So I’m a baseball fan. I absolutely love baseball. I also love science and all the “Moneyball” type thinking that has emerged over the past 10+ years. I like to think about different things and recently, I’ve started wondering why teams don’t use a four man rotation anymore.
In a five man rotation (used by pretty much all professional baseball teams), starts are divided up as follows: 33 starts for the front four and 30 starts for the #5 guy (approx.). The fifth starter normally loses a few starts due to scheduled off days, in which he is skipped so teams can get back to the Ace. Guys with 31 or more starts averaged 199 IP last year (let’s call it 200). The average is approximately 6 1/3 innings per start.
Sooooo…. I wonder what would happen if teams switched to a four man rotation, leaving teams with eight relievers (or maybe seven if you decide to go with 14 position players). The starts should work out as follows: Two guys get 40 starts and two guys get 41 (your long guy will have to spot start any double headers). Pitchers still pitch the aforementioned 200 IP as to not over use their arms. That averages out to approximately 5 innings per start, which is still enough to qualify for the win. The bullpen would be set up as follows: One of your eight relievers is your closer, leaving you with seven relievers to cover the remaining 3 innings per day (486 innings). 486 innings spread across seven guys is 69 1/3 innings per guy. (Obviously your closer is only going to close games when required so don’t take these numbers as gospel. A closer on a good team should get 45 – 50 save opportunities per season.)
Now you’re getting more quality starts without really increasing the amount of innings on your starter’s arms. If you want to get back to your Ace a day sooner, you can skip the fourth starter after an off day, or you can opt not to skip him and let your guys get five full days rest. By doing this you’ve just eliminated 30 (probably) sub-par, starts from your fifth starter. Plus, your four remaining starters are no longer throwing “tired” innings (6, 7 and 8) which means they can cut loose for the five innings they do pitch, thus (hopefully) making them more dominant during those five innings.
New sabermetric-type stat: Adjusted Strike Outs per Inning Pitched.
Adjusted K’s per IP or “Adj. K/IP” is: (K’s-(HR+H+BB+IBB))/IP
In 2015, there were 117 starters with who pitched at least 120 innings. Their average Adj. K/IP was -0.52. By knocking off your #5 guy, who is essentially your weakest starter, you’re giving more innings to your better pitchers. We know that starters with minimum 31 starts averaged 199 innings pitched and had an Adjusted K’s per IP of -0.43, a 0.09 improvement from the average. The 28 starters who threw at least 200 innings averaged an even lower value of -0.28. Obviously these guys are being entrusted with these innings because they’re elite pitchers in the first place but the point is, the closer you can get to these guys, the better Adj. K/IP you’re going to get.
Relievers are a little harder to judge as many of them made spot start appearances, some more than ten. But if we look at the 166 pitchers who threw a minimum of 45 innings, the average Adj. K/IP was -0.34, an improvement (from the average starter’s value of -0.52) of 0.18. What does this tell us? It tells us that on average, relievers were noticeably more effective than starters.
But wait, there’s more!
Of the 166 relievers who pitched at least 45 innings, 70 of them had at least 10 holds*, and were thus more likely to be a regular big league reliever. This group had an Adj. K/IP -0.27, an improvement of 0.07 over our previous group of relievers (-0.34). If we create a crude average using our starter value of -0.52 and our reliever value of -0.34 we get -0.41 (five starters at -0.52 + seven relievers at -0.34 = 4.98. 4.98 / 12 = -0.41). If we create the same average using our new four man rotation and eight man bullpen, we get -0.32, an improvement of 0.09. Here’s an idea of the calibre of pitchers falling in to these groups:
- Johnny Cueto -0.40
- James Shields -0.41
- Taijuan Walker -0.41
- Ian Kennedy -0.42
- Yordano Ventura -0.42
- Carlos Martinez -0.31
- Tyler Clippard -0.31
- Drew Pomeranz -0.31
- Clay Buchholz -0.32
- Adam Wainwright -0.32
- Kyle Hendricks -0.32
- Vince Velasquez -0.32
- Justin Verlander -0.33
- Carlos Villanueva -0.32
At face value these two groups look very similar, but let’s look more closely at the stats:
Group One vs Group Two
4.05 ERA 3.23
.250 AVG .233
.311 OBP .294
1.252 WHIP 1.172
1.12 GO/AO 1.04
8.66 K/9 8.26
3.29 K/BB 3.34
16.19 P/IP 15.92
Group two are more effective in every category except K’s/9 and GO/AO (ground outs : air outs). This may not seem like much but over a 162 game season these tiny differences can have a substantial impact on a team’s performance. Let’s pretend that these groups represent two actual pitchers and let’s look at their respective WHIP’s as an example. Every MLB team plays 162 games, a total of 1,458 innings (yes, when teams are winning at home they will not bat in the bottom of the 9th but let’s just pretend for a moment). Over 1,458 innings our Group One pitcher’s WHIP equates to 1,825 baserunners while our Group Two pitcher’s WHIP equates to 1,708 baserunners, a difference of 117 baserunners. In 2015, approximately 35% of baserunners ended up scoring so of these 117 extra baserunners, we can expect about 41 of them to score. It may not seem like much but, as with everything in baseball, it all adds up.
- Starter is hooked after the 5th inning unless he’s working a no-hitter or a perfect game.
- Away games: Starter does not pitch the bottom of the 5th if down by 3+ runs at end of the top of the 5th.
- Home games: Starter does not pitch the top of the 5th if down by 3+ runs at end of the 4th.
What’s that you say? These pitchers will no longer be likely to win the Cy Young award? Our top 28 pitchers averaged 212 1/3 innings pitched. Over the 41 starts that your Ace is likely to get (using the four man format) pitchers would throw 205 innings. They can be allowed to pitch past the 5th if they have a no hitter or perfect game so getting up to 220 innings is more than likely for most of these pitchers. This shouldn’t be too taxing on their arms as pitchers rarely carry a no-no or perfecto past the 6th or 7th inning. Thus they are likely to accrue the same aggregate number of innings pitched which would still enable them to put up the same type of numbers as when pitching in a traditional five man rotation.
*A hold is awarded to a relief pitcher who meets the following three conditions:
- He enters the game in a save situation; that is, when all of the following three conditions apply: (a) He appears in relief (i.e., is not the starting pitcher); and
(b) He is not the winning pitcher; and
(c) He qualifies under one of the following conditions:
- (i) He enters the game with a lead of no more than three runs and maintains that lead for at least one inning
- (ii) He enters the game, regardless of the count, with the potential tying run either on base, or at bat, or on deck
- (iii) He pitches for at least three effective innings.
- Records at least one out
- Leaves the game before it has ended without his team having relinquished the lead at any point and does not record a save.