Every sports league deals with the challenge of making the final weeks and months of its regular season seem relevant for as many teams as possible. They do this by offering various incentives -- playoff spots, playoff byes, seeding systems and home-field advantages -- so each team doesn't have just one, perhaps unreasonable, goal to shoot for (the most wins) but a series of them.
This is a complex challenge, since the contours of each season are unknown in advance. Sometimes these incentives work beautifully and a lot of teams are between, say, 10 and 90 percent likely to make the playoffs toward the end of the season. Any team within that range will see its chances vary with each day's results and be compelled to treat every day like the season might hinge on it. This makes rooting fun and viewing essential. But sometimes, even these incentives leave most teams, good and bad, sleepwalking through the end of the regular season.
On Sept. 1, 2015, for example, all five NL playoff spots had more or less been decided: The Nationals, Dodgers and Cardinals were all north of 90 percent likely to win their divisions. The Pirates and Cubs had basically each wrapped up a spot in the wild-card game. There were essentially no stakes to any National League game played in September 2015:
(All playoff odds and charts via Baseball Prospectus.)
The majority of the 2017 playoff spots were locked in by mid-August. The Indians, Dodgers, Nationals and Astros were all near locks to win their divisions. The Yankees and Red Sox were each overwhelmingly likely to make the playoffs, with the Red Sox's steady lead in the AL East heating that race only to a gentle simmer. The Cubs grew to be about as safe in the NL Central, and the Diamondbacks were all but eliminated from the NL West but all but sure of a wild-card spot.
The second wild card -- which had backfired as a tension-builder in the 2015 National League -- was our only salvation, keeping a dozen teams in the race for the final spot in each league. This was great for fans of the Royals, Orioles, Marlins and others teams who could maintain a little hope despite season-long mediocrity. But for fans of the leagues' best teams, with almost no reason to scoreboard watch, this looked like it would make for six boring weeks. No game the Dodgers or Indians played in late August or in September would shift their outlooks to make the Division Series. They were in the green room, waiting, knowing nothing would change.
Nothing did change in either team's division:
Yet, everything changed.
Aug. 10: Dodgers 25.6 percent likely to win World Series; Indians 15.2 percent
Playoff odds work like this: Baseball Prospectus, using a blend of projections, actual performance and playing-time estimates, calculates what it thinks a team's true talent level is. The Dodgers, on this day, were estimated to have the true talent of a .619 team -- a team that, if a brand new season were to suddenly break out that day, would win about 100 games. (They were actually on pace to win more than that, but show any projection system an outlier performance and its instincts are to regress it a bit.) Cleveland was figured to be a .577 team, which is about 93 wins in a full season.
BP then puts two teams' expected winning percentages against each other (along with adjustments such as home-field advantage) in all scheduled matchups and calculates the likelihood each team will win. It simulates entire seasons, and then postseasons, based on these expected winning percentages. A team's assigned chances of winning the World Series change throughout the season depending on how likely the team is to make the playoffs, how good a team it appears to be and how good the teams it'll face in the postseason appear to be.
After losing on Aug. 9, the Indians traded for Mets outfielder Jay Bruce. He would be the last player Cleveland would add. The Dodgers would acquire Curtis Granderson 10 days later, but for the most part, each team's postseason roster was set. Cleveland had a little work left to do to guarantee the division, but for the most part, each team's entry to the postseason was set. The Dodgers were by far the team most likely to win the World Series, according to BP's odds, because they were a lock to get to the postseason and because every bit of information we had on that date suggested they would be the best team in the postseason.
Aug. 25: Dodgers 27.2 percent likely to win World Series; Indians 15.8 percent
Two weeks after Cleveland added Bruce, and a week after Los Angeles added Granderson, the Dodgers reached their best World Series odds of the season. They had gone 11-3 in those two weeks.
The key detail here, besides the relatively small Granderson acquisition, is nothing had actually changed. The playoff odds had already, weeks earlier, given the Dodgers a 100 percent chance of winning their division and conceded them home-field advantage throughout the National League playoffs, so the 11 wins hadn't materially altered their path to the World Series. The roster was basically unchanged, and nobody unexpectedly got hurt or got healthy. The only thing that changed is that the odds assumed the Dodgers were even better than initial assumptions had been. With those additional 11 wins, the Dodgers had further convinced the computer they really were a historical anomaly. Their expected winning percentage was up to .628, a 102-win team over a full season.
Sept. 3: Dodgers 23.2 percent likely to win World Series; Indians 19.5 percent
This is the first day BP's playoff odds gave Cleveland a 100 percent lock on the AL Central. That reflected a real change to the real standings, and it bumped Cleveland's chances in a real way -- but only slightly, since it had already been between 90 and 99.9 percent likely to win the division.
More significantly, Cleveland had just won its 10th game in a row, and its expected winning percentage had been adjusted up to .597. It passed Houston as the American League team most likely to win the World Series. Meanwhile, the Dodgers had just lost their third in a row and eighth out of nine. These games didn't matter, but they counted as information. The Dodgers' expected winning percentage was back down to .618.
Sept. 8: Dodgers 20.6 percent likely to win World Series; Indians 21.2 percent
After five more Los Angeles losses and five more Cleveland wins, the odds officially declared Cleveland more likely to win the World Series. No players had changed teams, and no relevant change to the actual standings had occurred, but the longer the season went, the more information pointed to Cleveland being better: a .611 winning percentage, to the Dodgers' .608.
Sept. 14: Dodgers 20.1 percent likely to win World Series, Indians 23.8 percent
Cleveland hit its (to-date) high mark for the season on the same day it won its 22nd game in a row. Two days earlier, the Dodgers scraped 20.0 World Series odds, the lowest they had been since June 19, when the NL West was still competitive.
Sept. 21: Dodgers 20.1 percent likely to win World Series,;Indians 23.1 percent
The Dodgers continued to struggle, but their odds held steady mostly because their likely first-round opponent -- the Diamondbacks -- dropped three straight games. The Indians continued to roll, but their likely first-round opponents -- the Yankees -- had solidified their hold on a wild-card spot and improved their expected winning percentage.
Which brings us up to date: The Indians, once an underdog to the mighty Dodgers, are now the favorites over Los Angeles.
And it's all, basically, an illusion. In the first place, it was an estimate of probability based on a lot of educated guesses. Now, it's an estimate of probability based on slightly more informed guesses. On Aug. 9, we all thought the Indians were a team that scored 4.8 runs a game and allowed 4.0. Then we got more information, which adds up to better information, and we now know they're a team that scores 5.1 runs a game and allows 3.5. The Dodgers aren't a team that scores 5.1 and allows 3.3, as it looked like on Aug. 9; they're a team that scores 4.7 and allows 3.6. Still great. But everybody was a little bit wrong about them.
Now here's the head-trip part: If the season had ended Aug. 10 instead of Oct. 4, the Indians still would have been a team that scores 5.1 runs and allows 3.5, and they would still have been more likely to win the World Series than the Dodgers; we just wouldn't have known it. An omniscient god would have known, just like an omniscient god would know Mike Trout is a better baseball player than I am, even if baseball had never been invented. But we humans require observation and data before we can know things. After Aug. 9 we got more data. The data changed a great deal about what we know, but nothing changed about either team. Teams change data. Data don't change teams.
I don't think it feels that way to fans of the Dodgers or Indians, though. The other day, I was listening to a radio broadcast of some other game that truly didn't matter, between two out-of-it teams. The broadcasters got to talking about the Indians and Dodgers. "Would you rather be the Indians than the Dodgers right now?" one asked. It went without saying. A month earlier, the opposite would have gone without saying. The net result is that, for a month of baseball that had no tangible impact, Cleveland and Los Angeles fans got to care.
OK, maybe these games did change something. Maybe momentum is critical here, though research into the correlation between September performance and October performance has never turned up any relationship. And even if momentum is very real (plausible!), who could possibly assume that late-August momentum would still have any force in mid or late October? June and July momentum obviously had no force in late August.
For that matter, maybe the reverse could be true: Maybe the Dodgers' struggles were a wake-up call, a push to treat these next six weeks with more urgency, a sign that Pedro Baez shouldn't be on the mound in a key situation. The World Series odds at BP don't consider the momentum a factor in either direction.
So why do we do this? Why do I look with great interest at these odds shifting every day, even as nothing tangible changes? Why do Dodgers fans panic and Indians fans rejoice over outcomes that no longer shift any standings that matter?
I think it's this: Baseball produces two different rewards systems. For the players, the rewards of winning or losing, homering or striking out, are tangible. If they win, they get paid more money, the world remembers their names differently, the historical record of their worth is altered and their sense of accomplishment is directly tied to an action they did. For about three hours every day, 162 times a year plus games in October, they get to affect these real, tangible returns. After each game, they go home and have to wait for their next chance.
For the fans, winning or losing is nothing more than emotions. We have no actual effect on the outcome, and we get no actual return, other than the dopamine, endorphins, oxytocin and vasopressin that affect our moods. We get to feel happy, and we get to avoid feeling sad. It's a less tangible return, but it's also unbounded by the action itself. We can feel emotions 24 hours a day, and we can feel them all year.
All we have to do to collect these emotional returns is believe these probabilities are shifting based on how good we think the team is. A fully invested fan can convince herself a spring training game matters, so she can certainly believe that three late-season losses against the Phillies -- which technically, tangibly matter a lot more to the Phillies than to the Dodgers -- matter. A fan has incredible power to turn "nothing" into something. Since these games exist for fans, that makes it not nothing at all.