Some weekend links
January 25th, 2009 | Published in baseball, links | 10 Comments
by Myron Logan
At The Hardball Times, John Walsh takes his annual look at outfield arms. It’s, as always, a very straight forward, interesting analysis. Brian Giles comes in as the worst right fielder in 2008, at –13.6 runs (per 200 opps). Here are Giles’ numbers since 2004 (from Giles’ THT card):
2004: –3.5
2005: +1.2
2006: –3.6
2007: –7.6
2008: –13.6 -10.3
I know I brought it up a while ago, but personally I think it’s time to consider moving Giles to left field. It’s not just the numbers either, with regards to his arm. Here what the fans have to say:
| Year | Release | Strength | Accuracy |
| 2005 | 61 | 39 | 55 |
| 2006 | 52 | 39 | 59 |
| 2007 | 52 | 28 | 53 |
| 2008 | 59 | 19 | 47 |
You can see that, in line with the numbers, Giles’ arm has declined over the past few years. Note that by arm, I actually mean “ability to get to the ball, catch it, release it, etc.” Anyway, I think it makes sense to move Giles to left to hide his arm a bit. As MGL reports in this thread, the range in right field is like plus or minus 10 runs (per season), while in left it’s only +/- 4 or 5. You reduce the amount of throws that need to be made, important ones anyway, by moving Giles to left. Kind of like putting a bad hitter lower in the line up.
I know that some (including Geoff Young, IIRC) have argued that Giles plays the oddities in Petco’s right field well, and that may be tough for a new guy to learn. Also, Giles is probably well acclimated to the ball coming off the bat a certain way, and stuff like that. No doubt, overall he’s still probably a plus fielder in right, so it’s not really a major problem. I just think that it’s getting to a point where a move to left is necessary, or at least quite clearly beneficial to the team.
****
Sticking with THT, here are a couple of other interesting pieces from last week:
- Colin Wyers on total player value
- Greg Rybarczyk applies Hit Tracker to high schoolers
****
This discussion, which drifted over to The Book Blog, is pretty thought provoking. It’s about Jeff Francoeur’s value (or, more accurately, his marginal revenue product) in 2008, which J.C. Bradbury says was $12 million. The ensuing discussion is pretty fascinating, though I will admit to being too ignorant to really weigh in.
****
Baseball Prospectus announced that their 2009 annual is ready to go. The comments there (just about all ~170 of them) centered around the foreword that’s being written by Keith Olbermann. Seriously, I can’t believe that people care that much about the foreword … but what do I know; I guess they do.
The part I’m most looking forward to, outside of the projections and team analysis (and of course the foreword!), is Clay Davenport’s “dramatically revised fielding runs.” I wonder if he’s now using play by play data. My guess is that he is, because I’m not sure how dramatically he could improve the old stat; I believe it got just about all you could get out of traditional fielding stats.
****
Rotoworld is looking for writers. I thought about it for a second, then read the description: “…providing coverage for Rotoworld’s player news page that requires the ability to report news with instant analysis and recap games in a clear, concise style.” If that said like “delayed analysis and recap games in a long, rambling style” I might have given it a second thought : ) On a serious note, it’s a really good opportunity for someone. From Gleeman’s blog, it looks like the positions are still pretty open (as of Friday anyway). Past writing experience is a biggie.
January 25th, 2009 at 2:44 am (#)
I get -8.32 for Giles’ arm ratings if you weight then at .10, .20, .30, .40 from 05-08. If you weight his games played and UZR (not UZR/150) the same way you get 140 games and a total UZR of 4.42. If you extrapolate his weighted arms runs out for 150 games, you get -8.91 runs and take his weighted UZR and extrapolate it out over 150 games you get a rough average of 4.74. Basically, he turns from a +4 defender to a -4 defender. That’s a pretty decent swing with some effect on his value.
Take a -4 projection for 2009, his CHONE wOBA of .346 over the 73% playing time that projects him for, you get 1.24 WAR (13.06 RAR). If you give him his +4 defense you get 2 WAR (21.06 RAR if you’re going by the strict 10.5 runs/win.
That adds up to, given $4.4 on the open market this year +$400K to being $5.86 million for the bad arm incorporated and $9.2 million without the arm in the mix.
(assuming I didn’t screw up the late night math, obviously).
That’s a hefty difference of $3.34 million which in pro sports normally isn’t an astronomical amount but for a team like the Pads who need to shed payroll, and in an economy like this todays, $3.34 is a very big deal - one in which valuing his arm into his defense properly and placing him in what will hurt the team less in the field (hiding his arm in LF), to maximize the wins is vital.
January 25th, 2009 at 11:28 am (#)
Bah! I don’t believe you messed up, Mike, but I did. The 13.6 I had up there for 2008 was per 200 opportunities, the rest were just season totals. So 2008 should say 10.3 and, doing the same thing you did, I get a weighted average of -7. So, just a difference of like 1.3 runs, but noteworthy nonetheless.
Still, your post above pretty much holds true. He goes from being an above average defender to a slightly below average one.
January 25th, 2009 at 3:45 pm (#)
The Tango Fans Fielding thing is a load of horse hooey.
Anything that is based on untrained observers responses and a very small sampling size is certainly far from accurate.
That someone who claims to want to quantify fielding like Tango can even trot that out there as some kind of measure of ML players fielding prowess throws much of what he is trying to accomplish into question.
The fact is that allowing 60 un-vetted people throw their opinion into the ring is certainly not a scientific way to measure fielding. In fact any good pollster would laugh in Tango’s face if he tried to enter that as a poll. The error margins are so large as to make the entire thing invalid.
I really like HBT, but any quantitative FEILDING measure that starts with “Using play-by-play data” is going to be inaccurate.
Play by play data cannot take into account where the ball is hit or the dimensions of the ballpark. How can it accurately measure a players arm? The answer is it can’t.
Every game is on tape so why not just hire a researcher with knowledge of the game and the situation to look at the tape and measure the distance traveled to the ball and the distance, accuracy and efficacy of the throw.
To do otherwise is a disservice to those who are reading your supposed stats.
I work a more than full time job and still find time to do that for the Padres regulars. Yes I know I am a geek, but at least when I talk about the range or the arm of a player I don’t have to base it a formula based on play by play data. Its on evidence gathered by watching tape, the same way the teams do it.
The ONLY reason to move Giles to left would be to make the team better and since there is no one else to play RF or the other outfield positions, it would not help the team to move him right now.
Headley can’t play CF or RF. Gerut can play RF but Hairston, the only other player who can man CF is not as good as Headley with the bat.
Moving Giles makes the team weaker right now. Go pick up a good RF bat and trade Kouzmanoff for a good #2 starting pitcher and THEN move Giles to LF. THAT would make sense.
Moving him now just because some terribly inaccurate metric says he can’t throw is not smart baseball and the Padres are certainly not considering it at this point.
January 25th, 2009 at 4:27 pm (#)
I read that Francoeur article on JCs blog a while back, and considered writing a post on it. I’m glad to see Bradbury’s unintuitive claim is getting scrutinized so heavily. While his understanding of MRP as it applies to baseball may be right (I am actually not so sure, but lets just assume he knows better), his model is really not relevant. The concept of replacement level is real, and because it is, it makes no sense to say a player with below replacement level performance was worth $12 million. (Note: As many commenters at the book blog pointed out, it also makes no sense that pitchers are vastly undervalued compared to hitters).
January 25th, 2009 at 5:12 pm (#)
Web, you undoubtedly bring up some good points. You are going up against a sabermetric giant on both issues (Tango, of course), and while I don’t like just giving someone the benefit of the doubt, Tango has certainly built up a lot of credibility in my book.
Let’s start with the FSR first. My guess is that the people that vote on the report are, in general, pretty hardcore fans. They probably don’t review the tapes and go that in-depth, but my guess is that they watch most of the games and are pretty knowledgeable about baseball. I think there is certainly some value in gathering observations of a bunch of people who are knowledgeable about a subject.
My biggest question with the FSR has always been that the participants are too biased by past numbers, and that their observations are a bit clouded by that. Does anyone (with some familiarity with stats) really view Derek Jeter as an average fielder to start, and then change their opinion based on his play in the field? Or do they imagine him as a -20 run fielder, and then look for plays that justify that? My take has been that it’s more the latter.
Still, I think the fans scouting report has some value. It should be added to the fielding metrics, imo (I’m not sure how much you’d want to weight each; probably depends on how many innings the player has played). But, no doubt, I’d prefer a SSR (scouts scouting report) based on a bunch of scouts that follow the players or a video analysis team — but I think we have to work with what we’ve got, and nobody has that kind of access.
Regarding the arm ratings, I’m not saying it’s perfect (certainly far from it), but it’s about as much as you can do with the data, I believe. And I think that over a period of time, it’s going to show the best and worst arms quite often. I wouldn’t quibble about a run or two, as I’m sure there’s a large margin for error. But don’t all stats have that built in?
Can you really look at a guys hitting stats and tell how good he is? They don’t tell you the park, how the fielders were positioned, the pitcher, where exactly the ball was hit, etc. A double is a double, whether it’s a little bloop that falls in between two guys, or a rope that hits the top of the wall.
I think that the argument the stathead would use, would be that over a long period of time, things like that tend to even out and you get a pretty good representation of someones hitting ability based on his traditional numbers. But, yeah, I’d rather have speed off the bat, angle off the bat, spin, etc., as it would probably give you a more accurate picture of a hitters performance.
Anyway, I appreciate the comments. They make me think more : )
January 25th, 2009 at 5:14 pm (#)
Daniel, feel free to make that post still : ) Would like to hear your thoughts … I’m still having some trouble wrapping my head around the issue, though that should not be a surprise.
January 26th, 2009 at 12:47 am (#)
My biggest problem with JC Bradbury’s methods are that he doesn’t defend anything with anything except “it’s my model and I explain it in my book so I don’t have to do it anymore on this blog that I run.” That’s certainly his right, but then he says things like “point out my mistakes.” One can point out a mistake to Tango because he documents all of his stuff, but Bradbury’s claims are all based on his model. And he’s writing a new book and has claimed his model has changed. that book isn’t out yet, but we’re supposed to show where he’s made the mistake on Frenchy? What? How? Huh?
January 26th, 2009 at 1:44 pm (#)
Rally’s response to the Bradbury debate:
http://lanaheimangelfan.blogspot.com/2009/01/ballplayers-and-their-marginal-revenue.html
(h/t: Tango)
January 29th, 2009 at 3:33 am (#)
In the past I have posted elsewhere about this and maybe at some point I will actually take the time to research more than just one position to counter some of the rankings.
One thing I do know from my own research from watching tapes is that Gerut had two of the top 10 runs in distance and 6 of the top 50 in CF that resulted in catches, but the metrics don’t seem to show that he has that type of range.
No, I didn’t watch every game in the league. Before the season started I asked at least 2 people that are, as you put it pretty hardcore fans, from each team around the league to watch their teams games on MLB.TV and recommend games for me and my two employees to watch and we also watched every Padres, Dodgers, DBacks, Giants and Rockies games that were broadcast.
We used the positioning of the players and the dimensions of the parks along with a CAD like program to plot and then fairly accurately measure the distance the CF had to cover to make the catches. Not perfect, just fairly accurate.
Our original intent was to do this for all OF positions, but we already have hundreds and hundreds of hours into this and we are pretty sure of the CF data, but we haven’t even started on any other position.
Yes I know I am a geek, but I am so obsessed with having at least a small amount of fairly accurate fielding data based on observation, that I actually spent about 4 grand over the past 4 months on having a couple of my employees moonlight with me watching game after game and programming the software.
We were also just looking at ONE part of what goes into being a good fielder, range.
As a coach we measured release time (time from catch to throw) and % percentage of errors and % of balls in their areas of responsibility (we told them what their area of responsibility was based on their times in the 40 and long shuttle) that they caught and the number of assists and to what base, but I never charted where the player was on the field and where the ball landed on the throw from watching film.
I really don’t have the resources to take my project to the next level and I am pretty sure my wife will murder me if I even try to recreate what we have done this year, but someone should do it other than the teams.
February 18th, 2009 at 10:35 pm (#)
Web, I think this is the first comment I’ve entirely missed on this blog — until now, obviously.
I think what you are doing is great. If someone else could take it up, that’d be awesome. If you haven’t already, you should email Tango, mention what you’ve done and see if he has any ideas. He’s always very helpful.
Any chance you could post more of what you did, or what you found, on your blog or something? I am interested.
One question I have … did you measure how long the ball was in the air for? Obviously, a line drive compared to a high, deep fly, even if they land in the same spot, are of a different difficulty level for an outfielder.
Also, I think Petco holds the ball up in the air longer, on average, so I wonder if that is part of the reason why Gerut shows up near the top a few times.
Anyway, good stuff …