Two of the more interesting stories for our beleaguered team this year are closer Huston Street and setup man and AN author Brad Ziegler. Having read the title (and not being an Angels fan), you have probably figured out which player will be the focus of this article. AFTER you finish reading this article and commenting on it thoroughly, head over to Hardball Times, where our own salb918 has written a fascinating article on Brad Ziegler and why his slow pitch softball delivery is so effective. He has graphs … I like graphs … I have graphs too, so stay and enjoy my graphs … then go check out his article.
So Huston Street has been a pretty frustrating player for a lot of us. I had always seen him as the heir apparent to Dennis Eckersley. In celebrating that, I hung my signed, framed SI with Eck on its cover next to my signed, framed Beckett with Huston on the cover.
But, see, there is a difference between Huston Street and most closers. Health issues this year, aside, Huston Street is a really, really good pitcher. Over his career, he has done a great job of not allowing the opposing team to score very many runs. The problem, though, is that he has just been too darn consistent. Whereas his overall numbers have generally been in line with the bottom of the top tier or the top of the second tier (depending on how you define them) of closers, he has blown quite a few more saves than you would expect. Why? Because, in accumulating saves and avoiding blown saves, consistency is a bad thing. The more evenly his runs allowed are distributed, the more blown saves he will have. Because of that, though, the A’s win a lot of games that Huston blew the save in. They are 2-3 this year, for example. Because of that, saves and blown saves underrate Huston’s ability as a closer. How much, exactly, is a story for another column.
This year, though, it has been different. Huston has given up more runs, hits, homeruns and walks than he has in the past. Why? Well many suggest that it is due to a declining fastball, likely due to health problems. But is that true?
Well, at first glance, his fastball has definitely slowed down a bit over the years.
2005: 91.2 mph
2006: 91.7 mph
2007: 90.4 mph
2008: 89.8 mph
source: fangraphs.com
Yep, that is slower – but I am still not convinced. This is Staturday and we demand more rigorous analysis than that. Thanks to 74mk, that analysis is now possible. Note, this data does not include games in Japan, as those did not have reliable pitch data and runs through July 27.
As you can see, his fastball velocity has trended down over the course of the season but it is pretty all over the place and it is hard to make much out of it.
All of a sudden -- when we use a rolling average, instead of single games – the data becomes a lot more clear. Huston’s velocity plummeted in late May. In April and most of May, he was typically working at nearly ninety miles per hour. Since June, though, his velocity has typically been around eighty-eight. So, yes, I think it is pretty clear that Huston as lost velocity – but that leaves the question, is that loss of velocity the reason for his sub-par play?
To determine this, I separated games by the speed of his fastball. Based both on simplicity and comparable sample size, I broke the games into three categories – those in which his average fastball was 88.49 mph or less, those where his average fastball was 89.5 mph or more and those in between. The difference in results was stark:
When Huston was working in the upper range, he was flat out dominant – striking out more than a batter per inning and only allowing a hit every third inning, a batting average of roughly .120.
The middle group of games sees a small drop in strikeouts but a huge increase in hits, homeruns and, not surprisingly, runs.
The slowest group is event worse. He strikes out very few, gives up more than a hit an innings and allows a ton of runs. The only redeeming quality of this set is that he managed to keep the ball in the park.
So Huston’s fastball becomes very hittable when it drops below ninety? Not so fast.
Ninety-one of the batters Huston faced saw their plate appearance end with a fastball. In those at bats, batters put up a .200/.253/.341 line. Clearly, hitters have not done much damage to Huston’s fastball. What is more, it does not really matter how fast the fastball was thrown:
Batters have been even less productive against Huston’s slider, with a .194/.237/.333. No, the damage has almost entirely been done against his changeup, allowing a .280/.333/.520 line. It seems that the loss of velocity does not make Huston’s fastball any easier to hit – but the loss of separation from his changeup makes it much, much easier to hit. Perhaps Huston should start throwing fewer changeups until he gets his velocity back. He has not made that adjustment yet, though, as he has been throwing it about 16% of the time all season. Hopefully the couple of hours will not make this article someone else’s problem … unless it is Tampa Bay’s problem.
UPDATES:
Spurred on my grover's questions, here are a couple more charts.
Average velocity and movement of change up by outcome:
As you can see, there is no real difference in the velocity or movement of his change ups based on the results of that pitch.
Separation between average velocity of fastball and change up:
The results are a bit interesting, but somewhat rational. The less separation he has, the more hittable he is -- but in games with the least separation, he also has the best control. Curiously, it is the last group of games when he threw the most change ups. The first two were virtually identical, with 2.38 and 2.32 CU/IP, respectively. The group with the least separation, however featured 3.5 CU/IP. 
As far as Faust's question, it seems that his change up's velocity has more or less tracked his fastball velocity.
