In their ongoing effort to Make Life Better for Everyone, the good folks at Baseball-Reference.com have released an awesome downloadable spreadsheet that "that contains decade-by-decade records for teams when they enter an inning with a lead of 1 to 6+ runs (about 400 lines)."
Here's the sample, which lists the data that's probably the most interesting to many of us:
What's fascinating, of course, is that even as tactics have dramatically changed over the last 70 years, teams' winning percentages in these games -- games in which they enter the ninth (second column) with a one-run lead (third column) -- hasn't changed at all, sticking at 85-86 percent throughout (last column).
Now, the conclusion one might draw from this data is that modern bullpen tactics -- which is to say, the dedicated closer, the NINTH-INNING GUY -- are no more effective than whatever managers were doing from the 1940s through the 1980s, and further that managers today would fare just as well if they went back to the old ways.
I'm not sure I buy that. For one thing, the usual move in the 1940s was to just let the starting pitcher finish the game. You try that today, and the guy's going to be entering the ninth having thrown 130 pitches already. And I don't know anybody who thinks that would work well.
That said, you can reasonably argue that what managers did in the 1970s and early '80s -- routinely letting their ace relievers throw two or three innings -- worked just as well as what managers are doing today. But I will point out that the top relievers of that era, guys like Rollie Fingers and Bruce Sutter and Dan Quisenberry and Goose Gossage, really weren't top relievers for long, compared to guys like Trevor Hoffman and Mariano Rivera. Maybe you really can get great results from two-inning relievers who pitch 100-some innings per season ... but not for many seasons.
In the previous era, managers would bring their ace into the game at the first sign of late-innings trouble ... and then just keep them in the game. Which was, we might agree, rough on the pitchers in the long term. Really, what nobody's yet tried is bringing in the ace at the first sign of serious trouble ... and then taking them out of the game after three or four outs, even if the game's still close. On paper, that would give you the most wins.
But would it push your winning percentage up by even a full percentage point? And would your ace reliever hate you for using him in this somewhat haphazard fashion that results in many fewer saves than ace relievers are used to?
My guess is that while relief tactics aren't perfectly optimized, in the real world they're practically optimized. And you're just not going to do appreciably better than 86 percent, no matter what you try.