Sports Reference Blog

From time to time Baseball-Reference receives updates to the historical record from trusted sources. These are transmitted to us via the Chadwick Baseball Bureau, which obtains updates from the legendary sports researcher Pete Palmer.

Generally these sorts of updates are most numerous when our friends at Retrosheet release new seasons worth of data. However, earlier this month we loaded a smaller update that deserves a mention and some further reflection on our part on how to best proceed.

First some quick background: The sacrifice fly was first listed separately in statistics in 1954 (and has been every year since). Before this, while the sacrifice fly never constituted its own statistical category, it was accounted for in a variety of evolving ways.

Per John Schwartz's research at SABR (linked above), 'sacrifice hits' first appeared in baseball box scores in 1889. Here's a quick timeline from 1889 to 1954:

1889-1893: Sacrifice hits are credited for advancing runners on bunts, ground outs and fly balls. Sacrifice hits count as at-bats.

1894-1907: Sacrifice hits are only credited on bunts and no at-bat is charged.

1908-1925: Sacrifice hits are credited on bunts and also on fly balls (if a baserunner scores). No at-bat is charged for sacrifice hits.

1926-30: Sacrifice hits are credited on bunts and also on fly balls that advance a runner (including advancing to 2nd or 3rd base). No at-bat is charged for sacrifice hits.

1931-1938: Sacrifice hits are only credited on bunts and no at-bat is charged.

1939: Sacrifice hits are credited on bunts and also on fly balls (if a baserunner scores). No at-bat is charged for sacrifice hits.

1940-1953: Sacrifice hits are only credited on bunts and no at-bat is charged.

1954-Present: Sacrifice hits are only credited on bunts and no at-bat is charged. Sacrifice flies are credited separately from sacrifice hits on fly ball outs that score a run. Neither sacrifice hits nor sacrifice flies count as an at-bat, but a player with no ABs can have a hitting streak end if he has a sacrifice fly, but not if he has no ABs with a sacrifice hit. This distinction is notable in that it was one way that the two types of sacrifices were deemed different when it came to counting other statistics.

In 1979, through his work with the Sports Information Center, Pete Palmer introduced On-Base-Percentage as an official American League statistic. The original version of this statistic treated sacrifice flies just like sacrifice hits in that neither of them were used in calculating OBP. However, when the National League introduced the statistic in 1984 it began using sacrifice flies in the denominator for calculating OBP. It was published in the Sporting News guides and gained traction as the official way to calculate OBP in years since. In the interest of having a uniform definition for the statistic, Palmer eventually began using SF in the denominator of the OBP calculation, as well.

Now, back to the recent update: Thanks to game accounts unearthed by the amazing group of researchers behind Retrosheet, Palmer was able to split out sacrifice flies from sacrifice hits for the 1924 to 1930 seasons, and also the 1939 season. Since sacrifice flies count towards OBP (and sacrifice hits do not) this reclassification of some sacrifice hits into sacrifice flies impacts OBP and OPS for players who played during those seasons. These changes went live on Baseball Reference earlier this month and we're now attempting to decide if the changes should stay. For now, we are rolling the statistics back to how they previously appeared.

A good example of the potential change is Joe Cronin in 1930. Previously, we credited Cronin with 22 sacrifice hits, a .422 OBP and a .934 OPS that season. New data indicates that 19 of Cronin's 22 sacrifice hits that season were in fact sacrifice flies, which adds 19 to the denominator of his OBP that season, lowering his sacrifice hit total to 3, his OBP to .410 and his OPS to .923. As an aside, I'll note that the "official" record for SF in a season is 19 by Gil Hodges in 1954 so Cronin represents one of the more extreme cases.

What we need to decide is how to handle these changes. By the official statistical convention since 1984, these sacrifice flies would justifiably count as outs for OBP purposes. Though this logic only came to prominence in 1984, it's how we calculate OBP back to 1954 when sacrifice flies first became official. However, it is a little hairier trying to decide if we should also treat sacrifice flies pre-1954 in the same way. The decision would be simpler if we could apply this logic uniformly to all seasons, but the constant changes to rules around sacrifices makes this something we want to think on for a bit. This is something we will work out in the near future, but for now we just wanted to make our users aware of the situation, especially if you're confused by any numbers you noticed changing recently.

If you have any strongly held opinions on this matter, we encourage you to please comment on this post. Thank you.

This entry was posted on Tuesday, January 31st, 2023 at 5:17 pm and is filed under Announcement, Baseball-Reference.com, Data. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.