clock menu more-arrow no yes mobile

Filed under:

Do Opening Streaks Matter?

Yesterday I touched on the idea whether sweeping the opening series indicated how well a team did at the end of the year. It seemed like there was some indication, though the effect was greatest with teams that started the season with a 5 game or more win streak. Jim asked whether losing would act in a similar manner? Follow me after the jump and we'll take a look.

If we can hypothesis that an above average team is more likely to be good enough and consistent enough to go on a win streak, then we can also hypothesis that a below average team is more likely to suffer from a losing streak. A more detailed study on the mechanics and propensity of losing streaks is work to be done at a further date, but for now I'll assume that an opening losing streak is the indicator of a bad team.

To summarize yesterday's framework, I created a sample of 103 teams that started the season with a 3 or more game losing streak from 1980-2011. Again, I avoided the 1981 season for strike-related shenanigans. Like before, I noted which teams had more than a 3 game losing streak with +1 or more.

For a null hypothesis I again used .500 ball. Anything below .500 is guaranteed to miss the playoffs, and they're likely a bad team to boot. The t-test will determine how dissimilar my sample is from the null hypothesis, with an extreme finding suggesting (while assuming certain things I won't bog you down with) that a losing streak is an indicator of final season results.

So without further ado I present the table:


And now math! For the set ALL the statistically significant range would be t > -2.626, which the set obviously exceeds. For the set STRK3-4 the range would be t > 2.654, and this case we fail to reject the null hypothesis. There is not a statistically significant relationship between starting the season with a 3 or 4 game losing streak. Even if I used a less stringent test (in this case I'm assuming a one tail probability of .005, which is very stringent) it still would not be statistically significant. For set STRK > 4 the range would be t > 2.712, so the result is statistically significant.

Okay, math nerd, but what does it all mean?

If you start your season with a losing streak of more than 4 games, you're probably in a for a bad year. But if you start your season with only a 3 or 4 game losing streak, it means virtually nothing. You might be the 1998 Yankees, who went on to win 114 games. Part of this can be explained by how bad the teams that lose 5 or more in a row. These teams would start the season with a long losing streak and never recover, because they were inherently bad teams.

Losing 3 games in a row seems to be less damaging than winning 3 in a row is rewarding. I don't know if this result would continue to any streak in any part of a season, but it's suggests that winning might have some kind of particularly significant influence on baseball teams. Psychology is not my forte, however, so I'll leave the speculating to the Snakepit, assuming you've made it this far and didn't run screaming to the hills after maths.