Will James

[ShotgunDawg]

While I really like looking at stats and using them as a tool and support, the article written by Dave Cameron this morning really exposes the weakness of WAR.

WAR is a great great stat for players that have a significant track record in MLB, but fails to offer any credit to players who are still developing and are going to improve.

In this article, Cameron has a difficult time figuring out how the Fielder/Kinsler trade helps the the Rangers. However, the reason he struggles to see this is because Steamer only projects a +1.6 WAR for a young and improving Profar and projects a +3.3 WAR for a maxed out Ian Kinsler that has a track record.

Problem is, anyone that has seen Profar play knows that this WAR is complete crap because he hasn't even been an everyday player yet, is still developing, and most any scout would tell you that that he could be at least the player Kinsler is and possibly more.

Not calling you out, because I really enjoy your stuff, but this article perfectly illustrates why there must be a balance between scouting and sabermetrics in any front office and for any decision.

http://www.fangraphs.com/blogs/rangers-move-pieces-around-outcome-to-be-determined/

[Will James]

Well if you focus on WAR then yeah.... I don't though. I like per/plate appearance stats with a healthy sample size. WAR is cumulative obviously.

[bully99]

No Comment

[engie]

Regardless -- even with all the advancements in SABR, you'll never be able to statistically account for potential, particularly unrealized potential. That's not a shortcoming of the stat itself -- just a limitation in it's feasible usage in personnel decisions.

[dawgs]

wait, so young players can be expected to improve, prime aged players can be expected to roughly maintain, and older players can be expected to decline at some point? no shit. amazing that this guy figured this out.

hey guyz, i bet bryce harper's WAR will improve next year. shocking i know because of what his WAR was this season when he was 21 and banged up, but at some point he's gonna explode for a huge season and WAR doesn't predict that. you have to use your brain and look at his age and project improvement yourself.

no one uses WAR exclusively and without considering the ages of players to project whether they are soon to be on the rise or decline.

[ShotgunDawg]

Exactly, just found it interesting that Fangraphs main writer did exactly that.

[smootness]

wait, so young players can be expected to improve, prime aged players can be expected to roughly maintain, and older players can be expected to decline at some point? no shit. amazing that this guy figured this out.

hey guyz, i bet bryce harper's WAR will improve next year. shocking i know because of what his WAR was this season when he was 21 and banged up, but at some point he's gonna explode for a huge season and WAR doesn't predict that. you have to use your brain and look at his age and project improvement yourself.

no one uses WAR exclusively and without considering the ages of players to project whether they are soon to be on the rise or decline.

This. His issue isn't with WAR, it's with projections of any kind. But then, he also seemingly wants projection.

If you're making a decision solely on someone's projected WAR for a young player, then the issue is with your decision-making process, not with WAR.

[dawgs]

i don't. it's common sense that a top prospect is likely to see his WAR rise over his first few seasons, whereas a 30+ year old guy is likely to see his WAR decline over the next few seasons. anyone that looks at data without considering the variables is a moron. if you think fangraphs just looks at faceless numbers without considering things like age and projections when they discuss player values moving forward then you are wrong. it's why bryce harper is rated higher on a trade value scale than cabrera.

[dawgs]

This. His issue isn't with WAR, it's with projections of any kind. But then, he also seemingly wants projection.

If you're making a decision solely on someone's projected WAR for a young player, then the issue is with your decision-making process, not with WAR.

an algorithm isn't going to really account for a guy's raw talent gaining experience and his expanded role moving forward. i didn't realize people didn't already understand that.

[Todd4State]

dawgs is knocking the ball out of the park on this.

I like WAR. I think it's good for evaluating where a player currently is and their current value. I also like it because someone along the way started to think- "well, why don't we just bunch up everything together and then come up with a rating."

Like all stats there is no perfect stat and they all have weaknesses. One of the main ones with WAR is there are actually several different versions of WAR. Baseball-Reference has one version, and Fangraphs uses another, and then there are others. That's why when you see me refer to WAR, I mention which version I am using usually.

The other issue with it right now is it is an evolving stat. And by that, I mean they are going to tweak it some as they get better defensive stats and data as time goes on. And they may do some things with the offensive stats as well that factor into it.

When I have seen people use WAR as far as projecting a player, most people assume and hope that they are going to get about the same production or slightly less but it's certainly not foolproof. And it's almost impossible to do that with a young player that has one or two years in the league- other than you can maybe assume it will go up like dawgs is saying.

But people really don't do that when projecting players. They are basically looking at stats and how a player is performing, and also they rely a lot on the minor league coaching staff as well. What is the player's work ethic like? OK, he struggled at this point in time, and how did he handle it? Did he listen to the coaches to correct his issues, or did he tune them out? Things like that, you can't capture with a stat. That's why a guy with worse stats is maybe sometimes more highly thought of than a guy with better stats at the minor league level.