Brief Programming Note

I’d like to say thanks to everyone who has been reading and commenting to this point; the support has been tremendous.  In my initial post, I said I was going to try to post every day.  I’ve taken some flak over the last few days for not posting every day. Please allow me to clarify – every day means when I want every weekday.  Expect 5 posts a week, one Monday through Thursday, with one sometime over the weekend.  Hopefully there will be enough Knuckleballs to go around.  Thanks again for reading, and if you’re curious, posts over the next three days will be regarding Cole Hamels, predicting Cy Young award winners, and (per commenter Mac’s question about how much to spend on a closer/bullpen, which I poorly answered) bullpen chaining.

Matt Cain

When thinking about my next post, I wondered about Matt Cain’s 2009.  Consider the following: a pitcher strikes out fewer hitters, gives up more home runs, and sees his ERA drop by a run.  My first guess was to look at his walks allowed and BABIP, and immediately I saw what I thought was the answer.  Cain’s BABIP fell 35 points from the season before, but in three of his four seasons as a full-time starter, he’s beaten the league average.  He also walked almost a batter fewer in 2009 than in 2008.  Here’s a look at his career numbers:

Season	ERA	GS	IP	H	R	ER	HR	BB	SO	K/9	BB/9
2005	2.33	7	46.1	24	12	12	4	19	30	5.83	3.69
2006	4.15	31	190.2	157	93	88	18	87	179	8.45	4.11
2007	3.65	32	200	173	84	81	14	79	163	7.34	3.56
2008	3.76	34	217.2	206	95	91	19	91	186	7.69	3.76
2009	2.89	33	217.2	184	73	70	22	73	171	7.07	3.02

and a quick look at his advanced statistics:

Season	K/BB	HR/9	BABIP	LOB%	FIP	GB/FB	LD%	GB%	FB%	HR/FB
2005	1.58	0.78	0.166	82.9%	4.08	0.56	17.8%	29.5%	52.7%	5.9%
2006	2.06	0.85	0.279	69.8%	3.96	0.75	16.7%	35.6%	47.7%	7.1%
2007	2.06	0.63	0.284	72.9%	3.78	0.89	16.1%	39.4%	44.5%	5.5%
2008	2.04	0.79	0.304	75.3%	3.91	0.76	22.8%	33.2%	44.0%	6.8%
2009	2.34	0.91	0.268	81.6%	3.89	0.92	18.7%	38.9%	42.4%	8.4%

As seen in the second table in the LOB% column, Cain stranded almost 82 percent of the runners who reached base against him last year; the major league average was 71 percent.  This is a simple yet difficult-to-maintain way to not allow runs; when guys are already on base, don’t give up any more hits.  The average BABIP in the National League  has been 0.303 through the duration of his career, which Cain has routinely beaten.   He was able to sustain a drop in his strikeout rate by simultaneously and proportionally dropping his walk rate.  All of this suggests luck, but he’s done it several times; 2009 just accentuated the point.  He allowed about the same number of fly balls, and a higher percentage of those went for home runs.   The only number that changed that really stands out (and mattered for the Giants in 2009) is his earned runs allowed, which dropped by over 20 percent.

The fifth column in the advanced stats table is Fielding-Independent Pitching, and it can be a difficult statistic to wrap your head around.  The onus for run prevention shows up next to the pitcher’s name in conventional box scores, but a lot more goes into it than that.  Earned runs allowed depends greatly on the defense behind the pitcher, and it very context specific.  In an attempt to remove some of this subjectivity, this metric includes only home runs allowed, walks allowed, and strikeouts – the things pitchers have shown an ability to control.  It strips out the defensive ability behind the pitcher and the context-sensitivity of ERA.  Some would argue that the context makes all the difference in the world, and I would agree, for individual games.  To eliminate subjectivity, FIP is a useful tool.

FIP is not an attempt to replace ERA; it’s an attempt to enumerate the pitcher’s abilities in a context-neutral environment.  As shown above in the second table, Matt Cain’s FIPs have barely moved, and his ERAs have generally agreed.  In 2007, he gave up fewer home runs and had a low BABIP, beating his expected ERA slightly.  In 2008, he stranded a few more runners, and in 2009, he stranded A LOT more runners while giving up fewer hits on balls in play.  While his BABIPs might be sustainable (many good pitchers are able to keep their BABIPs slightly lower than league average over a long period of time), his strand rate is the driving factor for his excellent 2009.  Given the contextual nature of this statistic, either Matt Cain was really “clutch,” or really lucky with runners on base in 2009.  Sure enough, his BABIP fell to 0.218 with men on base, and 0.190 with runners in scoring position.  This leads to the high strand rate and is the real reason for the fewer runs allowed.  With no one on base, Cain’s BABIP was 0.299, almost exactly league average.

Lastly, I looked at Cain’s previous seasons to see if this was a pattern or a fluke.  For his career, Cain’s BABIP is 0.274.  That number has increased to 0.284 with runners on base and 0.285 with runners in scoring position.  BABIP for most pitchers will decrease slightly with runners on base and with men in scoring position, but no more than 10 points at the extreme (in 2009, the league average was 9 points).  This phenomenon is usually contributed to the idea of  “quality outs,” ground balls which advance runners already on base.  Unfortunately for Matt Cain, it appears he was really lucky in 2009.  There’s nothing to suggest that this will occur again in 2010; in fact, factors such as the decreasing strikeout rate seem to indicate a possible turn for the worse.  Unless Cain knows that he’s doing something different with runners on base, I think this is another result we can chalk up to luck.  Nevertheless, the uncharmed version of Cain is still valuable and if he keeps his ratio of strikeouts and walks the same (either the 2006/2008 version or the 2007/2009 version), he should expect continued success.

Season	ERA	GS	IP	H	R	ER	HR	BB	SO
2005	2.33	7	46.1	24	12	12	4	19	30
2006	4.15	31	190.2	157	93	88	18	87	179
2007	3.65	32	200	173	84	81	14	79	163
2008	3.76	34	217.2	206	95	91	19	91	186
2009	2.89	33	217.2	184	73	70	22	73	171