Great Sinkers of Our Time

Part One: In which we set the stage
Part Two: In which PECOTA takes the spotlight
Appendix: List of sinkerballers and control group

We have thus far examined the careers of pitchers who possess a statistical similarity to Chien-Ming Wang. Now it's time to get to the crux of the matter: today we're taking a closer look at sinkerballers.

First, we need to identify sinkerball pitchers throughout history. To do this, I used Lee Sinins's Complete Baseball Encyclopedia to make a list of every pitcher with at least 1500 innings and a positive Runs Saved Above Average (RSAA). This gave me 411 pitchers to work with, plus a few active guys who went over 1500 IP this year. I then looked each of them up in the Neyer/James Guide to Pitchers and narrowed the list to those who had a sinker listed as either their first or second pitch. We're now left with 52 sinkerballers, spanning the ages from Grover Cleveland (Pete) Alexander to Derek Lowe.

Okay, so what do these 52 pitchers tell us? Well...actually, let's just look at the numbers first. Here is the average career line for our sinker group:

W	L	ERA	IP	H	R	ER	HR	BB	K	HBP	CG	SHO	DH	DR
156	131	3.53	2515	2502	1125	987	187	742	1361	54	97	21	+5	-20

Pretty good, obviously. We are looking at only pitchers with long, above average careers after all.

The last two columns require explanation: Delta Hits (DH) is the number of hits a pitcher gave up versus what you'd expect from the defense behind him. Delta Runs (DR) is the number of runs a pitcher gave up versus what you'd expect from his peripheral stats. For both, negative is good and positive is bad.

That our sinkerballers gave up more hits than expected makes intuitive sense. Groundballs become hits at a higher rate than flyballs. That's the tradeoff: a lot more singles, but fewer extra-base hits.

Which brings us to the -20 in the DR column. While I don't know exactly how Clay Davenport calculates these figures, I'm going to guess exact figures for doubles and triples aren't included. Since groundball pitchers will generally give up less of these, it basically takes three singles to score one run. In addition, they should also induce more double plays. There's also the theory that the sinker sinks more from the stretch, even though that blew up in my face in Part One. I'm still not giving up on it. At any rate, there are any number of reasons why sinkerballers might give up fewer runs than the rest of their stat line indicates.

It's kinda tough to read anything into these numbers without context though. With that in mind, I created a control group. Returning to the aforementioned list of 400+ pitchers, I selected 52 of them who roughly matched up to our sinkerballers in ERA, IP & RSAA. I also checked them in the Neyer/James Guide to make sure none threw a sinker, even as a third or fourth option. Back to the chart:

Group	W	L	ERA	IP	H	R	ER	HR	BB	K	HBP	CG	SHO	DH	DR
Sink	156	131	3.53	2515	2502	1125	987	187	742	1361	54	97	21	+5	-20
No Sink	159	125	3.58	2509	2410	1135	999	204	830	1479	52	111	23	-36	-16

About what we'd expect. The sinkerballers pitched to contact more, with fewer homers, fewer strikeouts, fewer walks and more hits. The difference in the DH column underscores nicely the added singles from the sinker. Unfortunately, my theories on why sinkerballers were better than average in the DR column are shot down by the similar number from our no sink group.

Another chart:

Group	ERA	UERA	H/9	HR/9	UIBB/9	K/9	BABIP	TTO%
Sink	3.53	0.49	8.95	0.67	2.47	4.87	.281	21.2%
No Sink	3.58	0.49	8.64	0.73	2.87	5.30	.276	23.6%

It's interesting that the unearned runs are the same for both groups. Since most errors come on grounders, one would expect sinkerballers to have a higher UERA. Maybe the higher error rate is offset by the lower walk and home run rates, so opponents don't take advantage of errors as much.

UIBB is unintentional walks per 9 innings. (The sink group averaged 51 IBB v. 30 for the no sink group.) This is the biggest difference between the two, as sinkerballers walk 14% fewer batters than their counterparts.

BABIP, for those unfamiliar, is Batting Average on Balls In Play. This is yet another way of hammering home the point that sinkerballers allow more hits.

And the last column, TTO%, stands for the Three True Outcomes: strikeouts, walks and homers. These are all the times that the pitcher and batter solve things by themselves, without bringing all those pesky fielders into the matter. Sinkerballers have a TTO 10% less often, which equals a lot more balls in play.

I was initially surprised to see sinkerballers allow just 9% fewer home runs than their counterparts. But really, that's a nifty little difference. Remember, we're only looking at successful pitchers here, and it's awfully difficult to have a long, productive career if you're handing out gopherballs left and right. That sinker pitchers keep the ball in the park above and beyond already quite-good pitchers is rather impressive. If we could compare them to league average, the numbers would change quite a bit.

Well, we might as well do that. For each season since 1900, I got the league average for ERA, walks, strikeouts and home runs per inning off Baseball-Reference (note: I'm calculating AL and NL separately). Then for each player season, I compared his actual ERA, walks, home runs and strikeouts to the expected totals.

Quick example: in 2002, there were 0.122 home runs hit per inning. Derek Lowe pitched 219.7 innings, so we would expect him to give up 26.8 home runs. He actually surrendered 12, so he was 14.8 HR better than average.

Graph time!

(Note: these numbers are NOT park-adjusted. )

Group	ER diff	HR diff	BB diff	K diff	ERA+	HR+	BB+	K+
Sink	-128	-31	-161	-71	113	114	119	95
No Sink	-129	-2	-96	+144	113	101	111	111

(The first four columns are again averages of all 52 players per group, for consistency with previous graphs. The last four columns are rate stats, with 100 as average. So 119 means 19% better than average and 95 means 5% worse than average.)

That's more telling. As you can see from the last four columns, the two groups were exactly as effective, but in very different ways. Our No Sink group is, in fact, pretty much average at preventing the longball. They get their fair share of strikeouts and do a good job of avoiding walks.

The sinkerballers, on the other hand, succeed in spite of not striking batters out. Their walk and home run rates are fantastic, much better than their counterparts'.

So what does this all mean? Well, it means that strikeout rate isn't all that important for sinkerballers. That's not to say they can get away with no K's, but a below average rate isn't all that disconcerting. I mean, these are essentially the 52 best sinkerballers ever. If they had this level of success--right in line with non-sinkerballers--with a below average K-rate, then maybe we shouldn't stress out so much over Wang's inability to miss bats.

Yes, Wang needs to up the strikeouts somewhat. But he doesn't need to get to Kevin Brown or Brandon Webb levels. He doesn't even need to get to average. He can still be a great pitcher without the K's.