, we think the division is in for more of the same.
The AL East has a reputation for being the toughest division in baseball, while the AL Central has a reputation for being the weakest. Our odds agree with those two takes – but also think that those two divisions have the narrowest gap between the worst and best teams. A scarce 13 projected wins separate the Tigers and Guardians in our projections. The NL East and NL West both boast first-last gaps more than twice that large.
Similarly, only four wins separate the Guardians and White Sox, who we give the third-best chance of winning the Central. Only the aforementioned AL East has a slimmer first-third gap. Add in the Royals and Tigers, who aren’t seriously contending but are more evenly matched with the best teams in the division than, say, the A’s and the cream of the AL West, and you end up with a muddled division race with plenty of capable teams but no juggernauts.
With so little separating the teams in our eyes, the division might go to the team that is least affected by injury, which I believe favors the Guardians. The White Sox and Twins have depth issues, though credit to the Twins for acquiring Michael A. Taylor to minimize theirs. Cleveland, meanwhile, has a 40-man roster chock full of potential contributors. This looks to be a close one yet again, with a marginal edge to the team with the deepest roster.
Pity the Phillies
The Braves have been putting together a mini-Dodgers run of their own in recent years, winning the NL East year after year while mixing homegrown stars with shrewd trade acquisitions. Our playoff odds have long liked what Atlanta is doing – we gave them the second-highest win total projection in the game last year – but this year we think they’re the best team in baseball, period.
The Mets aren’t far behind. They signed a string of high-profile free agents this winter, and they were already an excellent team – they won 101 games last year while their best pitcher missed most of the season. I’d take a full season of Justin Verlander over 11 games from Jacob deGrom, and Kodai Senga is a high-variance but high-ceiling addition that makes their rotation one of the best in baseball. They didn’t really lose anyone on the hitting side, either, so they’ll surely be a force to reckon with again.
That leaves the Phillies out in the cold. We give them the 10th-best projected win total, but only the 18th-best chance of winning their division. They made the World Series and then added Trea Turner, but it looks like their path to the playoffs this year will be as hard as ever. A playoff bye isn’t likely at their projected win range, but they have much less margin for error than they would have if the best teams in their division weren’t so danged good. Oh yeah, the Eagles lost the Super Bowl last weekend, too. Pour out some liquid cheese for the rough week our friends in Philly are having.
What’s Up With the Orioles?
The biggest disagreement between our playoff odds and the ZiPS odds is what to make of the Orioles. ZiPS has them as a roughly .500 team, while we think they’re in the 75-win range. That’s enough to take their playoff odds from questionable to doubtful, in NFL injury parlance, or from 28.7% to 9.9% if you want the math of it.
That has knock on effects for the rest of the AL East. The better the O’s are, the tougher everyone else’s road gets. The Rays look like a Wild Card contender with upside for more, just like they are seemingly every year. That’s much easier to do if the Red Sox and Orioles are bad, though. It’s not just that Tampa Bay has to play them 13 times apiece; they also need to finish ahead of at least one AL East team in the standings, and likely at least two, to secure a Wild Card spot.
To my mind, Baltimore’s error bands are the widest in baseball this year. They have a team made up of young up-and-comers, which means both a wider range of potential outcomes and less certain playing time. If they’re in the race in July, they might add at the trade deadline, or their prospects might develop at a completely different rate than we expect. That goes both ways, though; they’re still in talent accumulation mode, and didn’t hesitate to subtract at the 2022 deadline in pursuit of long-term success. Maybe we’re low on their chances (for what it’s worth, Baseball Prospectus agrees with us), but I think they have the least certain outlook in baseball.
If you’d like to look at a distribution of outcomes, you can do so using our Distribution mode, which shows the distribution of wins across all of our simulations. These simulations all start with the same talent level, so they won’t capture the exact type of volatility I’m talking about, but they will show you how much a team’s fortunes varied from simulated season to simulated season, even with the same players contesting each game.
Fair and Balanced Scheduling
Every year, preseason articles make too much of strength of schedule. It’s only natural – authors are grasping for stories without much new to talk about – but it simply doesn’t impact the outcomes of games very much. A 30-point difference in opponent strength of schedule is worth something like five expected wins, but that’s a huge schedule differential. That’s like the Orioles compared to the Guardians last year, before the advent of the more-balanced schedule.
Projected strength of schedule is always narrower than observed (projections have narrower bands than observed standings, since they’re mean outcomes), and the new schedule is pushing that gap even tighter. The Cardinals have the easiest projected schedule in baseball this year, and their opponents check in at a .491 winning percentage, quite close to .500. The toughest schedule belongs to the Orioles, who sport the lowest projected win total in the toughest division in baseball, resulting in a .509 projected strength of schedule.
In other words, don’t give the schedule too much credit when you’re looking at playoff odds. The Cardinals aren’t projected for 88 wins solely because their division doesn’t have any competition. Give them the toughest schedule in the game instead of the easiest one, and we’d still project them for 85.5 wins. The same goes for the Brewers; they’re getting a small tailwind thanks to their division, but the gap just isn’t worth that much. If the Brewers, Phillies, and Giants compete for the last playoff spot, our odds will tell you that the difference in their schedules is worth less than a win’s worth of tailwind for Milwaukee. Team matters more than schedule, in other words, even in the two weakest divisions.
The Angels Aim for the Middle
Every year, I make note of the fact that the way we construct our odds favors teams with strong starting lineups, regardless of the depth behind those lineups. We use a single initial projected team winning percentage to run every simulation, and that projected winning percentage mostly assumes full health. Sometimes that means that a team you’d expect to struggle if any stars miss time (yes, I mean the Angels) fares better on our odds than you’d expect them to in real life.
Not so in Anaheim (they might be the Los Angeles Angels, but they still play in Mickey Mouse’s shadow) this year. The ZiPS odds, which account for injuries by simulating unique seasons with varying playing time for each player, like the Angels more than our odds this year. The team has finally done enough adding around the edges, as Mike Petriello recently noted, that it’s no longer just Trout, Ohtani, and a pile of below replacement-level teammates.
Their reward? An AL West field that’s difficult and only getting more so. The Astros just won the World Series. The Mariners are ascendant, having broken their playoff drought behind Julio Rodríguez’s triumphant debut. The Rangers might have the best starting rotation in baseball. The A’s… okay, the A’s are bad, but the other three teams in the division are good. The middle class of the AL is better than ever. Last year, we projected eight AL teams to win 80 or more games. This year, it’s 10. The path to an AL West title or a Wild Card berth is littered with solid competitors.
Heck, last year the Angels and Astros were the only AL West teams with a preseason projection of 80 or more wins, and this year we think every team other than the A’s will clear that bar. The Angels front office is doing what they weren’t able to accomplish in recent years: surrounding the best players in the game with competent major leaguers. Unfortunately, everyone else is improving too, and the end result is 39% playoff odds. That doesn’t mean they’re going to miss the postseason, but it does mean they have a tough road ahead.
Ben is a writer at FanGraphs. He can be found on Bluesky @benclemens.
The distribution mode is super cool but I am having trouble figuring out how they compare to each other. Is there an easy way to get some measure of the dispersion on these that helps us understand what’s going on? Like the 80th and 20th percentiles of the distribution?
We list the 25/75th percentiles next to the graphs in the table, in the brackets. Curious if that’s not what you’re looking for.
It looks like the default is to show it “compact” which on phones is smart, but it also means I didn’t see it! So if anyone else has the same issue, click “full” and you will get them.
Update: the interquartile range for every single team has an either 8 or 9 wins, which makes me wonder a bit about the distribution of outcomes. I don’t think it’s wrong or anything but it does make me wonder what is happening there.
That’s because we fix talent level for all the simulations. Every team should by definition have similar variance.
I know the playoff odds don’t incorporate injuries, but is the variance in distribution mode attributable to playing time or just to inherent baseball randomness–sequencing, etc.?
This was my observation as well.
Kind of reduces the feature to “looks neat, but no real information” IMO.
that’s a lot of stuff around here, tbh
In terms of variance, we do inject random variance into the Season-To-Date Stats mode, but it just flattens the bell curve.
I guess it might be interesting to take the individual reliability scores for projections and apply them to teams, or we could randomly inject injuries and give bench players more playing time as the season went on.
Though, coding that gets fairly complicated and potentially slow as we rebuild depth charts for each injury, redo the baserun calculation, etc… There are probably some shortcuts that could be taken. Maybe next year. 🙂
Yeah, and I apologize. Rereading my comment, it wasn’t my intent to be obnoxious.
That the 25th and 75th percentiles are 4-5 wins away is an interesting and useful result. I was just trying to express that I’m just struggling to learn anything more than that from the graphs themselves.
It seems weird that the Rays have a similar distribution as the Angels who rely so heavily on front line health, and Orioles who have lots of unproven talented players.
The Orioles definitely are an interesting team here for sure. And it fits- often times young teams make a jump and then regress some the following year. Not just in baseball, but other sports as well.
They should have done more with their rotation in the offseason. Grayson could possibly be a top flight pitcher right away, and that would be huge, but even then Gibson and Irvin are back end guys while the rest of the rotation guys are guys that don’t miss bats and are due for regression from last year. Who knows how Means looks when he returns summertime? Hall can’t throw strikes and I just read on mlbtr that he’s banged up. I think they’re probably better than the Red Sox but that’s it in division.
totally agree. Very underwhelming offseason for them.
Where I think they get lucky- they were one of the lower shifting teams last year so the new shift rules with those guys who don’t miss bats won’t hurt them so much.
I don’t even think they’re better than the Redsox, who I see as an 85ish win team
I think both the Orioles and Redsox are high variance but the redsox seem to have a higher floor with their vets
Also, I think expectations are a big thing for middle of the pack teams. Orioles coming off a “everything went right year” and now have expectation to meet and pressure. Redsox came off a bad year and now have something to prove. Which I think is an often overlooked aspect when looking at mediocre teams
The only way I can see Boston being better is if Sale gives the Red Sox twenty plus starts of being close enough to his old self, and Paxton and others pitching and pitching well and I just don’t see it.
I mean, weird things happen, so it’s possible they get hot and luck into the last wild card spot, but I certainly don’t think it is likely. Losing Bogaerts hurts and losing Story for half the year just adds to the pain. It really limits the upside of this team.
Interesting also checking out the PECOTA projections – they are a lot less conservative, with four teams projected for 95+ wins from PECOTA and none here, as well as five teams projected for fewer than the 65 wins of the Rockies in the FG playoff odds (though three of those are within half a win of 65.)
PECOTA always underprojects the Braves and overprojects the Mets for some reason. They have not selected the Braves to finish first in the NL East in any of the last five seasons, seasons when the Braves have won the division every time.
I am very skeptical of their projections. Someone pointed out that they forecast the Braves with only the 11th most runs in baseball, 71 runs less than last year which seems preposterous to me.
Untrue. Last year they picked the Braves to win the division with 92 wins and the Mets to finish second with 89.
Of course, they didn’t cover themselves in glory in 2021, when they predicted Mets-Nationals-Phillies-Braves-Marlins.
Oh OK, they must have made multiple 2022 projections then. Their first 2022 projections, pre lockout, had the Braves with 82 wins and the Mets 90. Braves were projected for 3rd place.
https://www.baseballprospectus.com/news/article/72737/pecota-2022-the-highly-unofficial-lockout-standings/
Free agency wasn’t close to done, so why does that standings mean anything in the least? That’d be like doing standings in a normal offseason about Dec 10. Fun, but utterly meaningless.
Yet the Braves didn’t sign anyone notable in 2022 free agency, other than Jansen, and actually lost freeman, replacing him with Olson, so not sure why their projections went up so much a couple of months later. They did less than most teams did after that initial projection.
It’s interesting to look at this and PECOTA projections on baseballprospectus. There are actually some pretty big differences of opinions. For example, they like the Dodgers a lot more, don’t like the Mariners much, believe the AL East is basically the Yankees and a gap until Rays/Jays, and essentially flipped the Braves and the Mets forecast compared to Fangraphs.
I would probably take FG on all of those, except maybe the Braves and Mets flipped (maybe). I actually think the projections on the Mariners are pretty low; it’s probably the one projection I disagree with the most, but in the opposite direction.
The team with the largest payroll—nearly 50% more than the 2nd-place Yankees after adding projected tax penalties (per Spotrac)—checks in with a 32.3% chance to win their division. This would seem to argue against the need for a salary cap in baseball.
For this argument, I think you would not want to add in projected tax penalties (for either team), since their ostensible purpose is to limit how much teams are willing to shell out in actual salary.
With the caveat that I think all luxury taxes and similar cap-like features should be repealed, the Braves have been on an enormous heater the last 6+ years in terms of amateur talent working out, so the fact that the Mets have nearly closed the gap with a heavy dose of free agent spending doesn’t really suggest to me that payroll can’t make a decisive difference. Look at their team and lop off maybe 40-50% of the wins expected from Scherzer, Verlander, Diaz, Nimmo, and Senga and they’re not really in the hunt for the division.
I didn’t say that payroll doesn’t make a difference, and I didn’t say that a single data point clinches the argument. I said it argues against it. Which I think it does. To make a more complete case you’d obviously have to get into more of the specifics, which you start to do here.
The only argument in favor of having a salary cap is if you want to take money out of the players’ pockets and put it into the owners’.
I hear this argument- but one counter and I am not sure the result- what if along with the cap there was an aggressive floor. A high salary floor I believe would somewhat hinder the market for the top tier free agents but I think would have the effect of substantially increasing the market for mid-tier guys. Cheap teams who avoid the 1.5-2 war veterans would be “forced” to spend that cash. I know the pro-labor argument for no salary cap, and I am sensitive to that, but the differences are getting massive and sure a few really smart front offices can make up that gap but I would be curious what a cap with a high floor would do.
Also could have a nice knock-on effect of driving cheap owners of middle market teams to sell their clubs.
In each league, the six teams with the best playoff odds right now are the six playoff teams from last year. And kind of like last year, the Twins and Brewers are the biggest threats to crack into the picture. So my biggest takeaway is the projection of no, or almost no, turnover even after a truly wild offseason.
Rangers might have the strongest rotation in baseball? I mean they have the best pitcher when he is healthy, and all the other guys are ok, not bad, but I’d take half a dozen teams over them if not more.
Depth charts has them second behind the Yankees. But more than half of the other positions are in the bottom half of the league, with LF dead last and RP fourth worst. Seems reasonable as a .500 team.
I agree. Would be shocked if Rangers starters finish anywhere near best in game. deGrom is amazing but I think you have to expect more like 15-20 starts from him and the rest of the rotation…Martin Perez, Gray, Eovaldi, and Heaney is all…fine but also kind of meh. The entire rotation made about 100 starts last year so Dunning and Odorizzi are likely to start a fair bit and they are even more meh. I’d much rather take the rotation of a team like the Mets (2 studs at top and then fine after) or Braves (just very good starting). Of course, if deGrom starts 30 games, their rotation looks much better.
A healthy Jake de Grom plus four league-average-ish guys is absolutely a contender for the best rotation in baseball (ask any Mets fan). As to what are the right odds to put on his health, none of us and no public-facing baseball research is even close to knowing the answer.
If they really get 175 or more innings out of deGrom, they might.
I’m definitely taking the Yankees, Mets, and Brewers, Braves, and Phillies over the Rangers, and possibly others too. I do not buy that deGrom is pitching 172 innings. And that’s important because in the moments where he is out of the picture and Jake Odorizzi takes over they go from a rotation that is Top 5 to one that is Top 20. So yeah, where they fall in between 5 and 20 depends a whole lot on how much you think deGrom plays.
I would generally expect with as many good teams in the NL, that the favorite from the AL would have the highest WS odds based on higher odds of getting there. I am surprised with how good the odds are for one of the Braves and Padres to being the NL pennant winner versus the Yankees and Astros in the AL.
I think its mostly just getting a bye
The Braves and Padres have better 2nd best teams in their divisions than the Yankees and Astros. I’m surprised the Astros don’t have the highest odds for a bye by a decent margin with the biggest gap between them and the 2nd best team in the non-central divisions.
The Orioles’ problem is their starting pitching. FG depth chart (correctly IMHO) has their starters contributing 7.5 WAR, 2nd worst in MLB. It’s hard to reach .500 with that kind of rotation.
And most of that is Dean Kremer, so.
“The biggest disagreement between our playoff odds and the ZiPS odds is what to make of the Orioles. ZiPS has them as a roughly .500 team, while we think they’re in the 75-win range….. That has knock on effects for the rest of the AL East. The better the O’s are, the tougher everyone else’s road gets.”
With the more balanced schedule, wouldn’t this effect be spread across the entire American League? Everyone is playing them an equal amount and they are all vying for the same wild-card spots, right?
It’s not actually an equal amount despite the ‘balanced schedule’ nomenclature. You play divisional opponents 13 times each and the other 10 teams in your league 6-7 times each.
Some of the IP for starters on the FG depth charts seem wildly optimistic. DeGrom is an obvious example (the whole Texas staff seems to be high), but there are may others as well.
All these comments and analyses, and yet, at this point, not one of you (this is FANgraphs, after all) has a problem with this being about PERCENTAGES and not ODDS. You know, like the title says, this is about ODDS. Am I way too far out in the correct (er, um, “right” – get it?????) field here??? All these variances and TLA/FLA about SIDGAFA, and not one person is suggesting that this is about something that is not giving us odds? Someone please (PLEASE) help me understand your understanding of this “%” is representing this “:”. And FG wonders why I do not subscribe. There is a difference, by definition.
Thank you kindly, all.
Odds can be easily expressed as a percentage? I don’t think I understand your diatribe.
I believe it translates roughly to “I do not care to divide 100 into various numbers, can someone please do it for me”
The information being conveyed is exactly the same, the only difference is presentational. Being pedantic is an EW thing, not a FG thing.
The odds are what you show when you can’t lie but you don’t want people to easily figure out what the actual likelihood is. So I’m cool with percentages.
I’m surprised the Yankees are favorites over the Astros to win WS even though their odds of winning the division are lower. I guess it’s what happens when one team loses Justin Verlander and other gains Carlos Rodon, but I still didn’t expect it given how much better the Astros looked at the end of 2022.
Rays seem underrated also. They have 9 position players who project as above average at their position. Injuries shouldn’t bother them as much as others. Top 3 starters projected ERA are 3.06, 3.24 and 3.58. This seems like a recipe to use depth to make it into the playoffs and then lean on the starters more heavily.
I wish Baltimore had signed a big free agent.
It’s because of the division. The Athletics are a likely 60 win team, and the Rangers are still probably a mid 80 win team at best, while the Angels and Mariners aren’t as good as Toronto/Tampa. Even with improved rosters of the Mariners/Rangers/Angels, there’s still no real threat to the Astros in that division. Could the Astros lose the division? Well yeah. But it’s not likely.
I think the rays bounce back from their “down” year in 2022 and win 95+ games this season and possibly win the AL East. They had a ton of injuries last season and their starting rotation looks absolutely nasty
The FG projections for the Dodgers are an outlier compared to BP as well as the betting odds. Vegas odds have Dodgers at 7.5:1 to win the World Series compared to 20:1 for FG. Vegas predicts Dodgers NL west win probability at 58% compared to 24% for FG. Similarly, the 2022 projections for FG underestimated Dodger wins by 18 games. Differences this significant call into question the Fangraphs methodology.
yeah when are the dodgers gonna get a fair shake in the analytics community