JP Caillault wrote:

Sorry to bother you with this, but in 1891 Mickey Welch pitched 14 CGs, of
which he won 5 and lost 9. The MLB.com, baseball-reference.com, and
retrosheet.org websites, as well as David Nemec's 19th century encyclopedia
and TB8, all have his record as 5 and 9.

Daguerreotypes (1968) has his record as 6 and 9.

Here're the relevant parts of the account of the game in question (from The
New York Times via ProQuest):

"Five pitchers took part in to-day's game, and the New Yorks won after an
interesting struggle. Clarke was injured in the opening inning and O'Rourke
was called in to catch. He couldn't hold Rusie's fast delivery, and Welch
was put in. ... In the third inning Gore hit for two bases, [go through most
of NY batting order until] Whistler, Bassett, and Welch followed with
singles ... In the eighth Welch was handled pretty roughly and Sharrott took
his place. ..."

Here's the line score:

NY - 0 0 6 3 0 0 0 0 2 - 11
Phi - 0 0 1 1 0 0 3 3 0 - 8

So, Welch came in VERY early, when it was still 0-0, and stayed until it
was, at worst, 9-8. Surely Rusie cannot have been credited with this win.
Was Sharrott given the win because he put out the fire in the 8th and shut
out the Phils in the 9th, i.e., what we'd call a save today? I know that is
similar to the reasoning Frank Williams applied in his famous article about
1900-1920 W-L decisions, but, well, I thought that only applied to
1900-1920.

And, finally, since W-L weren't such a big deal back then, and, hence,
authoritative decisions weren't made contemporaneously, who is it that
ultimately decided (or decides) who gets the W?

 

Dan Holmes replied:

For what it's worth, Elias, the official statistician of MLB, has Welch's 1891 record as 6-12.

 

And Jim Smith followed up:

 

Dan: The 1892 Spalding Guide showed Welch with 5-11 for the 1891 season. He was 1-4 vs Boston, 2-1 vs Brooklyn, 0-2 vs Cincinnati, 0-1 vs Cleveland, 0-3 vs Philadelphia and 2-0 vs Pittsburgh. The Reach guide didn't have a club breakdown, but showed Welch with Six (6) and didn't show any number of losses. See page 32 in Spalding & page 69 in Reach.

 

And "hatchersj" wrote:

 

The first thing we ought to do is determine who was awarded the
victory in such cases. If the award was proper under the rules and/or
practices of the day, then it should not be changed. Revisionism is
not what we should be all about.