Filter posts by...
Tuesday, May 22, 2012
Even before the injury, Strasburg was getting a quicker hook. This is what happened, start by start:
70 after 6, ended at 82 (12 pitches in last inning)
84 after 5, ended at 108 (24 in last)
61 after 5, ended at 93 (32 in last)
71 after 5, ended at 94 (23 in last)
82 after 6, ended at 101 (19 in last)
76 after 6, pulled
73 after 5, ended at 103 (30 in last)
74 after 4, pulled
78 after 4, ended at 90 (12 in last)
The basic story after the first five starts is that if he finished the inning with under 85-90 pitches, he was sent out for one more inning, even if he labored in the last inning for 20 or 30 pitches (they don’t pull him out until he gets all three outs). The exception to this rule was his first start.
In his next four starts however, it was a 50/50 proposition that if he was still under 80 pitches, that he would be sent out for one more inning.
Normally, I don’t like to look at the “last inning” of a pitcher, because the last inning of a pitcher is usually a pretty bad inning (that’s why it’s his last). With Stras though, it seems that they have a more specific plan, that they don’t really care about his performance. It’s actually a beautiful thing from an analyst’s viewpoint, because we always worry about selective sampling, and we don’t seem to have that here.
Anyway, I wonder if they decided that sending him out in starts 2 - 5 was telling the Nats that maybe they were overly ambitious. He threw a sh!tload of pitches in the last inning in each of his games.
So now, instead of definitely sending him out if he’s still at 70-80 pitches at the end of the inning, they MIGHT send him out for one last one.
A loyal reader was interested in hearing from the Straight Arrow readers:
I’d benefit from a comparison thread of phone operating systems: (iOS vs Android personally, but Windows is a viable alternative, too.
(3)
Comments • 2012/05/22
•
Web Admin
While I am sure I am biased, I’ll include hiring Dan Fox in the mix.
There are different identifiers, with age, experience, and production being among them, and that’s the focus Bill shows in this good piece. It could be that the focus would simply be on the identity of the hitter, but there’s still lots of work today between here and there.
This is a great way to use WPA to start to tell a story. They all have the same thing in common: the Brewers were tied or behind, and then after the HR, were ahead or tied. And it was in the late innings in most instances. We could have used THAT as the english-definition (ties or puts you in the lead, in the 7th or later innings). WPA basically refines that by quantifying it in a continous scale, rather than an either/or binary scale.
Cool piece by Troy, showing the percentage of called pitches were deemed strike, according to the umpire, and the percentage of called pitches were tracked as strikes, according to PITCHf/x. At the league average, both are 45.4%.
. YEAR Zone % Zone % Pitch f/x
2009 51.70% 50.80%
2010 47.80% 50.90%
2011 48.30% 52.90%
2012 43.50% 51.00%
Another knockout interview by David Laurilia. I have several hundred articles and blog posts in my reader every day, so I will usually skim headlines, and for a small portion of them, I’ll open them up, and then just skim the article. In a few rare cases, I’ll read every single word, and that’s what I did here.
I love it whenever a player talks about fielding, and you can tell how much Ryan cares about it.
The following is an exchange between a reader and Bill James, where the reader quotes a report in the Boston Globe to that effect, and Bill James describes a true story where a ball-sniffing dog would separate the two kinds of balls. The seam-thing is the first I heard of it, but, given different standards and different manufacturing plants, I wouldn’t be surprised.
I also remember one year Felipe Alou suggesting that the balls were different (might have been 1997 or 1998), that the balls were “slicker” (meaning harder for the pitcher to grip, either because of the material, or the seams).
Bill, per the Boston Globe, minor league baseballs have more pronounced seams than major league balls (which is why Dice-K uses the MLB variety in his rehab games). Is there really a difference? If so, wouldn’t every team and every pitcher be better served by using the real thing, rather adding adjustment to s different ball to all the other pressures and changes pitchers face when they reach the majors…
Asked by: greggb
Answered: 5/22/2012
The balls are different, but I didn’t know they would let you use major league balls in minor league games. I would presume there would be an economic barrier to the wide usage of major league balls in minor league games--if not an outright prohibition on it. Major league balls are more expensive.
This is a true story; the Red Sox minor league equipment co-ordinator used to have a dog that had been trained to tell the difference between a minor league ball and a major league ball. We would sort the balls into “buckets"--major league balls in one bucket, minor league balls in the other. Major league teams have dozens and dozens of buckets of balls around for batting practice and such like. Anyway, if there was one major league ball in a bucket of minor league balls, that dog would smell it, and he would remove all the balls from the bucket until he found the major league ball, put it in the major league bucket, then he would put the minor league balls back in the minor league bucket. True story.
Alan posted a letter from (.doc file) the testing facility that says:
The Major League balls are manufactured in Costa Rica and have a compressed cork sphere per the specifications. The Minor League balls are manufactured in China and have a cork center as specified in “1996 Minor League Baseball Proposal”. This cork center is the likely source for the decrease in performance, which results in a comparable Minor League ball hit of 391.8 ft under the same conditions as the Major League balls [at 400 feet].
It should be noted that an 8 to 9 foot drop in batted ball distance would lead to a 25% drop in the number of home runs. It’s all well-and-fine to say that there might be an eight foot difference in home runs due to the difference in ball configuration, and to you and me, that sounds like a small number, but the reality is that it has a tremendous impact.
There’s also this quote, that talks about how a ball could meet specifications, but would have a drastic impact:
This means that theoretically, two baseballs could meet the specifications but one ball could be hit 49.1 feet further than the other could be hit. This 49.1 feet is the combination of the increased distance of 8.7 feet for the ball being on the light side with respect to weight (i.e. 5.00 oz. as opposed to 5.25 oz.) and an additional 40.4 feet for the COR being biased to the high side (i.e. 0.578 versus 0.514). However, it should be noted that the balls investigated in this study did not exhibit this potential 49.1-ft difference. Thus, the tested baseballs indicate that the 1999 and 2000 baseballs fall within a tight range of batted-ball performance and that the 1999 and 2000 baseballs are for all practical purposes the same with respect to batted ball performance. The 49.1-ft value is purely academic—it was not seen in the balls tested.
Basically, the plants do a great job of producing balls within a tight range, but the specs allow for a huge margin of error.
Monday, May 21, 2012
I use Firefox out of habit. I have to use IE, because that’s the only browser my office supports for remote access. I’ve never used Chrome.
I’m interested to hear from those who have use Chrome, and how it compares to the other two.
(36)
Comments • 2012/05/22
•
Web Admin
I was wondering if I was the only one crazy enough to notice this, but, no, there is someone crazier than I am, who went back to the tape, and did screen shot after screen shot. If you think only us baseball nuts go overboard, Bill is our TV counterpart. Welcome to the club, Bill.
(1)
Comments • 2012/05/21
•
Blogging
Sunday, May 20, 2012
Here’s Lincecum covering home plate.
The runner was pretty nice to him on the slide. My question: had the runner launched himself at Lincecum, and taken him out (and likely injured him), would anyone have a problem with that?
I was emailing with a couple of former players, trying to understand the reluctance of the power-hitting LHH to bunt. The basic theme eventually led to money: they are paid to be power hitters, not bunters, much like we’ve heard in the past about power hitters hitting the ball and not taking a walk. Now, we’ve incentived the power hitter to take a walk by (a) rewarding the value of the walk and (b) realizing that if a player is more selective, not only will it increase his walk total, but he may end up hitting better overall too.
So, these power hitters are not learning to bunt, because there’s little incentive to do so. While this may be a team game, players are paid for individual production. As long as it looks like bunt singles don’t help them too much, they are not going to waste their time bunting. The defense knows this, and are exploiting this inefficiency by over-shifting. The response to the over-shift is to bunt, and this will quickly force the defense to not overshift so much. But, the power hitters are not doing it, and the managers are not forcing them to do it. (You’d hate as a manager to order a hitter to bunt, when said hitter doesn’t have the confidence to lay one down.)
Now, how do you incentivize the power hitter to better appreciate that bunting will not only drive up his batting average, but will eventually lead to the defense having a fielder alignment that will help him as a hitter when he does swing away?
MLB doesn’t allow bonuses, except at the team level. I was thinking that if you create a reward system so that you overweight the thing you want them to do, they will quickly move to that area. If for example you tell the hitter than any bunt will count as a double in contract negotiations, then this would appeal to the hitter because he gets to increase his “extra base hits” totals.
How do you get it through their heads so that it becomes a natural thing for a power-hitting LHH(*) to bunt in the face of an over-shift (contrained by current MLB guidelines)?
(*) Carlos Pena excepted. Maybe teams can hire him as a motivational speaker.
Saturday, May 19, 2012
Nothing available yet, but you can get a leg up by applying now. Tell ‘em you heard it from Tango, as it’ll help. And, if you are Canadian, that probably helps too.
I love it when I see a pitcher who just threw a complete game shutout say this:
Adam then asked Millwood how he felt in the bullpen before the game got started, looking for Millwood to say he felt sharp. “Terrible” was Millwood’s response. He didn’t have anything. He went out there anyway and threw a complete-game shutout in Coors Field.
Friday, May 18, 2012
A stream of consciousness post from the usually jocular Brandon McCarthy.
The reality is that the umpire, like everyone else, INFERS what he sees. Things happen fast, there’s obstacles, the glove gets in the way, the grass is a certain height, you have a certain angle. You then infer based on past experience what you see, and then you have to claim “I saw this”, when the reality is he should say “Bayesian inference would suggest that I saw this”.
And I think if an umpire actually said that to McCarthy, Brandon would have no choice but to accept it.
Cory Schwartz is our counterpart over at MLBAM, and I’ve dealt with him alot, in an official or unofficial capacity over the years. Wonderful guy. Anyway, a reader sent me a link (which I can’t hear at the moment), along with his comments:
He’s the VP of Stats for MLBAM. Nothing too earth-shattering, though the discussions of [an MLB insider] disdain for Pitch f/x and Field f/x being public was interesting. Schwartz basically defended the idea of keeping that info private, unfortunately.
Schwartz nicely responds to those [basic statistical] types of comments. And Schwartz had a couple of great points about pitcher wins, saves, and LI near the end.
Overall, some good, broad overviews of what’s going on with data, the people behind it, and teams using it. His segment runs from about the 6-minute mark to the 34-minute mark.
I follow the NHL, I occasionally keep up with the NFL, and if I pass by the NBA, I’ll stop to take a look. And when I was living in Canada, I was an avid CFL follower. None of the media that follows those sports ever make a big deal about inter-conference play. Indeed, they never make any deal about it. At the start of the season, you may be told how often a team plays their inter-conference opponents, but that’s it.
The media trips over themselves to bring out the “news” that there’s inter-conference play in MLB. It’s not 1996 any longer. Is it the remnant of the DH rule not being uniform? Is it that there’s only 10% games that are inter-conference, in comparison to the 20%-25% for the other sports? Or, does the media simply love “press release” reporting, because it’s the easiest way to make a buck?
Also pay attention to the “extra teams in the playoffs” “story” that will likely play out. That is also a falsehood. It’s a play-in game. The number of teams in the playoffs is the same. Whoever wins the division will have have practically the exact same chance to win the World Series, regardless of the existence of the play-in game. The extra game impacts only the wild card team, and it reduces their chance in the new format. That could only be a good thing, not a bad thing.
Reject the media narrative this weekend. It’s as lazy as it can get (like my post here).
Apparently the Indians’ view on batting orders intersect to a good degree with what’s in The Book.
Thursday, May 17, 2012
DER is a Bill James stat, called defensive efficiency record. It is a team’s Out Rate on Balls in Play. A team has 30 balls in play, it gets 21 outs, it has an out rate of .700. That’s all that DER is.
Now, what goes into an out rate on balls in play? There are four things, in order:
1. Random variation on balls on play
2. Fielding on balls in play
3. Pitching on balls in play
4. Park effect on balls in play
If you do things right, and you have all the data you need, you can separate all four things. And all four things will be INDEPENDENT of each other. That’s an important point.
What does it mean when two things are independent? Well, if one zigs and the other zags, then they are NOT indepedent. If one zigs and the other zigs, they are NOT independent. If one zigs, and the other alternatively zigs then zags, then are NOT independent. Being independent means that when your wife tells you to take out the garbage, you will sometimes watch TV, sometimes go online, sometimes play with your kid, sometimes take out the garbage, sometimes take a shower.
If a team has great fielders, what does that tell us about random variation, pitching, and the park? It should tell us nothing at all. That’s because a team that has great fielders won’t necessarily have good or bad pitching or play in a pitchers or hitters park or have more good or bad luck go their way.
The Toronto Bluejays have the best out rate on balls in play so far in 2012. They have made 27.4 more outs than the league has. Why? Well, it’s any combination of the above four effects. According to UZR, the Jays fielders are -2.2 runs, which I translate as -2.9 outs, relative to the league average. This means that random variation, pitching and park effects accounts for the difference between 27.4 outs and -2.9 outs. Those three parameters are worth +30.3 outs.
The Rangers are #2, and all four parameters account for +21.2 outs, of which UZR implies that the fielders are worth (coincidentally) +21.2 outs. So, the other three non-fielder parameters are worth 0 outs.
So, it’s a nice thing to see that there doesn’t seem to be any relationship between UZR and the other three non-fielder parameters.
We run a correlation of the outs saved based on UZR against DER the other three non-fielder parameters, and we see that there is a correlation of r=-0.25 (negatively correlated). We should have gotten 0. Now, maybe there’s a good reason that it’s not 0. Maybe a team that’s big on pitching doesn’t need big fielders. There’s also a bad reason, and that is that UZR doesn’t adjust eveyrthing well-enough. There’s a bias there. Seemingly anyway, given we’re only looking at 6 weeks. (If someone wants to run the same study for 2011 and earlier, feel free. I need other people to step up here.)
I decided to repeat this with DRS, and the correlation there is r=-0.62. That is a pretty large negative correlation. Of the top 10 teams in DRS (average of +15 runs, or +20 outs), their “non-fielding” parameters account for an average of -14 outs. Similarly, the bottom 10 in DRS, average of -19 runs or -25 outs, has their three non-fielding parameters account for an average of +13 outs. Basically, DRS seems to “over adjust”.
Below you will find my dataset. Feel free to replicate for past year.
Note that a few minutes ago, I asked Colin to send me his FRAA data. Once he sends me that, I can run it and compare to the others.
One interesting point is that there is an easy way to get r=0, which is our goal: just randomly make sh!t up! So, the perverse thing is that, the best way to get an r=0 is to either have a highly effective system OR a highly random system. The key is that we’re trying to find something with no bias. A highly effective system is the goal, but, random numbers will also show no bias. Presumably, all the smart fellow working on fielding systems are geared toward a highly effective system. But, I just want to put it out there that while r=0 is the goal, there is a little thing to remember here.
Read More
Whether by bunting, or going the other way, it all works.
from Amazon
Recent comments
Older comments
Page 1 of 344 pages 1 2 3 > Last »Complete Archive – By Category
Complete Archive – By Date