Graph of the Day: The Worst Teams In Baseball, By Position
by Jack Moore on Jul 21, 2009 6:00 AM EDT
in Visuals Above Replacement
Email
PrintUsing Runs Below Average (combining wRAA and UZR), these are the teams that are worst at each individual position. To find runs below average, the teams composite production at each position is added up. These can be found under the Teams tab on FanGraphs, by clicking on each position. Catcher is not included here because there is no fielding component to catcher production on FanGraphs yet.
Some notes:
- The Jason Giambi reunion in Oakland is not working out too well. 1B is a position that you need average production out of to compete (runs below average does not include a positional adjustment).
- The contending teams listed here are the Twins, White Sox, and Braves. The Braves have addressed their problem with Nate McLouth already, but the Twins and White Sox should definitely consider a trade to improve their respective situations at 2B and 3B.
- The Royals still desperately need a shortstop.
- The Padres outfield is... bad.
Runs below average data from www.fangraphs.com
48 comments
|
0 recs |
Do you like this story?
Comments
They are sooo close
They come in a mere half run better than the Padres left fielders with a 19.5 runs below average. All thanks to Ankiel’s positive UZR contribution in left. They are also dangerously close to worst 3B production as well coming in overall at 18.3 runs below average.
by stickman179 on Jul 21, 2009 10:46 AM EDT up reply actions
phillies
there has been a team to get worse production out of the SS position than the phillies?
you must not be familiar with the tales of Tony Pena Jr.
Try to look beyond your own backyard once in a while.
World Series Champs complaining. I outta punch you in the face!
Coffee. The NEW Performance Enhancing drug for Sport's Writers. Just ask Ken Rosenthal.
complaining?
who said i was complaining. i just know that the first half for rollins has been woeful and his fill in is batting .136. shocked that other teams were sending out a product just as bad as and as frequently.
by jamiethekiller on Jul 21, 2009 10:36 AM EDT up reply actions
Shame the padres are on here twice
also, the only reason I bet the Sox are here is because they have praticly been using rokkies all season in third base. Now they have Gordon Beckham there and his WAR will improve by the end of the season
Yeah, the negative contributions at 3B are mostly from Josh Fields.
He’s having a pretty legendarily bad season.
---
http://www.beyondtheboxscore.com
http://www.rightfieldbleachers.com
No Giants second basemen?
Is Juan Uribe being included here…?
Aaron King is still my homeboy... iffy mechanics and all
If Dustin Pedroia played in Seattle, not many people would be talking about him.
GET THAT VORP SH!T OUTTA HERE!!!
not a good graph
Just my opinion, but I don’t find that a useful subject to graph. Seems to me graphs should be used when they communicate something that can’t be easily picked up through numbers. In this case, the graph made it worse. I can’t tell what all those extra bars are for and they’re too closely aligned to be useful.
Again, just my opinion, but straight bar graphs rarely communicate better than a simple table of numbers.
The colors confused me -- I was trying to read meaning into them.
I find bars useful for easy comparison of more/less, but it’s tough to compare, say, all the UZR bars at a glance in this one. Unless looking at the x-axis, you also can’t tell which bars are in a group.
Data communication is an interesting topic, one that deserves some more attention.
Beyond the Boxscore Not a member? Sign up.
tufte
I strongly recommend Edward Tufte’s work if you guys haven’t read him. He’s the guiding light on the subject.
Regarding bar graphs, remember one of his rules: don’t use a graph when simple numbers will suffice.
Right, but "suffice" isn't as cut and dried as it sounds.
don’t use a graph when simple numbers will suffice.
Beyond the Boxscore Not a member? Sign up.
Not sure what you mean
What, in particular, is difficult about it?
Some people like pictures better than numbers.
@bs_uf15bosox9be:OverTheMonster-ALLERGEN WARNING:May contain PB.
Tufte
His graph of Napoleon’s retreat from Russia is just about the coolest visualization I have ever seen. Seconded very strongly.
by Tommy Bennett on Jul 21, 2009 8:49 PM EDT up reply actions
right
Yup, though Tufte didn’t create that. Still, he’d endorsed it as perhaps the greatest visualization ever, particularly for its time.
You're right
Shouldn’t have said “his.”
by Tommy Bennett on Jul 21, 2009 10:34 PM EDT up reply actions
Anything in particular we should start with?
by Dan Turkenkopf on Jul 21, 2009 9:05 PM EDT up reply actions
This is the classic
The Visual Display of Quantitative Information
It is a beautiful book.
by Tommy Bennett on Jul 21, 2009 9:45 PM EDT up reply actions
Agreed
Tufte is hard to categorize into simple rules, but this first book of his comes closest to being truly useful.
This website…
http://www.washington.edu/computing/training/560/zz-tufte.html
… is also a nice summary of his work.
I'm no expert,
But I think the biggest thing to keep in mind is that creating a graph or any kind of infographic shouldn’t be a quick process. Different situations call for different styles and different points of emphasis. We need to consider what the point is we want to get across and then think about how best to get it across.
Graphs, to me, have two purposes: one is to present data, but the other is to assist us in finding patterns and further courses of research. The second type should not be used for the first, at least without some serious reconsideration of what it would take to optimize the point we’re trying to get across.
Beyond the Boxscore Not a member? Sign up.
by Sky Kalkman on Jul 21, 2009 10:41 PM EDT up reply actions
It would have been much better as a stacked graph
Derosa.
by vivaelpujols on Jul 22, 2009 1:09 AM EDT up reply actions
true...
Sorry, I’ll leave one last comment, then stop. You can probably tell I have a passion for good graphics. And I also recognize that some of this is opinion; some is science.
Anyway, it would have looked better as a stacked graph, but it still wouldn’t have been a good graph. (I don’t mean to pile on, just using this graph as an illustration). The numbers in this graph don’t need to be compared. The point of the graph was to list the teams with the least value at each position. Unless the writer was trying to point out the difference between the lowest value at short vs. the lowest value at right field, for example, there was no reason at all for the graph. IMO, should have been a table.
I think this is the big take-away:
The numbers in this graph don’t need to be compared.
Graphs should have visual bang. Do we need to compare the players in the graph? Maybe, but that’s not the main point. The main point is that these guys all suck compared to the rest of the league. So if there was a visual point to be made, maybe we could have shown these five guys’ total runs below average (RBA), the RBA of a rep level player, the RBA of an average player (yes, zero), the RBA for mid-level All-Star, and the RBA for Pujols. Or something.
Or if the point was to show crappy players could fail at hitting, fielding, or both, that should have been more obvious.
Jack hasn’t mail-bombed me yet, hopefully that luck continues…
Beyond the Boxscore Not a member? Sign up.
by Sky Kalkman on Jul 22, 2009 10:28 AM EDT up reply actions
Actually the point was there was no graphs were scheduled, it was 2 AM, and this was something I could put together quickly
Definitely not my best work, and I understand that.
---
http://www.beyondtheboxscore.com
http://www.rightfieldbleachers.com
I'd also add...
I think a lot of people consider saber-style number crunching to be very bland and dense. Technical writers will tell you that long chunks of text and tables of numbers are daunting to the average person, and that breaking up paragraphs with images is a good idea.
While I agree with this idea in principle (I kinda have to, since I’m a technical writer), we have to be careful that graphics also communicate ideas effectively. One of my biggest pet peeves are those godforsaken Brooks Baseball pitchF/X strikezone plot graphs. If ever there were a prime example for graphs that need tomes of text written to properly communicate data, it’s those. Furthermore, drawing conclusions (even accurate ones) from hard to read graphs is likely going to make your audience tune out.
I think you’ve brought up a lot of excellent points here. Sabermetrics and sabermetricians are misunderstood by a great number of people — when I write, I personally make it my duty to explain myself as thoroughly as possible, and any misunderstandings are my fault. Obviously, you sometimes run across people who are just stubborn or (rarely) not smart enough to understand what you’re talking about on any level, but many times it’s a communication problem.
Why do you have a problem with strikezone graphs?
Those are visually simplistic and very informative.
Derosa.
by vivaelpujols on Jul 23, 2009 10:17 AM EDT up reply actions
I read his point is that those graphs are hard to read, harder than they need to be.
Color and shape are hard to distinguish and aren’t too useful.
Beyond the Boxscore Not a member? Sign up.
by Sky Kalkman on Jul 23, 2009 10:30 AM EDT up reply actions
I find them very useful
When I do my Pitch f/x analysis’ at Driveline, I usually just look at a game or two thrown by the pitcher. I show two strikezone plots: one is organized by pitch type, and the other is organized by outcome (swinging stike, ball, foul, etc.).
Then I put them side by side, or right after eachother. I find that is an excellent visual representation of how the pitcher was able to be successful or not during the game, and based on some comments from my readers, they find it pretty useful as well.
Derosa.
by vivaelpujols on Jul 23, 2009 10:41 AM EDT up reply actions
Sky's right.
It isn’t all strikezone graphs, but the Brooks Baseball ones are just misused a lot. When a pitcher throws upwards of 30 dissimilar pitches and you’ve only gone so far to differentiate balls, strikes, and contact (that’s what X is, right? I forget) by color, instead of using shapes as well? It’s just way too much information, and it isn’t differentiated properly.
Here, I’ll go through a random one I picked out.
This pitcher looks like he’s pretty wild, given the number of pitches thrown off the sides of the strikezone. But, are those pitches off the right-hand side indicative of a plan to pitch lefties inside? Abusing an atrocious umpire strikezone? Who knows.
Of course, it’d be more enlightening if I’d followed that particular game or paid attention to the pitch sequence, but that’s precisely my point: You have to have a lot of foreknowledge when you look at these particular kinds of graphs (and I’m not saying all strikezone plots — differentiating by pitch type, batter handedness, inning, etc. makes it much easier to wade through the data). So, when a writer uses a graph like this with 30+ undifferentiated pitches on it, and then draws conclusions about how skillfully or unskillfully a pitcher performed on that given night, it seems very much like voodoo — which is not what sabermetrics is about.
I should also add...
I think conclusions drawn by these graphs can be perfectly valid, but personally I would rather see them represented in a numerical way, like in Alex Krolewski’s BABIP, HR/FB, and Batted Ball Type by Pitch Location fanpost.
The data in these graphs would be vastly more approachable if you broke down the strike zone used in the Brooks Baseball graphs and simply wrote numbers out in those positions. Better yet, use both, and put a numerical chart next to the Brooks Baseball graph.
Intersting thought
How would you suggest applying that to strikezone plots though?
Derosa.
by vivaelpujols on Jul 23, 2009 10:52 AM EDT up reply actions
Break down the strikezone into nine buckets as Alex did, with four buckets for high, low, in, and wide
Then, you could either place a raw numerical pitch count in each bucket, or a percentage of pitches in each bucket (or both, with a raw number and a percentage in parentheses).
That could work
I think that would be best for analysis of a lot of data though; like comparing to a baseline.
Derosa.
by vivaelpujols on Jul 23, 2009 10:57 AM EDT up reply actions
I think that the charts can be useful if used right
As I told Sky, I use them in my articles frequently. Here is an example of where I think it was effective.
Derosa.
by vivaelpujols on Jul 23, 2009 10:50 AM EDT up reply actions
Yeah, that's a vast improvement.
Changing the shapes and using easily distinguishable colors makes those graphs much, much easier to read.
Colors = team colors, Sky.
Use of black for WAR symbolizes the numbers being standardized.
@bs_uf15bosox9be:OverTheMonster-ALLERGEN WARNING:May contain PB.
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Recent FanPosts
Follow us on Twitter!
BTB: @BtBScore
Justin: @justinbopp
Satchel: @SatchelPrice
J-Doug: @RationalPastime
Daniel: @CamdenCrazies
Peter: @CapitolAvenue
Josh: @garik16
Julian: @splashinpumpkin
Matt: @devil_fingers
Bill: @BillPetti
JD: @JDSussman
Lucas: @DBITLefty
Jeff: @jeffwzimmerman
Lewis: @LewsOnFirst
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
