So Why Do Our Playoff Odds Love the Royals?
This is the postseason of the underdog. The Angels, Dodgers, Tigers, and Nationals were all bounced in the first round. Both wild card teams advanced, combining to lose one game in the process, despite having burned their best starting pitchers in the play-in game. One of the remaining division winners won just 90 games. These are not the League Championship Series many people expected, and with the little guys advancing in each division series, we should be in for some pretty even match-ups. At least, that’s what one would think.
But if you look over at our Playoff Odds page, our depth chart forecasts don’t exactly see it that way. This is how those projections look right now, before the start of either LCS.
| Team | LCS Odds | WS Odds |
|---|---|---|
| Royals | 63% | 36% |
| Cardinals | 52% | 25% |
| Giants | 48% | 23% |
| Orioles | 37% | 16% |
Our projections have the Royals as a significant favorite over the Orioles, even though Baltimore was the better regular season team by just about any measure you want to use. But this isn’t another FanGraphs-just-hates-the-Orioles situation — we don’t, really, I promise — as our forecasts actually had the Royals-Angels match-up as essentially a coin toss, and see them as a legitimately strong contender, not just a Wild Card who snuck past the first round due to the randomness of October.
But the Royals certainly didn’t play like an elite team this summer. By BaseRuns, they were a .500 team, and only managed to snag a Wild Card spot because of their strong performances in the clutch. So what’s the deal? Why do our forecasts love the Royals so much?
Thanks to our Depth Charts overview page and our positional leaderboards we can actually go see exactly where the differences are between projected value and what the Royals produced in 2014. So let’s find out where exactly the projections are bullish on this roster.
| Royals | C | 1B | 2B | SS | 3B | LF | CF | RF | DH | SP | RP | Bat | Pit | WAR |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Projections | 4.4 | 2.6 | 2.2 | 2.2 | 3.4 | 4.4 | 3.5 | 2.4 | 1.6 | 10.3 | 5.0 | 26.7 | 15.3 | 42.0 |
| 2014 | 3.0 | 1.0 | 1.0 | 3.4 | 1.3 | 6.2 | 5.8 | 2.3 | (1.7) | 12.9 | 5.9 | 22.3 | 18.8 | 41.1 |
| Difference | 1.4 | 1.6 | 1.2 | (1.2) | 2.1 | (1.8) | (2.3) | 0.1 | 3.3 | (2.6) | (0.9) | 4.4 | (3.5) | 0.9 |
Overall, the forecasts are pretty optimistic about the Royals young position players, giving them league average or better marks at essentially every position on the field. But there are two notable forecasts that paint a significantly more positive view than just looking at 2014 performance: third base and designated hitter.
Let’s start at DH, where Billy Butler was a miserable failure, especially when he wasn’t playing the field. His overall .271/.323/.379 line is bad enough for a bat-only player, but even that was pulled up by solid production when Butler played first base; as a DH, Butler hit .259/.307/.335, good for just a 79 wRC+. The guys who filled in when he played first base weren’t a lot better, and overall, the Royals DH’s combined for the second worst total in the AL, with only the Mariners (-3.2 WAR!) getting less from the position.
But the Steamer forecasts — the engine powering our Playoff Odds models — aren’t really phased by Butler’s lousy 2014 season, and think he’s basically still the good-not-great hitter he’s always been. The 119 wRC+ forecast for him is actually slightly above his career average mark, as Steamer is still giving weight to his strong 2012 season, and at 28 years old, he’s right in the sweet spot of the aging curve. This season, Butler was awful, but the forecasts don’t see Butler as an actually awful player, and assuming that he’s classic Billy Butler and not the 2014 version gives the team a significant boost in the forecasts.
The story is similar at third base. Mike Moustaksas had a miserable regular season, posting a 76 wRC+ and getting himself optioned back to Triple-A for a stint, but Steamer sees him as an above average big league third baseman. In fact, his +3 WAR in 586 PA is shockingly strong given that, in nearly 2,000 plate appearances, Moustakas has produced a total of just +5 WAR over his career. Steamer is really bullish on Moustakas despite a poor Major League track record, and so to find out why, I emailed Jared Cross, the gatekeper of the projection system and asked him what was up. His response:
In addition to going into a peak age, I think he’s benefitting from having a slightly better year this year (than his career average) in terms of BB%, K% and a worse year in BABIP. BABIP not only gets regressed more than K% and BB%, but BABIP data from longer ago weighs in more heavily relative to data from the more recent season (although the most recent season still gets the highest weight, of course). So, his rough year in 2014 isn’t quite as bad as it looks, projection-wise, because it’s largely the result of a terrible BABIP.
Jared isn’t kidding; Moustakas had a .220 BABIP this year, the lowest mark of any hitter who hit at least 500 times this season. His high infield fly rate shows that this isn’t just bad luck, as Moustakas makes a ton of weak contact that results in easy outs for the infield. But Moustakas has always hit a ton of infield flies, and he’s never run a .220 BABIP before; his career mark is .260, and Steamer is only forecasting him for a few ticks above that, at .272.
But as Jared notes, if you don’t hold the entirety of his .220 BABIP against him, the rest of Moustakas’ line actually isn’t half bad. His walk rate was the highest of his career, and his strikeout rate was well below the league average, while he also posted a decent-ish .149 ISO. For comparison, Moustakas’ BB/K/ISO numbers are almost exactly the same as Jacoby Ellsbury’s, and actually a little bit ahead of guys like Lonnie Chisenhall, Starlin Castro, and Pablo Sandoval, each of whom were slightly better than league average hitters. This high-contact/some power combination, mixed in with a smattering of walks, is a decent offensive player as long as the BABIP is within the normal range.
And that’s basically what Steamer is projecting for Moustakas; a strong enough BABIP regression to make him a league average hitter, based on his solid enough underlying skills. Add in his defensive skills at third base, and Steamer sees Moustakas as a productive player, not the black hole he was in the Royals line-up for most of the year.
Interestingly enough, this is one of those times when the data and the scouts likely agree. The Royals believed themselves to be contenders this year based in part on their faith in Moustakas and Butler, and both underachieved relative to what the team and the forecasts believed they were capable of. The same could be true, to a lesser extent, of Eric Hosmer, Salvador Perez, and Omar Infante. The Royals expected to have a productive infield, and the forecasts thought this group should be pretty solid as well, but in reality, they were pretty lousy, especially if you consider Butler part of the infield group. But just as the Royals haven’t given up on their young core, neither have the projections, and their optimism about these young players performing better than their 2014 numbers has the forecasts buying into Kansas City as a legitimate contender.
A total projection of 42 WAR might not sound like a lot, because after all, it’s only 1 WAR higher than their 2014 total, but it’s actually the fourth highest projected total of any team in baseball, a tenth of a win behind the Dodgers. These forecasts look at the Royals and see a legitimately good team, not a .500 club that clutched their way into the playoffs.
If you go to the Royals team depth chart page, you can see how the individual forecasts add up at the runs level. The positive forecasts for the young players turns the Royals from a bad offensive team into an above average one, grading them out at +26 runs above average with the bats. Toss in another +34 runs for their fielding, and Steamer really likes the Royals position players. Pair that with a decent rotation and a great bullpen, and the forecasts think the Royals are clearly the best team left in the postseason, as good as any of the big boys who just knocked out in the first round.
Now, how much emphasis you put on these forecasts is a matter of opinion, and if you think that the only data that matters is what happened in the regular season, then our season-to-date Playoff Odds model probably aligns more with your expectations, with the Orioles as strong favorites to win both the ALCS and the World Series. If you think Moustakas, Hosmer, and Butler are more of what they showed this season than what the forecasts think, then the Royals probably aren’t a legitimately great team.
Personally, I’m probably somewhere in between, thinking the Royals are better than their 2014 performance but not entirely buying into the full improvements that Steamer sees for the Royals young hitters. But then again, I’m also the guy who would have had Anaheim, Detroit, Los Angeles, and Washington playing in the LCS, so Steamer’s doing better than I am this postseason.
Dave is the Managing Editor of FanGraphs.
The Cards have the “it” factor. Can’t measure it, of course, but you can see it on the scoreboard. And I am NOT a Cardinal fan.
Call me crazy, but I mostly see *runs* on the scoreboard
YOU’RE CRAZY!
I love that Guns n’ Roses tune. Especially the acoustic version on the “Lies Lies Lies” album. “Hey boy where you comin’ from, where did you catch that point of view?”
it’s TWTW measure that Hawk uses
http://chicago.cbslocal.com/2013/04/29/ofman-what-really-is-twtw/
They also have lots of clutch hitters.
They’re like a team full of Derek Jeters. Intangibles galore.
Pinnacle has the series odds at 56% for the Orioles, if we believe that to be a fair line. That is…not particularly close to 37%.
Yeah this. When your odds/projections differ so much from the betting markets, you should know something is wrong.
The point of betting odds isn’t to accurately predict sports
Yet they do
This is true but overstated. Since there is so much smart, sharp money out there, it is generally the case that the betting line is actually a solid prediction of the likelihood of teams winning. If it wasn’t, the sharps would break Vegas.
No, they don’t accurately predict the games’ outcomes and they’re not supposed to. The betting line is placed at the precise point where they think half the bets will be on one side and half the bets on the other. That’s it. Vegas doesn’t care who is actually likely to win. They care about accurately predicting which side people will bet on. That way, regardless of the actual outcome of the game, they make the same amount of money.
To say that the betting odds are an accurate predictor of game outcomes is to put your faith in the hands of the average dollar placed on a bet (which is somewhat different than placing your faith in the average gambler).
“Vegas doesn’t care who is actually likely to win. They care about accurately predicting which side people will bet on.”
And accurately predicting which side people will bet on (such that they get equal action on both sides of the line) generally gives you a solid prediction. This is because the average of all of the public’s opinions is generally a solid prediction of what will actually happen.
Fair enough. I guess that’s where we disagree, RSF. Whether the democracy of betting dollars is a better prognosticator than the a stats-based projection system.
Oh, I am not saying that the people will always beat a stats-based system. But, when the difference is this drastic, I am inclined to think that this specific projection system is off.
I’m curious as to the dollars bet by analysts and professional gamblers vs the dollars bet by fanboys and drunks and hopeless gambling addicts and the superstitious.
Isn’t this line of thinking the old “I’m smarter than everyone else, I can beat vegas”. Here’s a tip, you are not.
If the Vegas odds weren’t the best predictor of game outcomes, then it would be easy to become a winning sports bettor. Let me know when any of you quit your jobs and start making money by using Steamer projections over the Vegas lines.
There have occasionally been statistical or other betting systems that have consistently beaten the Vegas odds for a short period of time. Then they become widely known (or even published as research papers) and the Vegas odds adjust.
It’s not inconceivable that the Steamer method here is superior to what the current average dollar is doing (which includes referencing statistical methods– the Vegas odds are very close to Fangraphs season to date method), but over time it becomes unlikely. Just like it’s not inconceivable to find $20 bills lying on the sidewalk.
I’m not completely convinced that our ability to rate how prospects develop is well-developed. Luckily, there’s no way that this one ALCS will answer the question. People who do trust the Steamer method should make many wagers with it.
This is because the average of all of the public’s opinions is generally a solid prediction of what will actually happen.
Disagree.
This assumption does not take into account the actual desires of the general public. For instance, the odds of a horse actually winning the Triple Crown is always way lower than the amount of people betting it to happen.
Another recent example was the desire of the general public to see Peyton Manning win another Super Bowl last year. This raised the point line considerably higher than what it should have been by most statistics.
As someone who actually bets on sports for a living, and makes a nice living doing it, I thought I’d weigh in.
The sharp money generally drives the lines, especially in baseball. When you get any big public event (Super Bowl, BCS Championship, World Series) you’re going to have more square money so the line will be less efficient.
A projection system doesn’t have to be “better” than the market to win consistently. It just has to add value. With my model, I’ve found that to properly predict game outcome, I weight my system at about 50% and the market at about 50%.
The points isn’t to split the money bet when you’re talking about moneylines. The books want to (in theory) minimize their exposure to an outcome which can be very different than having an equal amount of money bet on each team (when odds differ from even money).
The betting market contains information from many statistical models, since sharps have their own models they’re betting with. It’s clearly the *best* predictor. Does that necessitate that differences between a statistical model and the market line have no predictive value? Of course not.
Of course bettors are going to favor the team with more wins. The premise of the post is that in this case, that doesn’t give a reliable prediction.
Occasionally odds differ from fair value. You will never ever find odds that differ so much from fair value that FanGraphs is suggesting.
In this case Steamer is betting on heavy regression to occur in a really small sample. I’d go with Vegas.
ya cuz doz odds iz alus u no a legit u no mehzer of the too teems anats awl nuthin tado with makin book doan kare hoo winz cuz books r peeyoor dis innarrested sports luver
Them r fightin’ words if you can translate English’ properly.
The Royals had a team BABIP of .302. The Orioles was .296. I understand that this might reflect on the talent in the organization, but it seems to me strange to adjust for those on the Royals who had a lower than expected BABIP like Moustakas and not those who had a higher than expected BABIP (e.g. Cain’s .380).
There are some important differences between the regular season and the playoffs- most significantly the extra days off. This leads to an increased value to the best 5 pitchers on the club. You could make a good argument that the Royals top 5 are better than the Orioles. You could make a good counter-argument that Buck Showalter is more likely to use his top 5 judiciously than Ned Yost.
Pinnacle’s odds seem about right to me.
“I understand that this might reflect on the talent in the organization, but it seems to me strange to adjust for those on the Royals who had a lower than expected BABIP like Moustakas and not those who had a higher than expected BABIP (e.g. Cain’s .380).”
Steamer would be regressing everyone on all clubs based on their model, including Cain. It’s just not mentioned in the piece explicitly.
Considering their speed, I think we’d expect the Royals to have a higher BABIP.
You mentioned the Royals’ advantage in the bullpen, but another thing is that the Royals’ contact-hitting fly-around-the-bases offense is ideal for October.
We’ll see how it plays out, but there are a couple factors in Baltimore’s favor: home field adv. in a HR hitters dream of a ballpark (and an excellent HR hitting lineup), and success limiting the SB and throwing out runnings via OF assists.
The Royals have an advantage in the bullpen? If so, it’s pretty small: Baltimore’s bullpen is deep and solid. Steve Melewski compared bullpens on his blog. O’Day, Miller, and Britton each have better WHIPs (I know I know) than Herrera, Davis, and Holland.
They are really similar teams in several areas, except that the Orioles have more power and the Royals have more speed. Given the way that the O’s pitchers and catchers have done a good job against opponents’ running games this season, I see the O’s power advantage as more significant.
The playoff odds on here are absurd. Fangraphs’ projection systems break when they encounter Baltimore. Doesn’t mean the O’s will win, of course, but it’s a lot closer than these projections.
Add to the slight bullpen advantage a MASSIVE advantage defensively. UZR likes the Royals outfield considerably more than Baltimore’s, so unless the O’s get a lot of pitches to hammer (not likely), there’s a lot to like about the Royals’ chances.
UZR and the Orioles outfield… it’s a can of worms, trust me. As long as it’s not Cruz or Young in LF for too long, I have no worries about the outfield defense.
In any case, both teams have strong defenses, and if the Royals have an edge there, it’s not that significant. And if Manny were still around… sigh…
KCDave, Royals have a great defense, but they don’t have a massive advantage defensively over the Orioles. The Orioles have more defensive runs saved and better range while the Royals have a better UZR (though the Orioles have a better UZR/150). In other words, they are pretty much neck and neck. The #1 and #2 defenses in baseball.
The projections do regress Cain’s unusually high BABIP, as evidenced by the fact that they have CF projected at 3.5 WAR but the CF’s actually produced 5.8.
You’re looking for a bias that just isn’t there.
The article is a bit frustrating because it conflates things. The article refers to Steamer’s 2014 position player WAR forecasts and these include both offensive and defensive contributions, and then refers specifically to Moustakas and Butler’s offensive contribution but not others. If Steamer projects offence differently from actual 2014 performance, measures of offence exclusively (wOBA, wRC+) would be more helpful.
If the basis for playoff odds projection is wildly different from actual season results, there is obviously a lot more to the story than the offensive contributions of Moustakas, Butler and Hosmer. All of the projection systems (Steamer included) had the Orioles as a .500 team or less at the outset of the season.
RIP 2014 Kansas City Royals
You have been tagged as a favorite by fangraphs
Speaking of jinxes, does the SI cover jinx still count since they are doing regional covers more often? Just wondering.
So long as it’s not the Giants or Cardinals, I will be content.
Being a Royals fan, I appreciate something to hang my powder blue hat on, but so many improbable things have happened this post-season, I have wonder how long it can go on. That said, Hosmer’s and Moose’s regression to the mean has really happened at just the right time, in addition to the qualities the Royals have shown throughout the year.
As an Orioles fan, I love this.
Imagine the color confusion of an Orioles Giants WS.
That sounded fun until I realized there are like 10 MLB teams with red/blue as primary colors.
last year was Cardinals-Red Sox
It would be the Halloween World Series.
To test each of these playoff series prediction models, we should have them predict each team’s record every single week during the season, and see how accurate they are.
Here is my favorite stat about this year’s playoffs: the ROYALS have now advanced to the LCS as many times as the A’S this century. The best laid plans of mice and Moneyball….Baseball playoff series are a crapshoot.
I always wonder about the A’s… is it set in stone that the A’s franchise simply had bad luck in the postseason, or could there have been some systemic problem with their roster and playing style that didn’t translate well to October? Cause they have now dropped what is a not-exactly-miniscule amount of ALDS’s. And the one ALCS they made they got swept.
Yes, their matchups were for the most part close. And this could be survivor bias or something similar. But surely one time they’d get an easy 4-gamer?
I like the A’s and Billy Beane’s style, it’s just an interesting thought to ponder. And it’s probably worth noting that the 2014 Royals play in a rather opposing manner to most A’s teams haha
Beane’s postseason match-ups by difference in fWAR:
2000 – Yankees (+1.8 WAR)
2001 – Yankees (+9.5 WAR)
2002 – Twins (+0.6 WAR)
2003 – Red Sox (-19.8 WAR)
2006 – Tigers (-7.7 WAR)
2012 – Tigers (-5.8 WAR)
2013 – Tigers (-11.9 WAR)
2014 – Royals (-1.4 WAR)
The A’s have basically be even or at a disadvantage every year with the exception of 2001. I know even SABR people love to fall in love with Moneyball, but it’s not like they’ve ever really entered the series as a favorite by SABR standards.
So are you saying they overachieved in most of their regular seasons? Why have their regualr season win totals been so high?
Shouldn’t the positive/negative indicator on “difference” be reversed? For instance, it appears at a glance at the “difference” line that the DH position outperformed their projection by 3.3 wins, rather than underperformed. It’s odd that a positive number would denote underperformance.
After seeing that ugly video of racist Cards fans, I just hope it’s not them.
I think you need to look deeper on BABIP for Moustakas. Isn’t he always shifted? And isn’t he notorious for hitting into the shift? And wouldn’t that kill his BABIP? Some players with low BABIPs these past couple years may never rebound because the shift has become more prevalent recently.
I hope you realize that there are bad eggs in every fan base. St. Louis doesn’t have every racist in the world cheering for them, they just happened to have a high-profile event happen close to their city so all the idiots could gather in one place.
Also, it’s ludicrous to me to project that a guy with CAREER .260 BABIP coming off a .220 season is going to somehow hit .272 BABIP going forward. I get adjusting for age and everything, but that is just silly.
Every fan base also has some incredibly loyal and passionate fans, but not every fan base goes around with the self-congratulatory “worlds greatest fans” shtick quite like the Cardinals do. I understand it is not overly rational but I don’t mind seeing a little egg on the collective face of all Cardinal fans.
C’mon man!
Well, now TWO high-profile events, one actually IN their city & the other in an area so contiguous & related as to be effectively indistinguishable from their city.
I haven’t thought out yet how one would approach an effort to objectively measure the EFFECT that these events, the associated protests & counters all happening in the contexts of a Mason-Dixon line city during the last national election touching on the only ever black POTUs, might have on the games in this series held in St. Louis; I’m not confident such a measure is even available. But it’s hard to accept there’ll be NO effect, or that the bulk of any negativity won’t befall the Cardinals. And in the context of a series against their closest comparator, that can’t make Cardinals supporters feel comfortable, particularly given how this makes them & their city look.
I’m actually thrilled to see Fangraphs still bagging on the O’s. Was starting to get nervous with everybody picking us in the LCS.
That’s a lot of words to write about a net <1 WAR difference between the actual and projection. 🙂
You seem to have left out the pitching side of the ledger.
In the playoffs, every game is a coin flip, just ask the four Cy Young winners who couldn’t get a win.
Royals speed and defense should play well in Camden Yards, Orioles pitching could be a little better at the K. It will all get down to who punishes a hanging slider or doesn’t make a routine play. That’s why they play the games.
There were actually five Cy Young winners without a win I believe; Verlander, Scherzer, Price, Greinke, and Kershaw. Peavy is the only Cy Young winner to make a start and earn a pitcher win. If you are talking about their team winning, then yes subtract Greinke from that group .
Even year, obviously the Giants are the favorite.
If these are the year of the underdog, Fangraphs put the analytical nail in the Royals coffin. Orioles in 5.
Man, something is definitely off here. 63% is an insanely high probability even if you don’t know anything about the teams. To have a 63% chance of winning a seven-game series, the favored team needs to have about a 56% chance of winning each game, which implies that it’s a HUGELY lopsided matchup.
I futzed around with these numbers a while back when I was trying to figure out how unlikely the outcome of the 2006 WS was. Using the Tigers’ and Cardinals’ 2006 Pythagorean records (97-65, 83-79) and applying Bill James’s log5 method, you get a 58.54% chance of the Tigers winning each game (for a 68% chance of the Tigers winning the series, and a below-10% chance of the Cards winning in five). Are the 2014 Royals almost as good relative to the 2014 Orioles as the 2006 Tigers were to the 2006 Cardinals? Hell no. Most of us probably wouldn’t say they’re the better team at all. I know that I’m kind of doing apples-to-oranges by comparing a prediction based on Steamer projections to one based on Pythagorean records, but my point is just that giving the Royals 63% odds implies that they’re the MASSIVE favorite.
But also obviously the Cardinals crushed the Tigers in 2006 so lol, can’t predict baseball.
Good analysis, I don’t understand that 63%, either. If you make some of the allowances for KC hitters being underestimated that the author does, you might favor the Royals, but a margin by this much doesn’t make sense.
And by the way, where was this analysis prior to the series with the Angels?
I too was curious how two postseason teams could ever realistically have a 63% chance of one team winning. I was really hoping Dave’s article would have some stuff more in the spirit of what you did here, and really get into the nitty-gritty of how their series predictions work.
So thanks for this.
” But then again, I’m also the guy who would have had Anaheim, Detroit, Los Angeles, and Washington playing in the LCS, so Steamer’s doing better than I am this postseason.”
I’m confused–if you agree that playoff results are dependent on the whims of small sample size production, then how can anyone come up with a good formula for predicting the playoffs? Steamer is great for the regular season, but isn’t it more or less irrelevent in a 5 or 7 game series?
I made the same observation on another thread and was told that I “don’t understand FanGraphs, especially the math part.” To save that dear reader the trouble, I’m passing along his wisdom to you.
Vegas does a decent job of working out the likelihoods of individual teams winning one-off games in all sorts of sports. Having done this, it’s not hard to work out the chances of a given team winning a series.
I love Fangraphs, and there’s certainly value in a piece like this. But to paraphrase Billy Beane, this shit doesn’t work in the playoffs.
It also didn’t do so well in the regular season when it predicted the Orioles for last place.
We’re still whining about this?
We’re still pointing out how there is probably something wrong with the prediction system, yes.
“But the Steamer forecasts — the engine powering our Playoff Odds models — aren’t really phased by Butler’s lousy 2014 season, and think he’s basically still the good-not-great hitter he’s always been.”
Fazed. They were shot by Captain Kirk?
All nonsense. Royals faced 2 cold hitting teams in the playoffs, the A’s and Angels. Just look at their numbers the last 2 weeks. Only hope for them is the layoff cooled off the Orioles bats.
I doubt the series goes 6 games.
Just want to point out that the Playoff Odds calculations here have been broken as recently as a few weeks ago, with results that were easily shown to be incorrect. I think it was the Mariners or Pirates that needed to win all their games, and have the Cards/Angels lose all of theirs in order to force another game. They were showing one of the teams as having a 10% chance to win the division or wild card, despite it requiring the team winning 5 or 6 coinflips, essentially.
If a team is something like a 63/37 favorite, they’ll have about a 10% chance of winning 5 “coin flips.” Maybe it’s hard to have expectation that high, but it’s not insanely outside the realm of possibility.
C’mon, though. Any prediction system that sees the Royals as such huge favorites is broken. I suppose if the Orioles do win, everyone will say, “well, 37% chance of happening, so it doesn’t say anything about our model.” But I think most people would agree that this model is doing something wrong to see one team as such a prohibitive favorite in this series.
Yeah, that’s really the key thing. Seeing the Royals’ projected WAR shows they’re a good team, but it’s hardly the projection of a dominant juggernaut that should theoretically be necessary to make them a 63% favorite.
This was the Pirates and the opponents of the Cardinals, both of which were no where near 63% to win. Both the Pirates and the Cardinals were at ~.550 on the same page showing the crazy 10% figure. By their own assumptions, the 10% number couldn’t possibly be correct.
Steamer has a known bias towards younger players. I think we all ( including Dave) think these odds favor the Royals overmuch
Oh, here goes Fangraphs again, predicting doom & gloom for the Orioles, who are obviously the luckiest team in the history of sports, for what, three years running now?
When the Orioles win the World Series, Fangraphs is going to predict it will rain on the parade.
No FG, you don’t hate the Orioles. Your projection system just sucks.
I’d say the jury has not returned a verdict yet.
Is there something analogous to confidence intervals for these projections, e.g., 63% Royals with a 95% CI of ___ to ___? Or even an 80% CI?
Anyone actually watched the games so far? Maybe the projections were not so far off.