Replacement Players

[AngelsWin.com]

By Robert Cunningham, AngelsWin.com Contributor - 

As saber metrics have become more prominent throughout Major League Baseball, the term “replacement level player” has become a part of the baseball vernacular.

The term applies to a generic “average” player that most teams could call up or trade for at any time to provide a neutral value in comparison to other players.

Wins Above Replacement, otherwise known as WAR, is a fairly comprehensive value that tries to value players above or below this generic, average player. The generic player typically has a value of zero WAR that is used as the baseline (however this isn't usually the case in reality).

FanGraphs also defines a replacement level player as a player who costs no marginal resources to acquire. This fits into the overall discussion of what a replacement player is.

This leads to the following blind, side-by-side player comparison (projected out to 600 at-bats):

[YouthofToday]

Good article.

[Brent Maguire]

Good article. Nelson is a serviceable guy at 3rd and if Green can end up being a 2-3 win player, the Callaspo trade is awesome. 

[Dave Saltzer]

Good analysis. I was never a big fan of Callaspo, and to see that Nelson is providing the same WAR is good to know. And, as someone else pointed out, if Green can do better, GREAT!

[Angelsjunky]

Nice article - its good to hear folks using WAR in a clear manner. That said, I have to take issue with a couple things that I think, in the end, mar your argument.

 

First of all, you say that a replacement player is a "generic 'average' player." Generic, yes, average, no. You clarify later on by saying that "a replacement level player as a player who costs no marginal resources to acquire." Meaning, every major league team has a slew of replacement players waiting in the wings - which is essentially what a AAA veteran is. But if this were an "average" player, then no major league teams would have less than average players at any position, which would in turn drive up what "average" means.

 

A replacement player is below average. If replacement level is -1 to +1 WAR, and a good platoon/marginal regular is 1-2, an average starter is around 2-3.

 

Secondly, I don't think the Nelson-Callaspo comparison is fair because 1) it implies that Callaspo's 2013 performance (0.0 fWAR - exactly replacement level) is his true talent level, when in fact he's averaged 3.2 fWAR over the previous two seasons, and 2.4 over his previous four years of being a major league regular, 1.9 if you include 2013. No matter how you look at it, I think it is safe to say that his true talent level is more in the 2-3 fWAR range, which makes him the definition of a league average starter.

 

Chris Nelson, on the other hand, has accumulated a -1.6 fWAR value in 234 major league games, making him nearly the definition of a replacement player (at best) - and thus "a player who costs no marginal resources to acquire."

 

So Dipoto did NOT swap one replacement level player for another, saving $5M+ in the process. He swapped a league average player for a decent prospect, and replaced the average player with a replacement level player, saving $5M+ and hoping that Callaspo's likely replacement next year - Grant Green - will be roughly equivalent to Callaspo's total production.

 

Don't get me wrong - I loved that trade and look forward to seeing what Green can do. But its easy to undervalue Callaspo, who has been a solid, league average player, which is significantly more valuable than a replacement level player like Chris Nelson, and not so easily replaced.

[Tyler]

Callaspo is way better than Nelson. Another victim of meaningless advanced stats.

[Angelsjunky]

Callaspo is way better than Nelson. Another victim of meaningless advanced stats.

 

Actually, it is those very "meaningless advanced stats" that support your point. Callaspo is having a sub-par year this year, but as I pointed out in the previous post his true talent level is more in the 2-3 WAR range, making him a solidly league average player, whereas Nelson is very much a replacement level player, which is well below average for a regular.

[Tyler]

I dont pay attention to WAR either. I go by what I see + actual stats + my knowledge of the game.

[DMVol]

Not sure any of this matter much because I think Green gets a shot at 3b before the season is over....he is playing 3b at SLC and I suspect they'll bring him up soon and give him a shot, over Nelson...

[Tyler]

Now with that said, I still like the trade because A) we're not in the race it forces us to play Nelson and add more L's to improve our draft position C) saves money to give to Trout and D) we got someone younger who has some upside, although him reaching it, like Nelson, seems to be improbable

[jsnpritchett]

I'm confused.  Where are you getting those stats for Nelson that you use to project his totals in 600 ABs?  

[AngelsFaninGA]

Yeah not sure where that .270 average is coming from. Nelson is hitting about .240 away from Coors in his career. 

[Mile High Dreams]

Was Nelson's WAR calculated at the AAA level? If so what's the co efficient to make this an apples to apples comparison?

[jsnpritchett]

Yeah, I really don't get this at all.  The slash numbers are kinda sorta close to his overall career numbers, but not really.  And if you're using his career stats rebased to 600 ABs, shouldn't you do the same for Callaspo?  Weird article.  

[Mile High Dreams]

Good analysis. I was never a big fan of Callaspo, and to see that Nelson is providing the same WAR is good to know.

But he's not

[ettin]

Yeah, I really don't get this at all.  The slash numbers are kinda sorta close to his overall career numbers, but not really.  And if you're using his career stats rebased to 600 ABs, shouldn't you do the same for Callaspo?  Weird article.  

 

I projected out his numbers based on reduced AAA numbers plus what he did this year. I also projected out rough AVG, OBP, SLG based on reduced minor league numbers (which don't translate as well into the Majors obviously). It just happened to come out to his career average.

 

If you look at his minor league numbers the stats I posted are NOT a stretch by any means. Just happenstance that Nelson's projected numbers were nearly identical to his actual career line to date.

[jsnpritchett]

I projected out his numbers based on reduced AAA numbers plus what he did this year. I also projected out rough AVG, OBP, SLG based on reduced minor league numbers (which don't translate as well into the Majors obviously). It just happened to come out to his career average.

 

If you look at his minor league numbers the stats I posted are NOT a stretch by any means. Just happenstance that Nelson's projected numbers were nearly identical to his actual career line to date.

So you made a bunch of guesses and didn't think it was important to reveal that in the article?  As Junky noted, your analysis led to some flawed conclusions.

[ettin]

Nice article - its good to hear folks using WAR in a clear manner. That said, I have to take issue with a couple things that I think, in the end, mar your argument.

 

First of all, you say that a replacement player is a "generic 'average' player." Generic, yes, average, no. You clarify later on by saying that "a replacement level player as a player who costs no marginal resources to acquire." Meaning, every major league team has a slew of replacement players waiting in the wings - which is essentially what a AAA veteran is. But if this were an "average" player, then no major league teams would have less than average players at any position, which would in turn drive up what "average" means.

 

A replacement player is below average. If replacement level is -1 to +1 WAR, and a good platoon/marginal regular is 1-2, an average starter is around 2-3.

 

Secondly, I don't think the Nelson-Callaspo comparison is fair because 1) it implies that Callaspo's 2013 performance (0.0 fWAR - exactly replacement level) is his true talent level, when in fact he's averaged 3.2 fWAR over the previous two seasons, and 2.4 over his previous four years of being a major league regular, 1.9 if you include 2013. No matter how you look at it, I think it is safe to say that his true talent level is more in the 2-3 fWAR range, which makes him the definition of a league average starter.

 

Chris Nelson, on the other hand, has accumulated a -1.6 fWAR value in 234 major league games, making him nearly the definition of a replacement player (at best) - and thus "a player who costs no marginal resources to acquire."

 

So Dipoto did NOT swap one replacement level player for another, saving $5M+ in the process. He swapped a league average player for a decent prospect, and replaced the average player with a replacement level player, saving $5M+ and hoping that Callaspo's likely replacement next year - Grant Green - will be roughly equivalent to Callaspo's total production.

 

Don't get me wrong - I loved that trade and look forward to seeing what Green can do. But its easy to undervalue Callaspo, who has been a solid, league average player, which is significantly more valuable than a replacement level player like Chris Nelson, and not so easily replaced.

 

In hindsight I probably shouldn't have used the word 'average' simply because it may cause the very confusion that it just did. What I meant simply was that most replacement level players are at or near the '0 WAR' mark give or take. There are certainly outliers to that statement for certain teams but generally it is close to the truth.

 

With regards to this years numbers, yes it is a bit of a cherry pick. However, Callaspo is starting into his decline years and his defense is only going to degrade over time, which is where a lot of his previous years value is originating from. I don't think it is inaccurate to project him based on this years numbers.

 

Nelson is the very marginal resource but I think it is unfair to look at his career averages since he has not really been given a consistent amount of playing time in his short Major League career. In his partial (basically half) 2012 season in Colorado he hit .301/.352/.408 with 9 HR's. Yes that is Coors Field but even accounting for park factors he had a productive 1/2 season. His minor league numbers support that projection. If anything my projected line might be selling him short a bit.

 

So with all due respect AJ I think you are completely undervaluing Nelson to make your counterpoint. Many of you preach about how WAR calculation for defense is inaccurate yet want to use it as the basis of the argument for why Callaspo is good (and that may not be you who does that AJ but others here do). I'm not saying that Alberto didn't have value, he did for a while there but what I am saying is that Nelson is equally good in a different sense and is a bit younger with more speed.

 

The real point I am trying to make is that did we really lose anything in the Callaspo trade? I definitely don't think we did, period.

[ettin]

So you made a bunch of guesses and didn't think it was important to reveal that in the article?  As Junky noted, your analysis led to some flawed conclusions.

That is a bit unfair of you to say pritchett. How does any projection system project numbers? They are made with educated guesses. No projection system is flawless and my "bunch of guesses" were made looking at Chris Nelson's Minor League and Major League numbers, making some adjustments for the Pacific Coast League's propensity for good offense (and Coors Field for that matter) and then making as educated a guess as I could about what those numbers would be.

 

I'd invite you to actually go look at Chris Nelson's Minor League and Major League numbers and decide for yourself if I'm really that off?

[Angelsjunky]

ettin, well I agree with that final point - but not because of Nelson but because of Green. I also agree that, if Callaspo's bat doesn't improve, his overall value is going to plummet as his defense erodes. Its also worth pointing out that his triple-slash numbers have gone down in each of the past two years. He's young enough (30, 31 next year) for that trend to halt or even regress upward, but it isn't a good sign.