The Games are Already Mattering
One of my favorite tools we have here is our Playoff Odds page. I used to make frequent use of Cool Standings, and I keep track of hockey playoff odds at Sports Club Stats. I like looking at the playoff odds because they give the clearest sense of where things stand, and in theory, the whole point is getting to October, so why not monitor how likely that is in real time? Of course, we’re not just in it to see the playoffs — if that were the case, most baseball fans would be pissed off all the time, if they’d be fans at all — but getting to the postseason is at the core of most of what we talk about. Why do we care about player analysis? Good players make teams better. Better teams are more likely to go to the playoffs and win the World Series. And so forth.
Eventually, on our Playoff Odds page, you’ll be able to click through and see how the odds have changed over time. That hasn’t rolled out yet, though, nor are people probably even thinking about it in the first week of April. The season is so new we’re all still trying to find our baseball-fan footing, and it feels like the whole thing is in front of us. Most of it is, to be sure, but what’s already happened counts. And it turns out, what’s already happened for a few teams is pretty significant.
This graph is going to say pretty much everything:
If you want to know where teams stand today, go to the Playoff Odds page. This graph shows how much the odds have changed, in percentage points, from just the start of the year. Note that this is through Thursday, and doesn’t consider Friday games already complete. Note also that I’m pretty sure the Dodgers/Diamondbacks games in Sydney are excepted, so, for them, this is showing the Dodgers’ change since being 2-0, and the Diamondbacks’ change since being 0-2. That’s why the Dodgers have lost a tiny bit of ground, and there’s also the Clayton Kershaw injury factor. Also, these come with error bars, because the odds aren’t based on enough simulations to completely eliminate the noise.
But at either end, there are some really interesting data points. The Mariners are already up eight percentage points, despite having lost in extras Thursday night. The Giants are right there with them, having taken three of four from a division rival. It’s never too early for higher-leverage games, and not only does beating a rival give you wins and give them losses — it reduces the number of opportunities for those games to be taken back. Arizona has lost ground to San Francisco, and the two now have four fewer scheduled head-to-head games to play.
Over on the right, we could talk about Colorado or Arizona, but really, this should be about the Angels. In three games — three games! — the Angels’ playoff odds have dropped by ten percentage points. They were swept at home by a rival, and now the Angels have the worst playoff odds of the AL West foursome. They have lower odds than the Indians, and lower odds than the Braves. A few days ago, it was entirely reasonable to think of the Angels as the most likely team to win the division. This quickly, they’ve fallen behind, and whatever you thought of the Angels before, now you have to include in that projection an 0-3 start.
The best way to think of what’s happened to the Angels? On average, if you allow a leadoff home run in the top of the first, you lose about ten percentage points of win expectancy. There’s still a lot of game to go — pretty much all the game to go — and people are stunned more than anything else, but from that point forward, you have to outscore the opponent by two instead of one. That run goes on the scoreboard immediately and it doesn’t come off, and just because it happened quickly doesn’t mean it matters less. The winning process can start out of the gate, and the same is true of the opposite.
After a slow start a year ago, the Angels were a better-than-average baseball team. After a slow start the year before, the Angels were one of the best baseball teams in the majors. In neither year did the Angels make the playoffs, so they understand better than most the importance of not doing what they’ve already started doing. The good news is they get the Astros on Friday, but then the Astros are much improved and our playoff odds account for schedule strength.
Hit a leadoff home run and you know you still have a full game to play. Allow a leadoff home run and you know the same thing. But a leadoff home run does change the dynamic, and sometimes games end 1-0. Just because you might not be ready for baseball yet doesn’t mean baseball hasn’t already started happening, and there’ve already been a few plot-changing twists.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.
Great article. Funny to point out that the bar in the middle in which one cannot see the franchise logo is that of the Astros’. Their projection changed to a whopping 0.3% by beating the yankees 2-1. Way to go!
I believe that’s the Nationals just to the right who have lost a bit of their chances, presumably from the Fister and Ramos injuries. (Or maybe Fister was already priced in?) Or anyway I assume it wasn’t their 3-0 start that made them slightly less likely to make the playoffs.
No. It looks like the Nationals are actually up 2.5% and are in between the blue jays and Red Sox on the chart.
Right you are… missed them there (the chart doesn’t seem to get any bigger than it appears in the post). Personally, I like them to come out of the East less (and I already thought the Braves had the inside track) after losing Ramos and Fister for at least a month apiece.
pretty sure that’s the twins between the astros and orioles
Are you sure about that Anon21?
It’d be really odd for you to think the Nats chances are not as good as FG calculates and the Braves chances are better.
You tell me, cass. Do you think the Nationals’ chances are better today than they were the day before Fister was DLed? I have a hard time even seeing an argument that that’s the case. It’s probably not a substantial swing—two players only matter so much, neither Ramos nor Fister is Mike Trout, and they each will miss only one month—but surely the Nats’ chances have gotten somewhat worse.
The Fister injury happened before the season, so it was already factored into the baseline. Since then, the Nats won three games against the Mets and Ramos got injured.
Not sure if today’s game was factored in.
What Anon21 meant is that, in his mind, the Nats chances of making the playoffs have fallen from 15% to about 8%. Since he doesn’t give them much chance of making it anyway.
Well Anon21 is the resident Braves homer who subtly complains here and there that the Braves aren’t getting their due and that the Nationals are overrated. And I say that as an enormous Braves fan.
“Well Anon21 is the resident Braves homer who subtly complains here and there that the Braves aren’t getting their due…”
Subtly – I don’t think that word means what you think it means. 🙂
AL West FIVEsome actually. Hard to get used to.
I think he is completely ignoring the Astros for the purposes of talking about playoff odds, which: fair enough.
I believe you mean the over .600 winning percentage Astros, thank you very much.
They are not ignored, the chart counts 30 teams. I _think_ they are the ones without a real bar, due to the quality of the image I am not that sure, though.
I meant in the text, though, which is what Ceetar was complaining about.
Astros play in the NL what the hell are you all talking about?
That Derek Bell is pretty good, no?
The implication is that there are four teams that could reasonably compete this year.
If you think the Astros are going to compete you might have more important problems than this graph.
This might be an ignoramus question, so pardon me…do the Playoff Odds have any kind of “confidence rating” that grows as the season goes on? Having an 87% chance of going to the playoffs today is great, but having an 87% chance in September would be better, right? Or am I missing something?
An 87% chance in October would be more meaningful yet.
I don’t think you’re missing anything. That “87% chance” today is based on a few games and then a whole hell of a lot of projection. The 87% chance in September is based on a ton of games with fixed, unchangeable outcomes and a little projection. The error bars would probably be pretty tricky to construct, though.
I disagree. I think an 87% chance means that 87/100 seasons simulated from this point result in the team making the playoffs. I think tomemos is suggesting that a team with an 87% chance in September is somehow more likely to make the playoffs than a team with an 87% chance today, and I don’t believe that’s the case. The odds are the same either way.
But aren’t the simulations more reliable later in the season, when there’s less of the season to simulate and more actual data? I know that a “true” 87% in April is as good as a true 87% in September, but it seems like simulations would get truer as the year went on.
By September we will have a better sense of each team’s true talent level, so in that sense it will be more accurate.
If you think projections are a perfect simulation of the future, that’s true. But we know they aren’t, so the less projection you need to factor into your playoff odds, the better off you are.
No, that’s not correct. An 87% chance in April is the same as an 87% chance in September, but a team would have to win a ton of games in April to get to 87%. The “confidence rating” is just the playoff odds themselves.
The confidence interval for a percentage is usually directly related to the value of the percentage and the number of observations (simulations). If so, then 87% means 87% no matter what the calendar date, assuming the number of simulations is identical.
You could quibble with the idea that the simulations themselves are equally reliable now and in September. If so, then it might be correct to say something like: “I am less certain about the 87 percent probability because it was calculated based on April simulations”. One could reasonably build this level of certainty into the confidence interval and say something like: “I am 95% certain that the probability team A will make the playoffs is between 77 and 97 percent”
The Nationals’ playoff odds were 87%, like, two days ago. They’re still at 85%, as are the Dodgers’. Apparently teams do *not* have to win a ton of games to get there!
tomemos: That’s because preseason expectations are being factored in. I was thinking of odds calculated using only the standings. My bad.
Right, my point being that the preseason expectations are much less reliable than the actual information about each team’s true talent level and place in the standings after ~130 games.
This is a great feature. One nitpicky note on simulation: you may have error bars due to simulation error (not enough repetitions). But even if you had more repetitions, you would still have error bars due to input uncertainty.
I prefer the coin flip mode. And even it shows that a few games can make a pretty big difference.
Now I’m not real big on all the words you use on this site — makes it awful talky so I skip most of it. But based on the colorful picture, I have placed a large sum of money on the Mariners beating the Giants in the World Series, after respectively dispatching the Tigers and Braves. Y’all better be right.
Playoff odds and change in playoff odds are not the same thing.
I believe the commenter was trying to make a funny.
Well, yeah, Tigers get off to 3-0 start, but that does not improve their playoff odds as much as some, given that their odds were already pretty good.
Rangers beat the Phillies 2 games to 1. Rangers fall farther out of the playoff picture than the Phillies. Bah Humbug!
It was never reasonable to think the Angels had a chance to win the division, much less that it was likely, shows how flawed the projections were to begin with.
How was it unreasonable to think the Angels had a chance to win the division? They still have a chance to win the division, three games is not a huge swing in the standings if true talent level remains unchanged. It is not like they are the Astros, they are still a talented team with Mike Trout.
Bunch of noise. By the end of the month, only a handful of games will be separating the top from the bottom. Everyone is sorting it out for real, with their 24 players. .500 is a good result at EoM. Not many teams see hope die before June…..except perhaps the LAAAofAAA
I think this is too much of as swing this early in the season. Difficult to believe that the playoff odds can change this much. I see the Angels in the playoffs, as much as I don’t want to believe it.
“Just because you might not be ready for baseball yet . . .”
Who could possibly not be ready for baseball yet?
I was ready in December.
Folks, the important thing is winning in September. Those games count a lot more.
Mike makes a good point here.
Also, I really think those Marlins will finish above .500 based on their first series against the Rockies.
(He said this! He actually said this! And is gainfully employed to speak intelligibly about baseball! Jesus damn.)
Also, two games in March tell me a lot more about a coach than their entire body of work beforehand. A LOT. Florida Gulf Coast wins a couple tourney games? Sweet, let’s sign that dude to coach Duke or UK!
A win now counts the same as a win later in the year….
…swing and a miss.
He’s spot on. A win in April = 1W. A win in September = 1W. A September win feels more important, but it counts for just as much – there is just more context.
…..strike two.
I feel like we’re in the middle of a JP Arencibia at-bat. Lots of flailing and missing.