Posted by Neil Paine on July 15, 2010
Over the years, many users have written to us expressing confusion over the assignment of saves for seasons prior to 1975, the year the current save rule was put into effect. This post will hopefully set the record straight on what changes the rule has seen over the years, which version of the rule we apply to older seasons, and why the save totals on a player's main page might not always match those you find in the player's game logs and splits.
Evolution of the Save Rule
As most of you know, there was no official statistic for saves prior to the 1969 season. But with relief specialists becoming more and more common during the 1960s, Chicago Sun-Times and Sporting News writer Jerome Holtzman successfully lobbied MLB to add the category to their official stats following the 1968 season. The original save rule, Rule 10.20, went into effect with three provisions:
- The relief pitcher must enter the game with a lead and hold it.
- He can be taken out of the game only for a PH or PR.
- If more than one pitcher qualifies, the scorer gives it to the "most effective" one.
The first change to the rule came after the 1973 season, in an effort to simplify matters:
- The relief pitcher must face the potential tying or winning run at the plate or on base, or pitch effectively for at least three innings, and in either case preserve the lead.
- Same as #3 above, one save per game, given to the most "effective" reliever.
This rule was only in effect for the 1974 season. Note that the stipulation about leaving for a PH or PR was removed, eliminating the requirement that the reliever must actually finish the game to earn a save.
In 1975, the rule was changed for a second and final time:
- The reliever must be the finishing pitcher in a game won by his club.
- He cannot be the winning pitcher.
- He qualifies under one of the following conditions:
- He enters the game with a lead of no more than three runs and pitches for at least one inning; or
- He enters the game, regardless of the score, with the potential tying run either on base, at bat, or on deck; or
- He pitches for at least three innings.
The basics of the 1973 rule were retained, except the reliever was once again required to finish the game to earn the save, the tying or winning run could also be on deck when he entered the game, the subjectivity of choosing the "most effective" reliever was removed, and a third way to earn a save was added (pitching for at least one inning with a lead of no more than 3 runs).
Application of the Rule(s) at Baseball-Reference
Now that we have box scores for every game since 1920, it creates many opportunities to retroactively expand our statistical knowledge about older seasons. However, with that comes the reality that we must make judgment calls for categories like the save, which didn't exist for most of baseball history and was defined differently at various times even after it became an official stat.
Our policies are the following, based on Retrosheet's procedures:
- For 1920-1949, saves are awarded on the "encyclopedia basis": a pitcher who finished a game his team won, but did not get the win himself, is awarded a save. This policy is so named because it was used by the first Big Mac encyclopedia and later by Pete Palmer.
- For 1950-1968, the policy is to use the 1969 rule (the first official one) and apply it retroactively. However, David Smith of Retrosheet admits that it hasn't been applied as rigorously to these seasons as the encyclopedia rule was applied to 1920-49, so there may be instances where saves were or were not awarded precisely according to the 1969 rule.
- For post-1968 seasons, the rule in place at the time is used.
Why Do Full-Season Stats Not Always Match the Game Logs/Splits?
We have two competing sources for our data: Full-season data comes from Pete Palmer, and the game-by-game data comes from Retrosheet.org. For some stats these are not always going to match up -- this goes for every stat on the site, but especially the save rule, where Retrosheet has in some cases applied a different rule than Pete did. David Smith will be reviewing the 1950-68 seasons this summer and will make the appropriate changes to the game-by-game stats, which will eliminate some of the discrepancies, but there will likely always be some differences between the datasets simply because of the nature of each one.
Hopefully this post has cleared up some of your confusion regarding our retroactive application of the save rule for seasons in which the stat was not originally tracked. But if you still have questions or comments, feel free to email us at email@example.com.