Early Exits

[DrungoHazewood]

Another thread inspired by a Bill James online article (subscriber only). Another thread that I may be the only person here who cares, but I think it's cool so I'll post anyway.

He did a little research bit based on retrosheet data on pitcher useage patterns. Not the first time this has ever been discussed, but James quantifies the practice of yanking a starter when he just didn't have his stuff on a particular day.

My general feeling is that today a starter will typically:

a) get through five innings unless he's really being bombed.

b) get yanked sometime between the 6th and 7th unless he's really cruising.

In the 50s and 60s things were a bit different. Starters, good, bad or indifferent were often:

a) pulled after a couple innings if they'd given up 2, 3, 4 runs.

b) left in for 10 or 12 innings if they were throwing well.

James went through the numbers and proved it:

 		Unexplained	       Exit Score		Early		       Less than	Decade	Starts	Departures       %       30	%1950s	18362	2072	        11.30%	1113	6.10%1960s	31730	3006	        9.50%	1473	4.60%1970s	39560	2764	        7.00%	1259	3.20%1980s	40674	2302	        5.70%	938	2.30%1990s	43188	1500	        3.50%	660	1.50%2000s	48582	1147	        2.40%	620	1.30%

"Unexplained Early Departures" and "Exit Score Less Than 30" were just two ways of defining a quick shower for the starter. Either way you look at it, a starter in the 1950s was about five times more likely than someone today to get pulled after a two-inning, three run start.

There's also another chart in the article showing that really long outings have gone down from just over 11% of starts in the 50s to 0.4% today. But I think we already knew that.

So whenever you start talking about the pitchers from days gone by completing all their starts, you also need to remember that about once every week and a half a team yanked their starter after 30 or 40 pitches. That almost never happens today. If Jeremy Guthrie gives up four in the first chances are he's going six or seven innings anyway.

And just for fun, I looked up Whitey Ford, and his gamelogs from 1955 when he led the AL with 18 wins. He had 18 complete games in 33 starts. And despite never allowing more than five runs in any game, he was pulled from four starts in the 4th inning or earlier. He also had six relief appearances and two saves.

[Frobby]

It is interesting that the number of early exits has gone steadily down, regardless of variations in runs per game over the years.

[DrungoHazewood]

It is interesting that the number of early exits has gone steadily down, regardless of variations in runs per game over the years.

I had a lot of thoughts I was trying organize, but I think it comes down to baseball being more and more professionalized, which leads to standardization. I think those forces are a big glacier smashing the little rocks, which represent year-to-year or decade-to-decade run context.

[RShack]

I had a lot of thoughts I was trying organize, but I think it comes down to baseball being more and more professionalized, which leads to standardization.

Right. When Whitey Ford got yanked in the 2nd inning, he was good to come out of the BP over the next couple games. But if Guthrie gets yanked early, the chances of using him as a RP before his next start are about zero.

Instead of a SP's early exit being a double-edged sword, something to help the chances of still winning today's game and something that would enable the SP to help the BP a day or two hence, it's viewed as just a very one-sided thing that's nothing but negative: killing the BP. So, when a SP has a bad day, he gets left in too long, and that game gets sacrificed. There's not enough recognition of something that seems very obvious: P's just have bad days... and if he's having a bad day on Wednesday, maybe he'll have a good day on Friday, so try him then and see...

[DrungoHazewood]

Right. When Whitey Ford got yanked in the 2nd inning, he was good to come out of the BP over the next couple games. But if Guthrie gets yanked early, the chances of using him as a RP before his next start are about zero.

Instead of a SP's early exit being a double-edged sword, something to help the chances of still winning today's game and something that would enable the SP to help the BP a day or two hence, it's viewed as just a very one-sided thing that's nothing but negative: killing the BP. So, when a SP has a bad day, he gets left in too long, and that game gets sacrificed. There's not enough recognition of something that seems very obvious: P's just have bad days... and if he's having a bad day on Wednesday, maybe he'll have a good day on Friday, so try him then and see...

But you also have the advantage of having a good pitcher in a game that he'd otherwise be out of. Most of the time leaving Guthrie in after a bad inning early on wins you more games than bringing in Mark Hendrickson to pitch four innings. It's not just killing the pen, it's subbing in your 12th-best pitcher for one of your top 5 or 6.

I think this is also a realization that a lot of the time "he just didn't have it today" was shorthand for "he's pitching the same as always but in the 2nd inning the other guys still scored three runs." Unless he's hurt you're better off leaving the starter in.