Winning Teams and Losing Streaks

The Tigers are having a good season - they’re 60-43 with 97% odds of reaching the postseason. But they’ve had a rough stretch - they’re current 1-9 in their last 10 games, including a 6 game losing streak. While I’m very certain they’ll play better than 0.100 ball the rest of the season, it does beg the question of how common this sort of stretch is for winning teams.

To answer this, I downloaded daily game data for each team from 2004 - 2024, using the baseballr package. This gets data from baseball-reference the conviently includes streak length data, which saves some time calcing that ourselves in R. I’d have gotten more seasons, but getting the data is a bit of a pain and 20 years should suffice, especially since we’re more interested in recent history.

First, let’s look at each team’s longest losing streak and plot their win percentage.

library(hrbrthemes)
library(gt)
library(gtExtras)
library(tidyverse)
library(zoo)

team_seasons <- read_csv('mlb_team_seasons.csv')

team_seasons_agg <- team_seasons %>%
  mutate(w_flag = if_else( gsub('-wo', '', Result) == 'W', 1, 0) ) %>%
  group_by(Tm, season) %>%
  mutate(ten_game_win_per = rollmean(w_flag, k = 10, fill = NA, align = "right")) %>%
  summarise(games = n(),
            longest_w_streak = max(Streak),
            longest_l_streak = abs(min(Streak)),
            wins = sum(w_flag),
            worst_ten_game_win_per = min(ten_game_win_per, na.rm = TRUE),
            best_ten_game_win_per = max(ten_game_win_per, na.rm = TRUE)) %>%
  mutate(win_percent = wins / games)

team_seasons_agg %>%
  ggplot(aes(x = factor(longest_l_streak), y = win_percent))+
  geom_jitter()+
  geom_boxplot(outlier.shape = NA)+
  theme_ipsum()+
  xlab('Longest Losing Streak')+
  ylab('Win Percent')+
  ggtitle("Teams with Longer Losing Streaks Aren't As Good",
          subtitle = "A Shocking Development")

This is perhaps the world’s least surprising chart - of course bad teams are going to have longer losing streaks. If they were shorter, they’d be better teams. There’s still a bit of a spread, though - teams whose longest streak was 6 can be as low as 0.365 (the 2024 Rockies) and as high as 0.654 (the 2019 Dodgers). The mean and median are both just a hair over .500. It would take a catastrophe for the Tigers to be the bad.

But as the current stretch shows, longest losing streak isn’t necessarily the whole story. Being 1-9 is brutal, but there is a win in there. We can look at a team’s worst 10-game streak and compare to their final win percentage to check that angle.

team_seasons_agg %>%
  ggplot(aes(x = as.factor(worst_ten_game_win_per), y = win_percent))+
  geom_point()+
  geom_boxplot(outlier.shape = NA)+
  theme_ipsum()+
  xlab('Worst Ten-Game Streak')+
  ylab('Win Percent')+
  ggtitle("Better Teams Avoid Long Losing Streaks",
          subtitle = "But good teams hit them, too")

The 2017 Dodgers had an 11 game losing streak and finished with a 0.642 win percentage. We see plenty of teams with pretty bad streaks finishing above .500, though the more you limit this the better - all teams who managed to go no worse than 0.400 in a 10-game stretch finished above .500. Let’s look a little closer as the summary stats.

team_seasons_agg %>%
  group_by(worst_ten_game_win_per) %>%
  mutate(over_500 = ifelse(win_percent >= .500, 1, 0)) %>%
  summarise(team_seasons = n(),
            over_500 = sum(over_500),
            average_win_percent = mean(win_percent),
            median_win_percent = median(win_percent),
            min_win_percent = min(win_percent),
            max_win_percent = max(win_percent)) %>%
  gt() %>%
  gt_theme_espn() %>%
  cols_label(worst_ten_game_win_per = 'Worst 10-Game Win Percent',
             team_seasons = 'Team Seasons',
             over_500 = 'Over .500',
             average_win_percent = 'Average',
             median_win_percent = 'Median',
             min_win_percent = 'Min',
             max_win_percent = 'Max') %>%
  fmt_number(columns = c(worst_ten_game_win_per, average_win_percent, median_win_percent,
                         min_win_percent, max_win_percent), decimals = 3) %>%
  tab_spanner(label = 'Final Win %',
              columns = c(average_win_percent, median_win_percent,
                          min_win_percent, max_win_percent)) %>%
  data_color(
    columns = c(average_win_percent, median_win_percent,
                min_win_percent, max_win_percent),
    method = "numeric",
    palette = "Blues"
  )

Worst 10-Game Win Percent	Team Seasons	Over .500	Final Win %
			Average	Median	Min	Max
0.000	76	4	0.412	0.414	0.253	0.642
0.100	222	67	0.469	0.469	0.290	0.636
0.200	208	149	0.526	0.531	0.401	0.660
0.300	93	87	0.569	0.574	0.463	0.654
0.400	11	11	0.635	0.623	0.584	0.685

Since the Tigers are in the 0.100 bucket right now, we’re most interested* in that. 67 of the 222 teams who had a 1-10 stretch finished above .500, with the 2016 Cubs having one such stretch and finishing with a .636 win percent and winning the World Series. While this doesn’t mean the 2025 Tigers are guaranteed the same results, in assuring to see that teams can go all the way even with bad stretches like this. I’m naturally concerned, but trusting the team’s talent and a regression bounceback for the remainder of the season.

*Well, I’m most interested, anyway. You may be more interested in other things, and that’s fine. You’re an adult.