Why the Cleveland Indians Were One of Baseball's Unluckiest Teams This Season
Remember that old game on Sesame Street "one of these things is not like the others"?
Well, cue the music for the top five teams in our MLB power rankings.
Leading the way is Toronto, who is 92-66 and has clinched the American League East, followed the by the 90-72 Dodgers, who won the National League West.
The Blue Jays (29.6%) and Dodgers (17.4%) are also the top two teams in our World Series odds.
Next up in the power rankings are the Astros and Cubs, who rank sixth and seventh in our championship odds, respectively, so there’s nothing out of the ordinary yet.
Who is the fifth-best team? That would be the Cleveland Indians.
That’s the right, the sub-.500 Indians, who were officially eliminated from postseason contention on Wednesday, are rated as about 0.7 runs above average.
What the heck is this 78-79 club doing above the likes of St. Louis and Kansas City? And what went wrong if they were so good?
Let’s take a look.
Cleveland is just 18th in runs scored per game (4.2) but is tied for eighth in runs allowed per game (4.0). They have a plus-21 run differential, and their average scoring differential ranks 13th.
They have been somewhat unfortunate in one-run games (they are 15-18) but have hardly underperformed at an extreme level here (unlike the also underachieving Oakland A’s, who are 19-34 in one-run games, and rank 16th in our nERD rankings despite MLB’s fourth-worst record).
Had the Indians played an average schedule, we might look at them as a slightly-above average team.
This, of course, is not the case, as they have had to face one of the toughest schedules in baseball.
Per ESPN, Cleveland’s schedule is tied for sixth most difficult, and by Baseball-Reference’s formula, it is tied for ninth, so the Indians look like a better team when you account for quality of competition.
Another One Bites the Sequencing Dust
Just as we can use run differential to calculate an expected winning percentage, we can also use underlying stats to derive an expected run differential, which is more indicative of true performance.
In the long run, expected records derived from stats like BaseRuns and Baseball Prospectus’ second- and third-order winning percentages tend to come close to a team’s actual record, but 2015 has seen unprecedented deviation between actual and expected records.
On one end of the spectrum, we have St. Louis, Minnesota, Texas and Kansas City. The Cardinals have won almost 11 more games than we would expect given their third-order winning percentage, while the Twins, Rangers, and Royals also have at least a plus-eight differential.
At the other end, we have Cleveland, which has won 12-fewer games than their second- and third-order winning percentages imply (second-order winning percentage uses underlying stats to derive a winning percentage, and third-order winning percentage adjusts this for opponent).
Based on this, and the fact there is no evidence “clutch performance” exists as a repeatable skill on the team level, we can conclude the problem for Cleveland has not been offense or defense, but the order in which events have occurred.
The Indians are seventh in baseball in offensive runs above average, according to FanGraphs, as they are tied for 13th in wRC+ and are sixth in baserunning runs above average.
This raises the question as to how they are only 18th in runs scored per game.
The answer, as you probably guessed based on the preceding paragraphs, is sequencing.
With the bases empty, Cleveland hitters have combined to post a slash-line of .258/.320/.410, and after park-adjustments, this translates to a 102 wRC+, tied for third-best in MLB in this split.
The opposite has been the case in Cleveland, though, as with men on base, the Indians’ wRC+ drops to 95 (tied for 20th in baseball).
Over-performing or under-performing with men on base does not seem to be a repeatable skill, so Cleveland's drop off here does not seem to be an indictment of actual talent. Since 2012, the correlation in a team's OPS with men on relative to its overall OPS from one year to the next is -0.08 (so virtually not correlation at all; a correlation coefficient of 1 implies a perfect relationship, and one of 0.0 implies no relationship).
The splits are also drastic on the defensive side of things.
Overall, the Indians are tied for the eighth-best ERA- (94) and fifth-best FIP- in baseball (90), while their BABIP allowed is .290 (their unadjusted FIP is 3.64, seventh in MLB).
With the bases empty, their 3.50 FIP ranks seventh, while their .270 BABIP is the lowest in the game. Opponents are hitting .218/.275/.368 in this split, translating to a .282 wOBA which is third lowest in the Majors.
When runners are on base, Cleveland’s FIP drops modestly (to 3.86) but still ranks ninth. The damage has come from balls in play, as the Indians allow a league-high .320 BABIP with men on, and their opponents are hitting .264/.334/.423.
In order to chalk this up to anything beyond sequencing randomness, you have to assume something about the Indians’ pitchers is making them pitch worse with men on.
If this presumed deficiency is real, it had never manifested itself before this year. Cleveland pitching was actually above average last season in terms of wOBA allowed with men on (.308, ninth-best).
This season, the three pitchers’ wOBA rose by .050, .088, .018, while the American League average increase was .016.
We can conclude, then, that their past success pitching with runners on is either a product of a skill they once had but suddenly lost, or more likely one that is the result of random variation.
Another way to look at Cleveland’s misfortune is the “Cluster Luck” rankings over at The Power Rank, which measure the runs lost to sequencing.
The Indians only the trail the Tigers in this metric, as Cleveland has lost a staggering 57.2 runs to clustering. Defensively, the Indians are minus-22.4 (behind only Oakland), and offensively, they are minus-34.8 (24th in baseball).
The end product is an above-average offense with a great pitching staff that was done in by failings in high-leverage situations (and don’t forget, these were exacerbated by a tough schedule).
Cleveland finally started to play up to its true talent over the last month of the season, going 23-15 since August 20, the sixth-best record in baseball during this span.
It was too little, too late though, so despite being one of baseball’s better teams this season, the Indians won’t have a record that shows it.