Comments on: The Anatomy of Every Start – 2010 This and that about baseball stats. Tue, 16 Jul 2013 17:01:55 +0000 hourly 1 By: John Autin Tue, 23 Nov 2010 05:09:17 +0000 @17-18 --
I am getting different numbers from the Play Index than you gents reported:

-- For IP>=6, ER=3, I'm getting a record of 264-303 (.466).
-- For IP=6, ER=3, I'm getting 98-122 (.445).
-- For IP>=6.1, ER=3, I'm getting 166-181 (.478).

Considering these discrepancies and my previous posts, I'm wondering if (hoping that?) the Play Index is malfunctioning.

By: John Autin Tue, 23 Nov 2010 04:50:29 +0000 Following up on #1 and 27 -- The mystery of why teams win substantially more in a start of 6.2 IP and 3 ER than a start of 7 IP and 3 ER is going to drive me nuts until I can understand it. Can anyone explain it? It holds true during the entire steroids era, even when the leagues are measured separately.

AL teams, 1993-2010:
-- 6.2 IP & 3 ER, .562 W% (241-188).
-- 7 IP & 3 ER, .510 W% (659-634).

NL teams, 1993-2010:
-- 6.2 IP & 3 ER, .556 W% (195-156).
-- 7 IP & 3 ER, .496 W% (754-766).

But if I look back to the 20 years from 1973-92 (post-DH, pre-steroids), the AL numbers look more rational:
-- 6.2 IP & 3 ER, .474 W% (200-222).
-- 7 IP & 3 ER, .539 W% (479-410).

So what is it about the current era that makes 6.2 IP & 3 ER look better than 7 IP & 3 ER?

One last table, including 6.1 IP & 3 ER:
1993-2010, both leagues:
-- 6.1 IP & 3 ER, .510 W% (397-382).
-- 6.2 IP & 3 ER, .559 W% (436-344).
-- 7 IP & 3 ER, .502 W% (1413-1400).

Can someone make sense of this? Please?

By: John Autin Tue, 23 Nov 2010 04:11:43 +0000 @25 --
[in my best Emily Littella voice]:
"Oh! ... Never mind!"

I am guilty of starting my tirade before reading to the end of your post. My apologies.

Moving on....

Your data table is interesting, but I don't think you said what year(s) it covers. If it's 2010 only, my check of the Play Index gives different results than yours. For example, in 2010, the team W% in a start of exactly 7 IP and 3 ER was 40% (not 44%), based on a record of 70-105. The team W% in a start of 6.2 IP and 3 ER was 57%, not 60%, based on a record of 24-18.

More to the point, though, the confounding downturn in W% from 6.2 to 7 IP should tell us that something is fundamentally wrong with the team-W% method of analyzing pitcher starts. There's no logical reason to say that a pitcher who allowed 3 ER in 7 IP did a worse job than one who allowed 3 ER in 6.2 IP -- but that's exactly what the team-W% method would say. Therefore, the method is suspect at best.

By: Albanate Tue, 23 Nov 2010 02:52:39 +0000 #11:

Cliff Lee gave up 6 runs in 9 innings on 7/10:

Edwin Jackson had the -5 GS:

Scott Kazmir had the -8 GS:

By: Zack Mon, 22 Nov 2010 21:40:18 +0000 @20 - But that was my point - perhaps I wasn't specific enough. I am attacking the minimums and not merely the concept of a "quality start" statistic. I think the stat is interesting, partly for the reason you cited (independence of team). It's the 3ER threshold that is my concern.

@21 - Yes - Quality starts earned a 1687-768 record.

By: Tmckelv Mon, 22 Nov 2010 21:15:07 +0000 JT @21,

A "Quality Start" is not (should not be) "official" until the starter's record is complete for any given game. So a pitcher could be "in line" for a QS and not get it, but he cannot have a QS and lose it due to the reliever's performance.

Sorry about all of the quotation marks, but since it is an unofficial stat, nothing is concrete (I guess that would apply to your comment as well, so my point is pretty moot).

By: John Autin Mon, 22 Nov 2010 21:05:47 +0000 @21-22, JT --

Re: relievers, I was referring to relievers "blowing the win"; for that purpose, I don't care if it's inherited runners or the reliever's runners who score.

And sure, your 2nd point is true, but that's a much smaller set than the "wins" and "losses" that accrue to SPs primarily because of the team's batting performance.

Bottom line: Quality Start is not 100% independent of the team's supporting performance, but it's FAR more independent than SP W's and L's.

By: Johnny Twisto Mon, 22 Nov 2010 20:45:41 +0000 Also, if the pitcher's team isn't hitting, perhaps he is more likely to be pinch hit for in the top of the 6th inning, before he can reach the QS threshold. So it's not completely independent.

By: Johnny Twisto Mon, 22 Nov 2010 20:44:11 +0000 I checked a few days ago, I think teams won two-thirds of the time they got a quality start.

But John, if a reliever allows inherited runs, those would be charged to the SP, and could change a QS into a non-QS.

By: John Autin Mon, 22 Nov 2010 20:24:08 +0000 @1 -- "Do[] many of us consider 3ER in 6IP a 'quality start'? I do not."

Well hey, it's the Straw-Man "Critique" of the Quality Start! Haven't heard that one in a few days. Welcome back!

It's "attacking a straw man" because you've set up the minimum requirements of the Quality Start as representing the entire category. You then condemn the category because those minimums aren't impressive.

But at least we can trust you to be consistent, right? So, you must not put any stock in starting pitchers' wins, since those can be "earned" with just 5 IP and any number of runs allowed.

For anyone who still doubts the merits of the Quality Start but retains an open mind, consider this:

In 2010 (as usual), the average Quality Start featured more innings and a better ERA than the average SP Win.
-- Quality Start average: 6.89 IP, 1.98 ERA.
-- SP Win average: 6.78 IP, 2.00 ERA.

Quality Start has the further advantage of being independent of what the pitcher's teammates do at the plate and what his relievers do after he leaves.