Rick Porcello and facing the same team after a bad outing

Rick Porcello takes the mound tonight versus the Braves, squaring off against the same team that took him behind the woodshed in his first start with the Mets. This gives us a chance to test out a theory of mine. My theory is that whenever a player faces a team that he just pitched against recently, that if he had a particularly good or bad outing against them the first time, he’ll do the opposite in the second go round.

It’s a theory because I’ve never actually looked at the results. So, let’s do that now. First, we’ll have to define some terms. What’s a particularly good or bad start? Let’s use the Bill James Game Score number to determine this, as found on Baseball-Reference. James considers several factors and throws them into equation where an average Game Score is 50. So, let’s consider a particularly good start as one with a Game Score of 60 or higher and a particularly bad Game Score as one with a 40 or worse.

Porcello had a Game Score of 14 for his outing against Atlanta on July 26.

What do we consider recently? For our purposes today, let’s use a definition of two starts in three games within a period of two weeks. So, if a guy faced a club and then went on the injured list for several months and then faced the same club again when he was activated, that won’t count. This actually happened with Noah Syndergaard back in 2017.

Before looking at the data, let me preface this by saying that there weren’t as many of these as expected. Partly this is because you typically only have the opportunities to do this in division games. It’s not like a Mets pitcher frequently faces the Dodgers twice in three starts. And the other is that a lot of starts end up with a Game Score in the 40s or 50s, eliminating them from our consideration. Steven Matz faced the Phillies last year in back-to-back starts late in the season but his first outing against them had a Game Score of 48, so it didn’t make our cut. Here’s our lists, beginning with the particularly good starts:

Date	Pitcher	Opponent	IP	ER	BB	K	GM Score		Date	IP	ER	BB	K	GM Score
81419	Matz	ATL	6	1	1	5	68		82519	6	1	1	6	69
72719	Matz	PIT	9	0	0	7	84		80219	3.2	5	1	4	32
62719	Wheeler	PHI	6	1	2	7	69		70719	5	6	2	7	32
61819	deGrom	ATL	8.1	2	0	10	75		62819	6	3	2	7	53
51119	deGrom	MIA	7	1	1	8	70		51719	5	6	0	3	26
42819	Matz	MIL	7	2	0	4	63		50319	5.2	3	1	3	41
82018	Wheeler	SFG	7	1	1	10	72		83118	7	1	0	9	74
42118	deGrom	ATL	7	0	2	10	77		50218	4	0	0	6	64
40718	Matz	WSN	5	0	2	8	65		41818	4	3	1	6	49
40518	deGrom	WSN	6	1	3	5	60		41618	7.1	3	1	12	65
91717	Gsellman	ATL	7	0	0	3	72		92717	6	1	1	4	59
83017	Montero	CIN	8.1	0	4	8	81		90917	5	1	5	5	55
40917	Syndergaard	MIA	7	1	0	9	70		41417	6	1	0	4	58
Sum			90.2	10	16	94	926			70.2	34	15	76	677
Per Game			7	1	1	7	71			5.1	3	1	6	52

Looking at the 2017-2019 period, there were 13 times that a Mets pitcher had a particularly good start and then faced that team again within his next two starts. In our initial start, the average Game Score was 71 and the follow-up start it was 52. That seems to work in favor of the theory. But let’s dig a little deeper.

There are two factors to consider here that work against our theory. First, you have the possibility of a good pitcher dominating a bad team. The 2018 Giants that Zack Wheeler put up consecutive strong starts against in 2018 went 73-89. This happened in the second half of the year, when Wheeler was lights out and the Giants went 23-41. And, there’s also the fact that pitchers tend to bunch good or bad performances together. You’ll frequently hear that Pitcher X is having a great month.

On the flip side, we have the overall tendency of regression. If a true talent 4.21 ERA guy throws a shutout, like Steven Matz did against the Pirates last year, chances are he’s not going to duplicate that performance the next time out. Essentially, the theory is saying that regression is a stronger factor than either the quality of the team or the streakiness of the pitcher.

Now let’s look at the particularly bad starts:

Date	Pitcher	Opponent	IP	ER	BB	K	GM Score		Date	IP	ER	BB	K	GM Score
81319	Wheeler	ATL	5	5	2	2	23		82419	6	5	4	3	39
61919	Matz	ATL	5	5	4	2	33		62919	2	2	0	3	45
51619	Wheeler	WSN	6	6	2	6	30		52119	7	3	2	6	61
51519	Font	WSN	2.1	5	2	2	25		52019	4	2	4	3	47
41619	Matz	PHI	0	6	1	0	13		42219	6	1	2	6	66
90818	Syndergaard	PHI	6.2	4	5	4	33		91918	4	3	3	6	45
42717	Harvey	ATL	4.1	6	5	1	25		50217	5.1	6	3	2	27
42617	Gsellman	ATL	4	5	3	2	19		50117	5	5	1	0	34
40817	Milone	CHC	4	4	1	2	29		41917	5	1	0	5	62
Sum			37.1	46	25	21	230			44.1	28	19	34	426
Per Game			4.1	5	3	2	26			5	3	2	4	47

There were only nine instances for the Mets in the last three years of what Porcello will be doing tonight. This really surprised me. And there was a tiny bit of unintentional cheating going on to get to nine. The information for Tommy Milone in 2017 was already entered before the realization hit that this came before he joined the Mets. Those two starts both happened when Milone was on the Brewers. Since it was already entered – and we needed the game for a better sample size – it was kept in.

The Game Scores for the particularly bad starts averaged out to 26, which is quite awful. The follow-up start produced a 47 score, a nice increase but not as strong as my expectation going in. Only one-third of our follow-up starts would qualify for our particularly good start list and one of those three was the cheat start by Milone. Maybe a larger sample would have produced different results.

Checking Porcello, he did not have a start the past three years which qualified for our particularly bad grouping and a recent follow-up against the same opponent. You have to go back to 2015 to find a case where it happened. On April 19, he had a lousy game against the Orioles, one that resulted in a Game Score of 14. He faced them the next time out on April 24 and put up a Game Score of 49. In that game, Porcello gave up 4 ER in 6 IP. After he went just two innings in his first start this year, the Mets would be happy with a 6 IP outing tonight.