The Dodgers Weren’t as Streaky as You Thought

This year saw two of the best teams in baseball go on wild streaks. The Cleveland Indians famously went on a 22-game win streak over August and September, winning 27 of 28 total games at one point. At the same time, the Los Angeles Dodgers (who had won 43 of 50 earlier in the season) accumulated an 11-game losing streak. Over the course of the season, they racked up separate 9, 10, and 11-game win streaks before going on their 11-game slide. Given these absurd streaks, I had a hunch that the Dodgers could be the streakiest team of all time. With all of these streaks happening concurrently, I set out to find the â€śstreakiestâ€ť team in baseball history.

One way to define â€śstreakinessâ€ť is by the number of â€śrunsâ€ť a team goes on over the course of the year. A â€śrunâ€ť in this case does not mean a baseball run, but a run of wins or losses. For example, the streak WWLLLWWWW has three runsâ€”two Wâ€™s, three Lâ€™s, four Wâ€™s. Another way to think about the number of runs in sequence is that it is the number of times we switch from a W to an L or vice versa + 1.

The expected number of runs in a sequence can be calculated by:

R = 1 + 2*(n-1)*p*(1-p)

where n is the number of trials (in our case, games), and p is the probability of a success (in our case, a teamâ€™s winning percentage). This equation is fairly common, with a good discussion and proof of it here.

Using n = 162 and each teamâ€™s individual winning percentage, we can calculate a teamâ€™s expected number of runs for each season.

I scraped data dating back to 1968 (to allow for 50 years of data) in order to find each teams true runs over the course of the season. Teams that have very few runs compared to their expected runs are â€śstreaky.â€ť Think about a season that consists of 9 games, with a team that has a 0.5 chance to win any given game. Using our equation, we expect 5 runs. A team that only has 2 runsâ€”WWWWLLLLLâ€”is highly streaky.

I omitted the strike shortened 1994 and 1995 seasons and then calculated the â€śruns below expectedâ€ť (RBE = expected runs â€“ actual runs) for each team.

The chart is below:

Interestingly, the below average 2002 Toronto Blue Jays were the streakiest team of the last 50 years by my metric. This team went 78-84, but their wins and losses both came in clusters, as they had 18.4 fewer runs than expected. Below is their game chart (courtesy of baseball-reference.com), where green bars indicate wins and red bars indicate losses.

Shockingly, this yearâ€™s Los Angeles Dodgers team was nowhere near the top, ranked 33rd with 12 runs below expected. I would have thought that their many long streaks (both winning and losing) would have vaulted them higher, but this was not the case. They lost relatively few games this, with 16 of those losses coming over a 17 game period. This yearâ€™s Cleveland Indians did not make the list, notching only 7.1 RBE. While these high profile teams did not top the chart, I noticed that many of these â€śstreakyâ€ť teams would likely receive no fanfare or recognition whatsoever. The list of streakiest teams is wildly unimpressive, littered with teams barely above or below .500.

When we filter the data to only account for â€śgood teamsâ€ť (over .600 winning percentage), the Dodgers are in fact the streakiest team. Other notable teams on this list of good streaky teams include the World Series winning 2005 White Sox and the 2002 Oakland Athletics, whose 20-game win streak was prominently featured in Moneyball.

One reason I (and the rest of the baseball world) may have been so convinced of the Dodgersâ€™ streakiness is that they were widely considered the best team in baseball throughout the season, so their streaks were consistently in the headlines. In streaky fashion, the Dodgers won their first six playoff games, until ultimately falling to the Astros in a seven game World Series.

Seeing the list of unimpressive streaky teams made me wonder if there was any correlation between RBE and winning. Looking at the scatter plot of RBE vs. winning percentage, we can immediately see that being very streaky, or the opposite (consistently trading off between wins and losses) has no correlation with winning percentage. Being streaky or not simply does not affect how much a team wins. AN interesting follow-up may be to look at how streaky teams do in the playoffs, where a hot streak could lead to a World Series title, but a cold streak at the wrong time could lead to a swift elimination.