clock menu more-arrow no yes mobile

Filed under:

Astros’ Extra Inning Record & Playoff Update

Thinking about the Astros’ poor extra inning record and continuing to follow the Playoff Odds

MLB: Boston Red Sox at Houston Astros
Red Sox batter Adam Duvall celebrates a HR during the 10th inning against the Astros.
Troy Taormina-USA TODAY Sports

There ain’t no genius here. Strategy in baseball is overrated. People say, ‘That Weaver, he plays for the long ball too much.’ You bet I do. Hit ‘em out. Then I got no worry about somebody lousing up a bunt, I got no worry about the hit and run - and that’s really overrated - I got no worry about base-running errors. And I can’t screw it up myself. —Earl Weaver

The Astros have a poor record in extra inning games. Most Astros’ fans are aware of that fact. In this article, I’ll investigate possible reasons why teams have bad W/L records in extra inning games. In addition, because the Astros are embroiled in tense battles for a playoff spot, we will take another updated look at the playoff odds.


After Wednesday’s 10th inning loss tot he Red Sox, the Astros record for extra inning games is 1 - 8. Not good.

The Astros have not been exceptionally good in extra inning games since the “phantom runner” rule for extra inning games was put in effect. The Astros’ yearly record since that rule went into effect is shown below (2023 is up to this date):

2020 2 - 7

2021 9 - 8

2022 5 - 6

2023 1 - 8

The cumulative record in extra inning games is poor: 17 - 29. The Win Percent is notably weak, given that the Astros have been the reigning AL Champions for the time period subsequent to 2020. However, another way of looking at it is that the Astros’ extra inning record is poor in 2020 and 2023, but the Win % in 2021 and 2022 is closer to a central tendency around .500.

The Astros have the second to worst extra inning record this year. The bottom three records for extra inning games, so far:

Pirates 2 - 7

Astros 1 - 8

Padres 0 - 10

By the way, the two AL West competitors of the Astros haven’t exactly been great in extra inning games: Mariners are 6-12 and Rangers are 2-5 . In order to explore possible causes of good or poor extra inning performance, I examined whether other variables are correlated with teams’ extra inning Win% in 2023:

  • Run Differential (i.e., average run margin per game) is negatively correlated (-.028) with extra inning win percent, meaning that teams with better run differentials tend to have a worse extra inning record. However, this correlation is small and doesn’t come close to being explanatory.
  • Overall W/L % should reflect the team’s quality or winning capability. Does it correlate with extra inning win percent? A moderate correlation (.18) exists between overall W/L% and extra inning Win%. However, R^2 indicates that season Win% can explain only 3% of the variation in team extra inning wins. So, team quality, in terms of overall record, is only a minor influence on extra innings Win%. Extra inning games tend to end with a low run margin, which may somewhat explain this relationship. Previous sabermetric analyses suggest that the quality of the team is more likely to be revealed at higher run margins, as opposed to “close” games.
  • A reasonably good correlation (.59) exists between Pythagorean “luck” and extra inning Win%. Pythagorean luck on the season explains 34.8% (R^2) of the variation in extra inning Win%. The Pythagorean Record is a formula which generally predicts the season record based on run differential. Pythag “luck” represents the extent that a team’s actual season record is above or below the Pythagorean prediction. (The Astros are one run below their Pythag prediction, according to Baseball-Reference.)
  • A reasonably good correlation (.514) exists between a team’s one run game Win% and extra inning Win%. A team’s record in one run games explains 26% of the variation in extra inning Win%. (The Astros are 17 - 17 in one run games, according to Baseball-Reference.) Since extra inning games frequently end in a one run margin, it’s possible that this result reflects the contribution of extra inning games to the overall one run game record.
  • Does some common team characteristic exist which causes or contributes to Pythag luck, one run game Win%, and Extra Inning Win%? Maybe. But it’s a matter of speculation to imagine such a team feature. Another possibility is that the causation works the opposite direction: Extra Inning Win% may partially explain why a team under- or over- performs its Pythag.

It’s not for lack of trying that we can’t demonstrate what causes teams to over- or under-perform its Pythag. That subject has been studied extensively by analysts. But with very little to show for the effort. In a recent Dan Symborski Fangraphs chat (12:57), he humorously listed the meaningful variables with a blank, and said that good late inning closer is the only variable with a minor correlation with Pythag luck.

Bill James stated that winning one run games is mostly random, i.e., successful one run game Win% is frequently followed by a losing one run record in the succeeding season. An old Fangraphs article concluded that bullpen K/9 and BB/9 are two variables which have some explanatory power for one run game W-L records.

Considering that bullpen capability is frequently cited by analysts as an influence on Pythag performance and one run game success, I performed the same correlation analysis between Extra Inning Win% and various team bullpen statistics. My conclusions:

  • Reliever K/9 and BB/9 are negatively correlated (-.22 and -.19) with Extra Inning Win%. I’m not sure why K/9 is inversely related to Extra Inning Win%, unless it’s due to the correlation between reliever K/9 and BB/9 (.258). A strike out pitcher would seem to be a good thing in extra innings—unless the strike out capability is accompanied by higher liklihood of walks. The R^2 statistic indicates that the bullpen’s K/9 and BB/9 explains 5% and 4%, respectively, of the variation in Extra Inning Win%. That’s not a lot, but it’s something.
  • Reliever HR rate has a mild correlation (.14) with extra inning Win%, but the explanatory power is less than 2%. The remaining bullpen variables tested (batting average against, LOB%, ERA, FIP, SIERA, K-BB%, WHIP) have little or no correlation with Extra Innings.

These results seem to support randomness or luck rather than bullpen performance as the primary determinant of extra inning Win%.

Given the influence of the phantom runner in extra innings, it would be interesting if one could develop a set of variable to measure a team’s capability to advance runners. But I don’t have any stats to show players’ ability to hit ground balls to the first base side of the infield in important runner on 2d base situations. But it’s worth noting that all three of the most unsuccessful teams in extras (Pirates, Astros, Padres) are normally above average in their ability hit sac flies and sac bunts.

Finally it would be interesting to test the quality of a team’s bench hitters versus extra inning performance. But I don’t have any ready variables to examine that question. But it seems like bench strength might be important, given the role of pinch hitters and substitute players in extra innings.

In summary, I’m left with the unsatisfactory answer that maybe the Astros have just been unlucky in extra innings.

Update on Playoff Odds

As I have done in past articles, I will look at the Astros’ current probabilities for the playoffs. For the reasons stated in previous articles, I will rely on the Fangraphs daily playoff odds.

As of Aug. 25, the Astros show the best odds for winning the AL West. Although the Astros record has been lackluster over the last week, the Rangers’ losing streak has impressed the Fangraphs model the most. Seattle now has slightly higher odds than the Rangers for winning the AL West. All in all, the AL West division race is very tight and involves all three teams.

(Projected Wins, Division Odds, Make Playoffs)

Astros 90.9 Wins, 44% Division Odds, 82% Playoff Odds

Mariners 89.6 Wins, 28.2% Division Odds, 71% Playoff Odds

Rangers 89.8 Wins, 27.8% Division Odds, 71% Playoff Odds

The Fangraphs model shows the Astros with an easier strength of schedule for the remainder of the season than either the Mariners or Rangers (.494 Astros SOS vs. .502 Mariners and .518 Rangers). Despite that advantage, it appears that the Mariners have a near term easier schedule. As some have noted, potentially the Astros may fall behind the Mariners in the next week. But the schedule should still favor the Astros in later games.