Lineups: The Sabermetric Way

Lineups: the Sabermetric Way

            When you think of a lineup how do you imagine it?  You put the fastest guy batting first, your number 3 hitter is the best overall hitter, 4^th is the best power and the 5-9 slots are the best to worst remaining right?  Well baseball has evolved with advanced stats leading the way in more than just scouting players and plotting trends on hitters and pitchers; lineups can now be analyzed and we can see what is the most effective lineups.  More importantly we can see what spot in the order is the most important.

            “To understand the impact of your possible choices, you have to understand the environment in which you are working.  Context.  Whenever you are trying to figure out what to do, take a step back, and ask yourself, “What’s the context?””  This is a quote form “The Book” by Tom Tango, Mitchel Lichtman and Andrew Dolphin.  This is a great quote to remember as we move forward.  Take a look at the chart coming up, this is a chart of the average number of PA given the position in the lineup.

Batting Order	AL Parks	NL Parks
1	4.83	4.80
2	4.72	4.68
3	4.61	4.56
4	4.49	4.46
5	4.39	4.35
6	4,26	4.23
7	4.14	4.10
8	4.02	3.98
9	3.90	3.86

As expected the person at the bottom of the lineup bats less than the guy on top, but we expected this.  Each batter getting roughly 2.5% percent more PA’s than the batter after him and with that logic we should have the best hitter on the team batting leadoff since they’ll acuminate more PA.  This is why I brought up the context quote, if this WAS the case we would see guys like Paul Goldschmidt and Robison Cano leading off not batting 3^rd.  Context.

            The next factor to look at is the number of runners potentially on base.  Baseball is a game of percentages; the way to win is to put the odds in your favor (insert your Hunger Games joke here).  While it seems obvious the lead off hitter is going to see the bases empty at a higher percentage than the rest of the team, roughly about 64% of the time in fact.  Right here we can see why you would never want to have your best hitter batting 1^st, he’d never see many men on base.  What position in the order sees the most batters on base?  According to tradition it’s the 3^rd batter; according to the stats it’s the 4^th and 5^th.  Odd how we put guys like Pujols and Cano batting when they statistically see batters less then that clean-up spot.

            So did anyone notice what I said above, that the 5^th batter is just as valued as the 4^th and more than the 3^rd?  Well its all do to a simple sabermetric stat called 24 base/out state.  To simplify it this the odds of a run in a specific scenario; example, the value of a single with a man on 1^st and 3^rd with 2 outs is 0.94 while the same scenario with 0 outs puts the value at 0.74.  The run became more valuable when the situation became more important IE the number of outs to play with decreased.  Now back to the 4^th and 5^th batters, I’m now going to show you 2 charts that have relevance the first is the runs value by event and batting order.

Batting Order	1B	2B	3B	HR	NIBB	HBP	K	Out
1	.468	.733	1.019	1.291	.350	.373	-.299	-.298
2	.478	.743	1.023	1.349	.340	.369	-.300	-.301
3	.469	.742	1.013	1.384	.319	.352	-.302	-.300
4	.504	.802	1.090	1.436	.337	.368	-.323	-.319
5	.513	.809	1.106	1.438	.348	.381	-.324	-.323
6	.494	.782	1.077	1.411	.344	.377	-.314	-.314
7	.489	.777	1.068	1.407	.340	.372	-.312	-.312
8	.488	.772	1.060	1.398	.337	.368	-.311	-.309
9	.485	.766	1.053	1.388	.336	.366	-.309	-.308

From this chart we can see that the runs value is higher with the 5^th batter than the 3^rd or 4^th.  The 5^th batter runs into more scenarios to score more often making him have higher run values overall.  But this chart doesn’t include a valuable detail; it doesn’t include PA in the formula and remember there’s a 2.5% difference from the batter to the man on deck so lets see how this affects the look of our chart.

Batting Order	1B	2B	3B	HR	NIBB	HBP	K	Out
1	.515	.806	1.121	1.421	.385	.411	-.329	-.328
2	.515	.799	1.100	1.450	.366	.396	-.322	-.324
3	.493	.779	1.064	1.453	.369	.369	-.317	-.315
4	.517	.822	1.117	1.472	.377	.377	-.332	-.327
5	.513	.809	1.106	1.438	.381	.381	-.324	-.323
6	.482	.763	1.050	1.376	.368	.368	-.306	-.306
7	.464	.738	1.014	1.326	.353	.353	-.296	-.296
8	.451	.714	.980	1.293	.340	.340	-.287	-.286
9	.436	.689	.948	1.249	.329	.329	-.278	-.277

            Notice the changes?  This is the chart that helps show the value of each event by each position in the batting order and according to the stats the most important and best hitters should bat 4^th, 5^th and 2^nd.  Didn’t see that one coming did you?  In fact the one spot in the lineup which most people use as the “get the runner over” guy is way more important value wise than the 3^rd batter commonly thought to be the best hitter.

            Now there is a lot more stats an info that can be used to further this argument but lets cut it here for now and ask another important question; why is Tony Gwyn Jr. leading off?  Or even more important why was B.J. Upton batting 2^nd for so long!  Coaches don’t seem to notice the data actually is pretty clear that these players shouldn’t be in these roles; Tony Gwyn Jr. owns a career .245 BA and a OPS+ of 75 (25% worse than the rest of baseball) and he lead off for the Phillies.  B.J. Upton had as many strike outs this year than most of the team combined, not counting how terrible he was last year and still batted second.  Coaches seem to want to hold to the old ways on lineup creation and not look at the stats.  While this isn’t the case for every team, the Houston Astros have been batting Jose Altuve 4^th; he owns a .257 SLG but Altuve is their best hitter and since his call up George Springer has been batting 2^nd and 4^th.  This is an example of an unconventional lineup creation but statistically sound.

            Every time I see Billy Hamilton batting lead off or 2^nd I shack my head; every time I see Xavier Nady bat 4^th I shack my head; every time I see Kole Calhon bat lead off I shack my head.  There are always new examples every day and it makes me wonder why this info isn’t being used!  Teams could be doing so much more with the lineup to get more production.