How Common is a One-Win Above Replacement Player?
by Justin Bopp on Nov 16, 2010 11:00 AM EST
in Visuals Above Replacement
Email
Print
please expand by clicking the graphic.
Notes after the jump.
To answer the question posed in the title, some clarification is in order. "How common is a one-WAR player?" Well, if you're saying "exactly one" or "1.0," not very common. In fact, there are only TWELVE exactly-one-WAR players. It's my belief, though, that the question isn't really asking that and is instead attempting to find the commonality of a range of slightly-above average players.
To do just that, I grouped the data into sections which include every tenth's place after zero for the integer in question. There were 18 players between -2 and -1, 326 between -1 and 0, and so on. In case it's not explicit, each circle represents a single player at a single one-tenth iteration of a particular WAR. For example, the section at the top of the graphic which is listed }5 represents all players from 7.0 - 7.9* Wins Above Replacement (of which there are 5). Where there is a grouping of players on a single line (see the zero line), each has the same WAR.
Back to the question in the title. The better answer is, "38% (369) of all players in 2010 earned between zero and one WAR, while only 83 players (9%!) earned between one and two." Clearly and obviously, and I think the entire scope of the curve captures this well, players that earn greater than one WAR are increasingly rare.
Data sourced from Fangraphs, "all players" for 2010.
* Josh Hamilton's 8.0 was accidentally included in the 7.0-7.9 grouping. All other integer groupings start at x.0 and go to x.9.
44 comments
|
4 recs |
Do you like this story?
Comments
Where was zero put?
I imagine 0.0 WAR is a fairly significant number (low PA & IP guys)… how did those guys get split up between the -1 to 0 and 0 to 1 groups?
On Twitter: @baseballtwit
Good question.
Everyone at 0.0 and above is included in the 0 to 1 list.
Everyone at -0.1 to -1.0 is included in the -1 to 0 list.
Read: every group of data includes the every instance of a player at that WAR that begins with the digit in question. (0.0-1.9, 2.0-2.9, 3.0-3.9, etc)
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
by Justin Bopp on Nov 16, 2010 11:37 AM EST up reply actions
And it looks like I made an error at the very top of the chart by
including Josh Hamiltion’s “8.0” (that’s pretty good, guys) in the 7.0-7.9 data.
Everything else appears consistent.
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
by Justin Bopp on Nov 16, 2010 11:39 AM EST up reply actions
If anyone requires explanation,
It’s because
Josh Hamilton is such a great team mate, he didn’t deserve to be left alone.
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
by Justin Bopp on Nov 16, 2010 12:26 PM EST up reply actions
Are pitchers included by the fWAR as batters?
There were 5 batters in 2010 with an fWAR from 7.1 to 8.0, as reflected in your chart, but Cliff Lee was a 7.1 fWAR pitcher in 2010… so it seems that someone’s missing.
I’m guessing that you used the “all players” batter fWAR… which means that poor Cliff is included as a -0.1 fWAR hitter… is that right?
Yes -- this is "all players" batter fWAR.
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
by Justin Bopp on Nov 16, 2010 11:42 AM EST up reply actions
That really skews the data
Most pitchers batter fWAR is less than 1, while their pitcher fWAR is typically significantly higher.
I think this is a great concept and I love the idea, but I’d love to see how that data looks with either pitchers removed, or pitcher’s fWAR for pitchers.
"I signed up for this job, the day I was born" - Brian Wilson, Ninja
by Giant Torture on Nov 16, 2010 4:18 PM EST up reply actions 1 recs
I think this is a valid criticism.
A quick review of the data shows something like 300 players with less than 30 PAs. That’s no good.
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
While we're at it,
The way FG has their data, I can’t just cut out pitchers (why no position, guys?), but I can sort by PAs. So what should be the cut off?
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
You can do the opposite, though...
…by selecting “pitchers” after selecting “all players” you get just the pitchers’ batter WARs…
…I’m trying to think of the simple way to take that info and use it to eliminate pitchers from the initial dataset….
wrangling the data into a common format
is always my biggest issue. Why can’t data hosts know what I’m thinking?
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
If you set the cut off at 71 PAs, you would get only 20 pitchers.
…that would also eliminate a few september callups, I suppose, but maybe that’s actually an improvement…
The position adjustment will be 0.0 for pitchers and non-zero for others
(Well, probably 99% true — if a pitcher played another position or a non-pitcher happened to have his position adjustment balance out to zero, you’re screwed.)
Don't you want essentially "full-season" guys?
Or maybe that’s what I want ;)
I’d like to know the answer to this similar question: How common is a 1 WAR player who is a starter? (Which is to say, I don’t care if Chris Getz or September Callup Player A gets between 0 and 1 WAR, so I’d like to filter them out.)
The only people who really know where [the edge] is are the ones who have gone over it.
by SagehenMacGyver47 on Nov 16, 2010 4:54 PM EST up reply actions
We could just limit it to "qualified for the title,"
if you’d like?
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
Considering Sagehen's comment...
…which seems like a reasonable way to define the desireable dataset for this question, I would suggest:
If we assume there are 12-13 position players on a Major League team, we should expect there to be roughly 375 batters in the sample, once it is defined correctly.
You can get a sample of that size by using a cut-off of 170 PAs. Since no MLB pitcher gets that many trips to the plate, you will also be assured that the pitcher data is no longer an issue.
And I don't want to forfeit the original graph, which is interesting in its own right
just would like to see the “qualifiers”, too.
The only people who really know where [the edge] is are the ones who have gone over it.
by SagehenMacGyver47 on Nov 17, 2010 7:11 PM EST up reply actions
I want to see this for pitchers.
Glad I came, just wish I hadn't stayed so long.
People ask me what I do in winter when there’s no baseball...Rock Chalk Talk
I'll elaborate briefly
While the display of the graph is very cool, I would like an alternative that eliminates the inconsequential players (i.e., Chris Getz or September Callup Player A) and shows only starters (and/or shows WAR/150, but that is its own can of worms).
The only people who really know where [the edge] is are the ones who have gone over it.
by SagehenMacGyver47 on Nov 16, 2010 4:57 PM EST up reply actions
Another way to point this out...
…Most of the NL Cy Young candidates show up in the -1 to -2 range:
Tim Lincecum: -1.1
Roy Halladay: -1.2
Ubaldo Jimenez: -1.3
Josh Johnson: -1.1
That’s not surprising, since you have to be a pretty good pitcher to get as many plate appearances as these guys got… which further emphasizes the reason that most pitchers have a batter fWAR pretty close to zero.
That said, I hypothesize that the conclusion remains -
“one-WAR players (or thereabouts) are relatively rare.”
Removing a good majority of the zeros by excluding pitchers and/or players with less than some number of PAs (30? 50?) will actually serve to accentuate this.
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
Probably true...
…but I do suspect you find a fair number of relievers in the 1-2 WAR range if you use pitcher WAR. Not likely enough to change your conclusion, though.
Very telling.
I think most fans underrate 1 WAR players and under appreciate the role they play on every team.
When you're drowning, you don't say 'I would be incredibly pleased if someone would have the foresight to notice me drowning and come and help me,' you just scream.
Indeed
2/3rds of MLB players are actually between +1 and -1 WAR. I sort of assumed (because I am too lazy to actually go and look) that average teams would have a few negative, 0 and 1 WAR guys, but the bulk would be at above 1. That’s not the case; average teams are made from guys hovering around the 0 WAR mark with the odd sprinkling of >1 WAR players. The better teams may have more >1 WAR players and worse teams less.
2/3rds of MLB players are actually between +1 and -1 WAR.
Actually, it’s closer to 3/4ths (there are 695 players in that range out of a total of 944). But, 311 of those 695 players are actually pitchers, and batter WAR only reflects their contributions while holding a bat… so the overall picture is a little distorted in that regard.
Lots of those pitchers are relievers who likely didn’t generate a whole lot of WAR while on the mound, but still, all MLB pitchers combined generated 472.4 fWAR in 2010. So, if you wanted to draw overall conclusions about the value of a 1 WAR player, you’d have to account for those 472.4 wins in your assessment.
No
Nope. Pretty sure that relievers get their fair share of accolades :)
by goeaglesxxxix on Nov 16, 2010 12:25 PM EST up reply actions
Pedro Feliz, -2.1 WAR.
That’s what happens when you don’t have any power, don’t walk, have an extremely low BABIP due to an incredibly low line drive rate and a very high infield fly rate, and play bad defense at a non-premium defensive position.
Baseball is my preferred sport. It should be yours, too.
I'm an editor for Beyond the Box Score, an SB Nation blog.
by Satchel Price on Nov 16, 2010 3:16 PM EST up reply actions
and actually get enough playing time to be allowed that much of an impact
see: Tony Pena, Jr.
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
(though he played middling defense at a premium position)
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
Yeah
That’s a problem I have with the whole “replacement level” concept, actually.
by BlueEyes_Austin on Nov 16, 2010 4:06 PM EST up reply actions
No its not
You have a problem with the replacement level that Fangraphs is using.
Come check out Bullpen Banter!
Follow Bullpen Banter on Twitter
Follow me on Twitter
Remember: baseball guys... baseball...
No
it’s more philosphical than that, actually. Don’t get me wrong, I think the notion of “replacement level” is a useful one. I just think that finding freely available talent that can perform at 0 WAR isn’t as straightforward as it suggests.
by BlueEyes_Austin on Nov 16, 2010 4:53 PM EST up reply actions
<3
The baseball season doesn't have to end! Create your own players, coach your own teams, and join your friends in THE premier baseball MMO. Two Out Rally opens October 25th!
Two Out Rally, BASEBALL MMORPG | Facebook | @2OutRally
There's a difference...
between players who perform worse than replacement level and players whose true talent is worse than replacement level.
by Sky Kalkman on Nov 16, 2010 5:02 PM EST up reply actions 2 recs
There's also a distortion at work here based upon the way that actual rosters work.
Replacement level players are, by definition, “freely available,” and Major League Rosters are constructed in a way that takes advantage of that fact. So, when Scrubby McScrub isn’t cutting it in the Show, he is eventually replaced by Johnny McNewscrub. This means that at the end of the year, we have two scrubs in our dataset representing only one roster spot (note that there are 750 regular roster spots in the league, but 944 ballplayers in our dataset).
That suffling isn’t happening to the 2-WAR players, so when you go through the league just counting the number of guys who are or are not near replacement level, you should expect a large concentration near zero. Add some distribution based on sample size (the scrubs’ numbers are, based on small samples), and you would expect to see large numbers slightly above and below the “zero-line” …exaclty what the data here shows.
And adding to this is the “pitcher issue” discussed above, by which roughly 300 ballplayers are being added to that cluster around (and actually just below) the zero-line. I’m getting really curious to see what this graph would look like with a reasonable “minimum plate appearance” threshold applied.
by kcemigre on Nov 16, 2010 6:46 PM EST up reply actions 1 recs
Quick and dirty fangraphs data look
I got 442 players if you limit to 100 PA.
85 were sub-replacement level
128 were between 0 and 1 WAR
81 were between 1-2 WAR
200 PA drops that number down to 346 players.
45 are sub-replacement level
80 are 0-1 WAR
72 are 1-2 WAR
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
Josh: @garik16
Julian: @giantsnirvana
Bill: @BillPetti
Lucas: @DBITLefty
Jeff: @jeffwzimmerman
Lewis: @LewsOnFirst
