On the “Craziness” of a Four Pitcher Limit

Yesterday, Dave put forward a proposal about how Major League Baseball could possibly improve its pace of play and run scoring in one fell swoop: limit the number of pitchers allowed per game to four. He couched it by saying that it was an admittedly crazy idea. But after compiling a grid of how many pitchers are used per game, I’m not so sure that it is.

What I wanted to see is a grid of how each team used its pitchers. How many games with two pitchers, three pitchers, etc. Thankfully, Baseball-Reference’s pitching game logs are very accommodating in this regard. In order to get a representative sample, I scrubbed out extra inning games, as well as games that were shortened for some reason (most likely rain). That leaves just the games where the pitchers threw eight to nine innings. Now, there’s certainly a chance that there was some weird game that was stopped for rain after eight innings, but barring that, this should be a sample of all the “regulation” games from last season. No team had fewer than 141 of these games, and no team had more than 154. Most of the games removed were extra-inning games, there were just a handful of shortened games.

Enough talk, let’s get to the grid:

Number of Pitchers Used Per Game, By Team, 2014
Team 1 2 3 4 5 6 7 8 9 8-9 inn games % 4 or less % 5 or less
CLE 6 11 23 38 40 13 9 1   141 55.3% 83.7%
BOS 3 14 40 50 27 6 1     141 75.9% 95.0%
BAL 2 14 46 44 24 11       141 75.2% 92.2%
PHI 2 22 47 43 23 3 1     141 80.9% 97.2%
OAK 7 16 48 41 25 4       141 79.4% 97.2%
LAD 6 23 26 40 34 8 4 2   143 66.4% 90.2%
MIA 3 15 35 54 26 10       143 74.8% 93.0%
TB 3 16 36 47 29 12 1     144 70.8% 91.0%
LAA 3 15 37 35 35 9 9 2   145 62.1% 86.2%
WAS 5 16 53 45 20 5 1     145 82.1% 95.9%
CHC 1 4 46 50 30 12 3     146 69.2% 89.7%
NYY 4 17 38 48 29 9   1   146 73.3% 93.2%
ATL 5 10 43 58 21 7 2     146 79.5% 93.8%
NYM 1 13 43 57 27 4 2     147 77.6% 95.9%
SD 4 8 47 53 29 4 2     147 76.2% 95.9%
CHW 2 16 48 49 26 6       147 78.2% 95.9%
ARI 2 14 40 48 35 9       148 70.3% 93.9%
PIT 2 17 53 46 23 7       148 79.7% 95.3%
COL 1 6 42 39 41 15 5     149 59.1% 86.6%
MIN 2 13 42 50 33 8 1     149 71.8% 94.0%
STL 8 11 36 52 31 9 2     149 71.8% 92.6%
MIL 3 12 42 61 20 9 2     149 79.2% 92.6%
SF 8 19 38 39 30 12 2 1   149 69.8% 89.9%
TOR 3 24 43 43 25 9 2     149 75.8% 92.6%
CIN 5 30 39 45 26 2 2     149 79.9% 97.3%
KC 3 18 52 45 19 11 1   1 150 78.7% 91.3%
SEA 2 22 36 39 36 11 4   1 151 65.6% 89.4%
DET 5 17 44 40 30 11 3 1   151 70.2% 90.1%
HOU 7 21 42 49 21 10 1     151 78.8% 92.7%
TEX 5 14 34 58 28 14   1   154 72.1% 90.3%
TOTAL 113 468 1,239 1,406 843 260 60 9 2 4,400 73.3% 92.5%

Do these results surprise you? I have to be frank (well, I have to be Paul, but you know what I mean), these surprised me a little. I figured the percentage of games where teams used more than four pitchers would be a lot higher. But just under three-fourths of games last season ended at the four-pitcher mark that Dave proposed. I also included five or fewer as a percentage as well, and that’s where I am especially encouraged. Fewer than eight percent of games played last season featured a team using more than five pitchers — 331 in all. And 96 of those, or 29 percent, came during September, when rosters expand. Now, not necessarily every game where managers use six or more pitchers in regulation in September can be attributed to the expanded rosters, but we’ll simplify and blame it on that anyway.

Whether you work off that lower 235 number or the 331 number, that’s not a lot of games where managers really would have had to change strategies in order to be in accordance with this new rule change. That’s somewhere between seven and 11 games per team when managers really would have had to squirm. That doesn’t really seem like that much, right? Make sure one pitcher in the bullpen can go really long, and otherwise your bullpen is unaffected. Perhaps that’s an oversimplification, but it’s probably not too far off the mark. And there are other potential benefits — perhaps teams could trim their bullpen to 11 pitchers overall, freeing up a spot for another bench bat.

The real key though is those games were managers used five pitchers. Could they have used one fewer pitcher in those games? That’s where the rubber meets the road with this rule change. Looking back at the above table and cross-referencing with the total number of relief games, there are encouraging and discouraging signs. The first encouraging sign is the Cleveland Indians. Manager Terry Francona is one of the brightest managers in the game. Last season, he tried to make lemonade when his starting rotation didn’t pitch as expected in the first half, and consequently his team had more relief games than anyone else. Corey Kluber did just swell, but no other starter posted a first-half ERA below 3.83.

As such, Francona worked his relief corps pretty hard. But in the second half, the rotation picked up things in a big way, and the Indians’ percentage of games finished with four or fewer pitchers improved dramatically. In the first two months of the season, just 52 percent of the team’s games finished with four or fewer pitchers seeing action. In the final two months, that jumped to 64 percent. So, if you hold to the belief that the Indians’ starting rotation will look more its second half iteration than the first, there is a big chunk of games that could come back on the side of the ledger we’d like to see.

One discouraging sign is the Rockies. Seemingly perpetually trying a new pitching strategy, the Rockies might have to pivot more than most teams under this rule change. It’s simply a lot harder for them to sustain durable performances out of their starting pitchers. To wit:

Number of Starters Pitching 6+ Innings, 2010-2014
Team Games
PHI 589
CIN 569
CHW 568
OAK 546
LAA 545
ATL 542
NYM 542
DET 541
STL 541
SF 537
TB 529
SEA 527
BOS 524
ARI 522
LAD 522
MIL 519
NYY 519
KC 505
WAS 498
TEX 497
SD 496
MIA 494
TOR 493
HOU 489
CHC 486
PIT 471
CLE 467
BAL 454
MIN 454
COL 394

Now, part of the reason why the Rockies are so far beneath every team is their ill-fated 2012 experiment. That season, only 36 pitchers worked six or more innings. But even if you double that total, which would bring it up closer to 2013’s 77 games total, they’re still in last place, and are hopelessly far from the average (512). Of the 150 team seasons in the past five years, by this metric the Rockies ranked 97th, 123rd, 136th, 147th and 150th (that’s 2012, obviously). The problem, as you probably know, is altitude. Even when pitchers are going right, sometimes they just can’t make it deep into the game. Under this rule, the Rockies would have to be more laser focused than most on getting pitchers who limited walks. And one look at their team history shows that they have never really had a lot of starters who could do that, because, again, the altitude.

In nearly 93 percent of the games played last season, five or fewer pitchers took the mound for their team. Could teams work those 19 percent of games where they used five pitchers down to four? Is that reasonable? It certainly doesn’t seem unreasonable. Teams would save time in those 19 percent of their games, and likely much more, because managers would be a lot more careful about pitching changes in the 71 percent of games where they used two, three or four pitchers. This time savings should far outweigh any potential messes created from leaving in a pitcher who isn’t up to snuff in the other eight percent of games where teams had used six or more pitchers. As Dave pointed out yesterday, this rule change would need a couple of loopholes (injury, extra innings) but other than that it seems like a great compromise — it would speed up games while still allowing managers the freedom to make the changes they saw fit. And the added intrigue from trying to think along with the manager, knowing that he does have a finite number of pitching changes left in each game would make watching the games that much more entertaining.

We hoped you liked reading On the “Craziness” of a Four Pitcher Limit by Paul Swydan!

Please support FanGraphs by becoming a member. We publish thousands of articles a year, host multiple podcasts, and have an ever growing database of baseball stats.

FanGraphs does not have a paywall. With your membership, we can continue to offer the content you've come to rely on and add to our unique baseball coverage.

Support FanGraphs




Paul Swydan used to be the managing editor of The Hardball Times, a writer and editor for FanGraphs and a writer for Boston.com and The Boston Globe. Now, he owns The Silver Unicorn Bookstore, an independent bookstore in Acton, Mass. Follow him on Twitter @Swydan. Follow the store @SilUnicornActon.

newest oldest most voted
James Fan
Guest
James Fan

Good stats article. Why not make it five pitchers or less per game? Granted, that’s just an incremental change and not a big bang, but impacting 25% plus of all games is a fairly large number. If you go to a five pitcher limit, you are impacting around 8% of games directly and a number of them indirectly – which may be more palatable.

Hurry Up!
Guest
Hurry Up!

I still believe that the best way to shorten games is to limit the number of trips to the mound – from the bench or from the field. make it, say…5. Anything after 5, and the pitcher has to be removed and it’s a balk. No more 5 pitcher/catcher conferences per inning. No more trips to the mound from the bench or catcher just to allow a reliever to get ready. The pitcher and catcher should be discussing how to handle a particular hitter before a game or between innings, not in the middle of an at-bat. Most visits are just delay tactics, anyway.