Sunday, June 1, 2014

The Astros in May 2014: Starting Pitching Edition.

Last month, we looked at the Astros starting rotation in comparison to the surprising Royals and the current Champion Red Sox.  We found that the Astros had a surprisingly good starting rotation, pretty much equal to what the Red Sox were putting out there on a team-wide basis, but that the Royals had been nothing short of excellent at preventing runs from their starters.

This month, I have changed teams around.  We will revisit the Royal and Red Sox later in the year, but I decided to look at two other teams.  Both of the new teams had the narratives attached that they had the basis of a good starting rotation, but they needed to be active in accumulating SP talent from around the the Majors during the offseason (mostly due to dry farm systems).  And they did.  Meet the Angels and the White Sox, to be compared to the Astros:

Houston Astros (average age 26.3, average salary approx. $2.1 million)
29 starts
IP: 182, H: 176, R: 80, ER: 70, BB: 52, K: 141, HR: 14, ERA: 3.45
K:BB Ratio: 2.29,  WHIP: 1.39
Groundballs: 250, Flyballs: 149, Groundball: Flyball Ratio: 1.68
Wins: 12, Losses: 13, Quality Starts: 18, Disaster Starts: 2
Record for Quality Starts: 10-6, Record for Disaster Starts: 0-2.

Los Angeles Angels of Anaheim (average age 27.6, average salary approx. $5.7 million)
28 starts
IP: 175, H: 155, R: 83, ER: 75, BB: 58, K: 146, HR: 16, ERA: 3.86
K:BB Ratio: 2.51,  WHIP: 1.22
Groundballs: 197, Flyballs: 177, Groundball: Flyball Ratio: 1.11
Wins: 13, Losses: 10, Quality Starts: 16, Disaster Starts: 3,
Record for Quality Starts: 11-2, Record for Disaster Starts: 0-3.

Chicago White Sox (average age 26.8, average salary approx. $3.35 million)
27 starts
IP: 151, H: 165, R: 87, ER: 80, BB: 53, K: 115, HR: 21, ERA: 4.76
K:BB Ratio: 2.17;  WHIP: 1.44
Groundballs: 174, Flyballs: 152, Groundball: Flyball Ratio: 1.14
Wins: 6, Losses: 12, Quality Starts: 14, Disaster Starts: 3,
Record for Quality Starts: 4-5, Record for Disaster Starts: 0-2.

* - Disaster start defined as <5IP, 4 or more runs or earned runs given up.
** - Each of the three teams used six starters in the month of May, so the average age and salaries of the rotation are calculated using these six players.

The White Sox are clearly the most inferior of the three rotations on raw numbers.  Some of this is likely due to Chris Sale's absence (he pitched only three starting innings in May due to injury), but the lack of a true ace pitcher - probably the best pitcher on all three teams - is unlikely to account for the extent in the skew in numbers.

Of course, these numbers are not park-adjusted.  The Cell is rated the 21st overall park for suppressing run scoring (Minute Maid is fifth, and Angel Stadium is sixth), so the park-factor, if anything, protects the White Sox rotation in comparison to the other two.  The Astros and Angels pitch in one of the hardest parks to suppress runs in the majors, making their raw numbers look more impressive.

(These raw statistics do not account for strength of schedule, either.)

Further comparison reveals that the White Sox have an ERA that is around 136% of that of the Astros, and a WHIP of around 115% that of the Astros.  They have tossed the fewest innings of the three clubs looked at, given up the most runs, the most walks, the most home-runs, and stuck out the fewest batters.  As May has progressed, they have continued to hover around .500, and their starting pitching has clearly nor assisted much.  Still, most Astros fans would kill for "hovering around .500" - in a franchise with recent history like Houston has, that would be cause for a party this late in the season.

The two teams lending themselves to the most direct of comparisons are the Angels and the Astros.  The raw numbers are very similar for starting pitching, and we will look at the raw batting and base running statistics tomorrow.

The Angels entered May 14-13, having won three games in a row, and exit May at 30-25, comfortably in second place in the AL West, and now 3.5 games behind the juggernaut A's.  Overall, their starting numbers are very decent, and this has been helped by excellent May results for Weaver (8IP/1ER, 6.3IP/1ER, 7IP/2ER, 9IP/1ER and 6IP/3ER) and Wilson (8IP/3ER, 6IP/5ER, 9IP, 0ER, 6.1IP/1ER, 7.2IP/3ER) and solid May for Garrett Richards (6IP/3ER, 7IP/2ER, 7IP/0ER, 7IP/5ER, 7IP/3ER... but 0.2IP/5ER in his last start of the month).  Thirteen out of sixteen quality starts for that trio, and one start lasting less than six innings-pitched.  Not shabby.

The other metrics like the Angels as a staff, too.  Their ERA sits at a solid, non-park-adjusted 3.78, and the WHIP sits at a very solid 1.20.  They strike out over two-and-one-half more batters than they walk.

That said, the Astros staff has pretty much their absolute equal, if not slightly better.  While the Astros' WHIP is a little higher (1.25), their ERA is a quarter-run lower, possibly by virtue of their defensive shifting and their ground-ball centric staff (1.66 GB:FB ratio, compared to the Angels' 1.12).  Both staffs have very similar numbers of innings pitched, quality starts and disaster starts.

The Astros threw 8 more innings, gave up a 21 more hits, walked six fewer batters, and struck out five fewer hitters.  The Astros gave up two fewer home-runs.  They started one extra time, so that could account for the innings difference, but not the other differences.  So, perhaps the Astros are just a little more hit-prone at this stage, but again this is likely a reflection of their policy of pitching to contact, and making the contact of the ground-ball variety.

The Astros and Angels starting staff were credited with roughly the same number of wins compared to losses (12-13 compared to 13-10).  Clearly, the win and loss stats are influenced by both pitching and hitting, but all things being equal, the Astros have the wins as losses tilted against them in comparison to the Angels.

The real advantage to the Astros comes when the average age and average salary is weighed against the the overall production.  The Astros are the cheapest and youngest of the rotations by a considerable margin, and are arguably the most effective of these three rotations.

In last months' article, we removed the statistical outliers from the various rotations, then re-analysed the rotations to try and get a better feel for how much of an effect the outlier had on the numbers of the overall staff.  Last month, Lucas Harrell, who made four starts, three awful, for the Astros was removed as a "bad" outlier, along with the Red Sox and Royals' worst starters.  Predictably, we found the Lucas had dragged the Astros down somewhat in terms of their overall starters' stats, and once his numbers were excluded, the Astros were very similar to the Red Sox.

I want to continue to remove the leading statistical outlier on the Astros staff, and reanalyse.  However, this month the outlier on the Astros is not the worst pitcher, but most likely the best.  Dallas Keuchel has been the focus of a few articles expressing wonderment recently, because of his excellent month.  (Imagine what the month could have been like without seventh inning sudden loss-of-control versus the Tigers (3 May), a two-out, 2 run triple (May 19) against the Angels, and a fifth-inning loss of control versus the Orioles (31 May).  Yikes!).

But, until Keuchel can prove that he dominate for half a season or more, his current performance remains somewhat outlier-ish, so lets remove him, and the other two best starters from each team (Jose Quintana and Jered Weaver) and re-run the digits:

Houston Astros (minus Keuchel)
IP: 136, H: 144, R: 68, ER: 59, BB: 45, K: 107, HR: 13, ERA: 3.90
K:BB Ratio: 2.38; WHIP: 1.39
Groundballs: 169, Flyballs: 127, Groundball: Flyball Ratio: 1.33
Wins: 8, Losses: 11, Quality Starts: 13, Disaster Starts: 2,
Record for Quality Starts: 6-5, Record for Disaster Starts: 0-2.

Los Angeles Angels of Anaheim (minus Weaver)
IP: 139, H: 133, R: 75, ER: 67, BB: 48,  K: 119, HR: 13, ERA: 4.35
K:BB Ratio: 2.48; WHIP: 1.31
Groundballs: 164, Flyballs: 129, Groundball: Flyball Ratio: 1.27
Wins: 9, Losses: 9, Quality Starts: 12, Disaster Starts: 3,
Record for Quality Starts: 7-2, Record for Disaster Starts: 0-3.

Chicago White Sox (minus Quintana)
IP: 117, H: 130, R: 73, ER: 66, BB: 42, K: 89, HR: 21, ERA: 5.09
K:BB Ratio: 2.12; WHIP: 1.47
Groundballs: 127, Flyballs: 123, Groundball: Flyball Ratio: 1.03
Wins: 4, Losses: 9, Quality Starts: 10, Disaster Starts: 3,
Record for Quality Starts: 3-3, Record for Disaster Starts: 0-2.

Obviously, with the best starter removed, all the teams' major statistical measures take a bit of a tumble.  The Astros and While Sox increase their ERA's approximately 0.5 of a run per nine, whereas the Angels hold their increase to just under 0.4 runs per nine.  The Astros experience the largest increase in WHIP out of the three teams, proving just how efficient Keuchel has been in keeping batters off bases (7 walks in May, one in the four starts between his May 3 start and his May 31 start).  The Astros also experience the largest tumble in GB/FB ratio (0.33, compared to the White Sox 0.08 and the Angels' improvement in GB:FB ratio of -0.19), showing (i) that Keuchel is the most effective on the Astros staff in keeping the ball on the ground, and (ii) Weaver is an extreme fly-ball pitcher who relies on weak contact.  The Astros experience the largest reduction in strikeout to walk ratio with Keuchel removed (0.35, compared to the White Sox' 0.05 and the Angels very slight improvement: -0.01 (again, proving how much of a statistical outlier Jered Weaver is).

Regardless, the characteristics of the starting staffs described above remains broadly intact.  The Astros outperform both other teams in ERA, whilst remaining somewhat of a ground-ball staff which gives up more in the way of hits to the Angels.  Both the Angels and the Astros outperform the White Sox.  All three starters have done a great job of keeping themselves in the contest, and doing a good job of avoiding disasters.

On the surface, the Astros' pitching staff has been very solid.  But while most of the brilliance has been provided by Dallas Keuchel, the rest have been plenty solid with the occasional dominant performance thrown in.  All Astros starters have pretty much kept the team the game, which is something, especially over a long season, and especially given the relative youth of the group.

Finally, lets compare April 2014 Astros with May 2014 Astros:

April Houston Astros
IP: 157, H: 146, R: 79, ER: 75, BB: 66,  K: 126, ERA: 4.28
K:BB Ratio: 1.90;  WHIP:  1.34
Groundballs: 235, Flyballs: 138, Groundball: Flyball Ratio: 1.70
Wins: 7, Losses: 13, Quality Starts: 13, Disaster Starts: 6,
Record per Quality Start: 6-1, Record per Disaster Start: 0-6.

May Houston Astros (rehashed from above)
IP: 182, H: 176, R: 80, ER: 70, BB: 52, K: 141, HR: 14, ERA: 3.45
K:BB Ratio: 2.29,  WHIP: 1.39
Groundballs: 250, Flyballs: 149, Groundball: Flyball Ratio: 1.68
Wins: 12, Losses: 13, Quality Starts: 18, Disaster Starts: 2
Record for Quality Starts: 10-6, Record for Disaster Starts: 0-2.

An across-the-board improvement for the Astros between April and May, with perhaps the only peripheral that is not in their favour being the slightly increase number of hits (especially when compared to innings pitched) in May.  That said, there will be some random variability there, and it is most likely a combination of noise, and Feldman regression (who was filthy for most of April, and has struggled a bit in May).  Arguably, the May Astros are a tiny bit less ground-ball-y as well, but again, most likely random variation plus subtraction of Lucas Harrell.  Singificant improvements in disaster starts (but remember three April disasters were attributed to Lucas Harrell), and quality starts are noted as the season has progressed, and are certainly welcomed.

The biggest improvement is the significant ERA drop between April and May.  Also, remember that the minus-Harrell ERA was 3.84 in April, so this continues to represent a significant improvement, even with Harrell excluded.  This improvement demands ongoing monitoring.

So, in summary, the solid April that the Astros starting pitchers put together have continued into a solid May.  The Astros starters also compare vary favourably with a rotation with two expensive free-agents on a club that (as a team) has had a great month, and finds itself contending through the first third of the season (the Angels).  The Astros continue to minimise base runners and home runs, and keep the ball on the ground at a far better rate than any of the teams that we have looked at this year.  The longer this starting pitching continues like this, the more likely that it will continue, as the statistical noise gets eliminated, and regression becomes less of a concern.  Some of the starters still have room to improve, given their ages, and improvements in their peripheral stats, but also when the individual box scores are looked at.  Some of the starters may also turn into pumpkins - and if that happens, the organisation still has some arms that they will want to see having an opportunity in a ML rotation.

Tomorrow, I will put up a much, much briefer post, using the advanced statistics to compare the rotations, once Fangraphs bangs them up.