The Playoff Odds, Now That We’re There
The postseason begins tomorrow, and we have our 10 contenders for the crown: Anaheim, Detroit, Baltimore, Oakland, and Kansas City in the AL, and Los Angeles, St. Louis, Washington, Pittsburgh, and San Francisco in the NL.
Two of these teams will be eliminated by the end of the day on Wednesday. Let’s take a look at the early odds for the Wild Card games, based on our depth charts model and the season-to-date stats model.
The depth chart results are based on the ZIPS/Steamer forecasts, so multiple years of data are being taken into account, rather than just what a team did in 2014. Based on those calculations, the model sees the A’s as 53% favorites to win the AL Wild Card game, even though the game is in Kansas City. While the Royals finished with the better record, the forecasts believe the A’s are the better team, and the gap is large enough to overcome the Royals home field advantage.
The gap gets even bigger in the season to date stats mode. That model gives the A’s a 58% chance to win, as rather than using projected data, it’s based on an algorithm that puts significant weight on a team’s run differential, and the A’s finished with the best run differential in baseball. Of course, run differential can include the effects of timing as well, and the A’s don’t slightly lower by BaseRuns, where they finished with the third best mark in baseball, behind the Angels and Nationals.
The A’s/Royals match-up is interesting for a number of reasons, and the difference in performance for wins versus expected wins is one of the more notable story lines. The A’s underperformed their BaseRuns expected total by seven wins; the Royals overperformed their expectation by eight wins. The Wild Card game matches up the team with the largest positive differential against the team with the second largest negative differential, and all of our playoff models are going to put more emphasis on the underlying performances rather than win-loss totals. Thus, the A’s grade out as slight favorites on Tuesday, even given their recent struggles and the fact that they are on the road.
Over in the NL, the Pirates are 52/48 favorites by both models. Given that they are probably a slightly better team than SF (more wins, better run differential, better BaseRuns expected record) and have home field advantage, this seems like a small margin, but the Bumgarner/Volquez even things up a bit. The Giants have a clear advantage with the first guy on the mound, and the game might very well be decided by how quickly Clint Hurdle gets into his bullpen.
But realistically, both of these games are too close to call. 52/48 or 53/47 isn’t really much of an advantage, and all four teams have a very good shot at advancing to the division series. So let’s look at the overall odds of each team winning the World Series, as of now. First, the depth chart forecasts.
| Team | WS |
|---|---|
| Nationals | 20% |
| Tigers | 19% |
| Dodgers | 14% |
| Angels | 11% |
| Cardinals | 9% |
| Athletics | 8% |
| Orioles | 8% |
| Royals | 5% |
| Pirates | 4% |
| Giants | 3% |
The forecasts give a slight edge to Washington as the favorites, with the Tigers coming in just behind them as the most likely champion out of the AL. The Angels certainly have built a strong case for themselves as well, and the Dodgers are expected to give the Nationals their best run for their money in the NL. Not surprisingly, the wild card teams mostly have the worst odds, except for the A’s, who the forecasts still like a lot despite their second half troubles.
Don’t like forecasts, and prefer to use the season-to-date numbers? Then things change quite a bit.
| Team | WS |
|---|---|
| Nationals | 18% |
| Orioles | 17% |
| Angels | 16% |
| Dodgers | 14% |
| Athletics | 9% |
| Tigers | 8% |
| Cardinals | 6% |
| Pirates | 5% |
| Giants | 4% |
| Royals | 4% |
Washington is still the favorite, but the Orioles jump to the top of the AL heap, and the Tigers fall down below three other AL teams. The Angels and Dodgers both get small boosts, but Baltimore gets by far the largest bump by using 2014 numbers only. If you think teams should be evaluated solely by what they did this year, then this is the playoff odds model for you, though I think we know more about players like Justin Verlander and Steve Pearce than just what their current numbers suggest, and think the depth chart model is probably a better predictor of future performance.
The interesting thing about the 2014 season is that we don’t really have juggernauts. The Nationals are the closest thing we have to a consensus favorite, but in both models, they’re only slight favorites over the #2 team, and it certainly won’t be any huge shock if they don’t win it all. While the Wild Card teams certainly have the hardest roads to a championship parade, each of these 10 teams have a real shot at this thing.
Dave is the Managing Editor of FanGraphs.
Queue the Orioles fans crying about lack of respect…
Cue even…
Cue the queue
Curly queue?
Que?
Yeah fans that complain are the worst! We should talk about them more….
sure, only the Nats deserve respect for their season. gotcha.
You’re right. This is all BS. I think the only fair way to do this is to give each team in the division series at least a 20% chance on winning the world series and then the wild card teams about 10%.
Even the Astros and the Rangers won a few games against good teams this year. The SSS of the playoffs means these alleged odds are about as predictive as a ouija board.
“Alleged” odds…sick burn bro/ma’am. You sure put these forecasting models in their places.
If you’re gonna insult people, maybe work on the spelling.
Queue can work in this context. As in, ‘line them up to complain’, as opposed to ‘hear them complain in 3, 2, 1…’
What are the odds that that’s what he meant?
I calculated them to be at 14.7154282190816690635110471% but there may be some rounding error in there.
SSS, it’s hard to say
I can’t help but feel that throwing volquez in the WC game is going to cost them, especially against bumgarner. But hey, anything can happen in one game.
I guess you haven’t seen Volquez pitch much at all this season, especially in September when he’s put up the line of:
33.1 IP, 1.08 ERA, 13 BB, 31 K
He’s been the Pirates’ best pitcher for the past two months and has had a resurgent season.
I’ve watched him. He’s pitched well. His control’s been much better over the last month. I’d still rather have Cole for an elimination game.
Cole just pitched on Sunday…
And a 3.75 xFIP in that span.
It’s less about Volquez becoming an amazing pitcher recently and more about a .229 BABIP and a 3.6% HR/FB. These things aren’t sustainable.
It’s fair to say he has been better lately, but he is far from the Pirates’ best pitcher, lately or otherwise.
While those numbers are not sustainable long-term and it could end at any time, it’s also important to remember that we are not looking at the long-term right now. If he can keep pitching better than his peripherals – keep throwing strikes, keep the ball in the park, benefit from very good defense – then in a one game playoff, he’s just as likely to pitch well and beat Bumgarner as he is to pitch poorly as lose.
That’s the beauty (and the short-comings) of the one game Wild Card round in baseball.
Well we just had 2 articles on FG last week discussing the Bucs ability to outperform FIP. It is kind of silly to dismiss 2 months of consecutive starts and conclude that the next start will certainly regress. Whatever EV is doing it sure has been working for quite a while, lucky or not.
Phoenix–I agree. I think Hurdle should have held Cole for the WC game. The odds weren’t good enough for them to waste Cole yesterday. Too many things had to happen: beating Cueto, who was going for 20 at home; Wainwright losing to AZ; beating STL in STL on Monday with Locke. The better play would have been stockpiling pitching for the WC game. It might work out, but I don’t get Hurdle’s process here…
So I just ran the numbers on the pirates using 50% wc, 45% ds, 45%cs, and 40% ws and came up with roughly the same thing. Going into the season I would have probably given the pirates about 25-1 odds of winning the series so I’m guessing my earlier intuition was off about their actual odds because I would have thought they would have gone up after actually making the playoffs.
Because your estimate included a chance to win the division which would of roughly doubled there chances tow in the World Series at this same stage
I’m going to go with the Pirates and Royals simply because of the bullpens.
This is what I have been saying for years.
If there’s one guy who can be counted on for intelligent bullpen usage, it’s Ned Yost.
I can see it now: Aaron Crow loses the WC game with two outs in the 6th. Yost “frustrated” that they couldn’t hold lead into 7th for Herrera.
Because you put in relievers according to what inning it is!
OK, I’ll pick the Nats & Angels since they only have to beat two opponents each and have home field advantage both times.
How much are we betting?
He was probably talking about only the wild card game…
Ah right, my bad. Probably.
Jon Lester is pretty much the worst possible matchup for the Royals. Just based on the fact that he is starting for the A’s gives the Royals about a 15% chance of winning.
Perhaps a little hyperbole there. The Royals would have a (much) better chance than 15% facing Clayton Kershaw…
Why is Lester a particularly bad matchup for the Royals? They hit a little better against lefties this season, and they had the fewest batter strikeouts in the league (frankly I’m not 100% sure if that is good or bad where this matchup is concerned, but I’m inclined to think it’s good for the Royals).
It’s because Lester has dominated the Royals for years (and is something like 3-0 with a 1.25 ERA and 20 Ks, 5 BBs, in 20.2 IP this year)
In a so-small-its-virtually-meaningless three game sample, Lester did the following versus KC this year:
6.2 IP, 9 H, 3 R, 1 BB, 3 K
6 IP, 6 H, 3 R, 2 BB, 9 K
8 IP, 4 H, 0 R, 2 BB, 8 K
Totals – 20.2 IP, 19 H, 6 R, 5 BB, 20 K. So in three starts he was OK, OK, and excellent. In total? Pretty good. Which is a fair summation of John Lester in general.
What does all that tell us for a 1-game, win-or-go-home scenario? Very little, I would venture.
It’s not just this year, you know; Lester has been around since 2006. And with a career ERA against of 1.84, has always absolutely owned the Royals
Because those 2006 Royals looked a lot like this 2014 version? And he had the same teammates in both years, so it’s a perfect match from his supporting cast as well.
I will buy an argument that past results against the Royals may also have some validity. But those from 2006, I would weight at 0.000000000. (And probably 2007, and 2008, and…)
I’m guessing that if the A’s win their wild card game, their odds the rest of the way would approximately double, putting them in the top 4 or 5 teams?
That would put them at the top of both lists. There’s no way that happens.
Yeah, their chances would improve by a factor of 1.72 = 1/(.58). This would bring the A’s odds up to between 18% and 20% (depending on what their prior probability was before rounding.) The Nats, Tigers, and A’s would be considered co-favorites.
This makes more sense, probably, when you consider that Oakland was a 99-win team, tops in the league, by Pythagorean W-L. There’s reasons to think that they’re not a 99-win team right now, but the projections don’t know about most of those reasons.
I am sure that Oakland would like to use their pre-Lester trade lineup that generated all of those runs; the one that was healthy and included Cespedes. Even a dominating starter needs runs. Whoever the Royals put on the mound Tuesday is going to look like vintage Sandy Kofax.
“I am sure that Oakland would like to use their pre-Lester trade lineup that generated all of those runs; the one that was healthy and included Cespedes.”
This is obviously true, but unless Yoenis Cespedes went to medical school in Cuba, I’m pretty sure the Lester trade and injuries have nothing to do with each other. And probably 80% of the Oakland decline in offense is due to the latter.
My point is the team that created the awesome run differential is not the one that made the playoffs.
“Given that they are probably a slightly better team than SF (more wins, better run differential, better BaseRuns expected record)”
Pirates: 88-74 +51 runs
Giants: 88-74 +51 runs
He did say slightly.
Part of his point was that… more wins… yeah, that didn’t happen.
You keep using that word. I do not think it means what you think it means…
Pirates BaseRuns run differential: +89
Giants BaseRuns run differential: +41
I think that is what he was referring to.
But if that is what he is referring to, why would he word it like “(more wins, better run differential, better BaseRuns expected record)”? It should just say “better BaseRuns expected record”. Keep it simple.
The Orioles having the same odds as the A’s in the first model seems absurd since the A’s have an extra round to get through. In the simplest model a team’s chance of winning the WS is 0.5^a, where a is the number of playoff rounds a team plays. Thus, in simplest model, the A’s would have a 6.25% chance and the Orioles would have a 12.5% chance. You’d have to believe that the A’s are a far superior team to the Orioles to think they have equal odds despite “coin flip” to get through. That coin had better be heavily weighted.
…should read “despite the extra “coin flip” to get through”.
Well, yes, that’s clearly what the model believes. It’s not so implausible. 55%^4 ? 45%^3
Well of course the model believes that. That was the result. I do think it is pretty implausible in real life unless the A’s are the best team in the playoffs (favoured no matter who the opponent) and the Orioles are the worst team (underdog no matter who the opponent). In the real world, I would put the A’s as underdogs against Angels, Dodgers, Nationals, Tigers, Orioles.
Yup, that math works out Aaron. Though it’s very hard to believe that if the models only think OAK has about 55% chance with their ace to beat the Royals in KC that they’ll also have a 55% chance with their ace only getting one start to beat LAA (also on road) and then again against DET (also on road). I can pretty much guarantee they’ll be underdogs in both of those series, or at worst even money, and we all know that Vegas is pretty shrewd with this stuff.
Yeah, the betting market has the A’s about 13/1 – barely better than chance from here on in: http://www.betfair.com/exchange/baseball/market?id=1.111930893
If you do some rough math backwards, it appears as if that model really, really likes the Tigers over the Orioles in the first round — something like 75%-25%. And 25% is the A’s “coin flip” chance of advancing to the ALCS, so the A’s and O’s being equal makes some sense. Whether or not the Tigers should get so much love (or the Orioles so little) is, I suppose, a big question.
As an O’s fan I’m not at all disturbed by these projections. We have a an annoying tendency to beat projections and it keeps things exciting. Anything can happen in the playoffs. No matter what team wins the WS, they will have beaten the odds and defied expectations.
I am thrilled with the O’s season and I was happy to see we won thirty more than we lost and captured the Division title for the first time in 17 years.
Very well said! I’m an O’s fan too and I have no idea how they managed to get this far but it’s been a hell of a trip, and who knows? Maybe Wei-Yin Chen will be WSMVP! 🙂
Ubaldo Jimenez will be the ALCS MVP and Chris Davis will be the WS MVP. You heard it here.
FanGraphs has been publishing game odds all season. Are we going to see a report on how accurate they were?
Do you think anyone was tracking that? Feel free to go back and check archives for each date if available and get back to us.
I’d certainly hope someone was tracking it, yes, or else how would Fangraphs know whether the predictions were accurate?
The in-season odds are a moving target though, as players get injured or traded, and rookie call-ups happen more and more throughout the second half. It might be interesting to look at how accurate the odds were in April, vs. May, vs. June, etc…, as well as what events during the season had the biggest impact driving the odds up or down.
Tracking the game odds?! For every game, all season? Who has the time or inclination for that?!
I believe you will, yes.
Weird things happen in the playoffs. Like Cody Ross or Delmon Young.
… or Kirk Gibson… Ryan Zim’man gets my pick as Kirk Gibson of this off season.
How is this year like 2010, and 2012?
It’s even?
Same first three digits even!
FWIW,5 dimes and most bookmakers have Shields -112 and Lester +102.People accepting bets around 5,000$ are not chumps.My thought is they have many guys like fangraphs brightest, and I love fangraphs.World Series odds are Dodgers+460,Det +500 Angels +500 Nats +525 Balt +630 Cards +800 Oak+1250 Giants +1300 Pitts +1500 KC +2000.
series ods
“People accepting bets around 5,000$ are not chumps”
You said it. I will pay you next week, I swear. I’m just a little short right now.
Hm.
I think I like the Nationals at +525, actually.
If I was going to bet, that’d be my pick, in terms of my odds to win it vs. the odds they’re giving. Best value. Cardinals at +800 seem like a good value, too.
but do those models account for the cardinal’s pact with the devil? as a cardinals fan i have to say that i didn’t sell my soul for nothing.
It looks like our deal with the devil didn’t make it to the person who drew up the odds for the second list in the main article – where the Cardinals are 3% less likely to go all the way than a team that has to go through the Wild Card.
One of the best things about this game is there are so many variables, and at the end of the day, in a small sample sized window…it could come down to a dice roll.
or a good performance. or a managerial decision. or mental error. or a goofy interpretation of a replay. a rowdy home crowd rattling a pitchers nerves. or just straight up luck.
is it a cop out to say I know enough to not know?
Only if you hate Socrates
If you replayed these games enough times, say 100 (1,000?), the odds would probably be pretty close. Selecting a subset of 1, 5, or 7 games out of the 100 and predicting that’s what is going to happen is a lot different.
If you average the projections from the two models you get:
Nationals 19%
Dodgers 14%
Angels 13.5%
Tigers 13.5%
Dodgers / Tigers would be cool.
How ’bout them A’s?