Is WAR (Wins Above Replacement) Accurate?

[Jordan Tuwiner]

On Monday I'll be publishing a small study I did comparing teams' actual wins to their projected wins based on WAR (Wins Above Replacement).

Before it goes up how accurate the ON community believes WAR is.

If you're unfamiliar with WAR I'd recommend reading this page first then digging into Dave Cameron's 15 part series on WAR:

Background on WAR - Offense

• Part 1 - Batting
• Part 2 - Fielding
• Part 3 - Positional
• Part 4 - Replacement
• Part 5 - Converting Runs to Wins
• Part 6 - Dollars
• Part 7 - Additional Info.
• Part 8 - Team Context

Background on WAR - Pitching

• Part 1 - Introduction
• Part 2 - FIP
• Part 3 - Replacement
• Part 4 - Run Environments
• Part 5 - Converting Runs to Wins
• Part 6 - Park Adjustments
• Part 7 - Calculations

What are your biggest issues with WAR? When I publish my study, do you expect the conclusion to be that WAR is accurate or not accurate?

[ofahn]

Would someone either give me a link to an explanation of WAR or post it?

I understand the basics but I would like to know more. Thanks.

[Jordan Tuwiner]

Would someone either give me a link to an explanation of WAR or post it?

I understand the basics but I would like to know more. Thanks.

I'd start here: http://www.fangraphs.com/library/index.php/misc/war/

Then check out Dave Cameron's 15 part series on WAR:

Background on WAR - Offense

• Part 1 - Batting
• Part 2 - Fielding
• Part 3 - Positional
• Part 4 - Replacement
• Part 5 - Converting Runs to Wins
• Part 6 - Dollars
• Part 7 - Additional Info.
• Part 8 - Team Context

Background on WAR - Pitching

• Part 1 - Introduction
• Part 2 - FIP
• Part 3 - Replacement
• Part 4 - Run Environments
• Part 5 - Converting Runs to Wins
• Part 6 - Park Adjustments
• Part 7 - Calculations

[BuckMagic]

My only problem with WAR is this: it tells me Alex Gordon was more valuable than Adrian Gonzalez this season. It tells me Howie Kendrick and Brandon Phillips were more valuable than Prince Fielder this year. It tells me JJ hardy was more valuable than Mike Stanton. I could go on and on but does that make sense? and if so how?

[Jordan Tuwiner]

My only problem with WAR is this: it tells me Alex Gordon was more valuable than Adrian Gonzalez this season. It tells me Howie Kendrick and Brandon Phillips were more valuable than Prince Fielder this year. It tells me JJ hardy was more valuable than Mike Stanton. I could go on and on but does that make sense? and if so how?

That's mostly due to the positional adjustment. It's much harder to find a good hitting 2B than it is to find a good hitting 1B, for example.

[Jordan Tuwiner]

My post on WAR will be up at 12:00 AM. Here's some food for thought:

• 18 of the 30 teams were within one standard deviation.
• 28 of the 30 teams were within two standard deviations (9.32). Only two teams -- Boston (11.9) and Kansas City (10.1) -- were more than two standard deviations away.
• WAR standings would give us the same order of standings for two divisions: the AL West and NL West.
• The top nine teams in WAR wins includes all eight playoff teams.
• If Boston had made the playoffs, the top eight teams in total WAR would have been the eight playoff teams.

[BuckMagic]

My only problem with WAR is this: it tells me Alex Gordon was more valuable than Adrian Gonzalez this season. It tells me Howie Kendrick and Brandon Phillips were more valuable than Prince Fielder this year. It tells me JJ hardy was more valuable than Mike Stanton. I could go on and on but does that make sense? and if so how?

That's mostly due to the positional adjustment. It's much harder to find a good hitting 2B than it is to find a good hitting 1B, for example.

That makes sense. So it's not as much value, but helps adjust to the fact that Kendrick is a better 2B compared to the rest of the 2Bs in the league?

[BuckMagic]

My post on WAR will be up at 12:00 AM. Here's some food for thought:

• 18 of the 30 teams were within one standard deviation.
• 28 of the 30 teams were within two standard deviations (9.32). Only two teams -- Boston (11.9) and Kansas City (10.1) -- were more than two standard deviations away.
• WAR standings would give us the same order of standings for two divisions: the AL West and NL West.
• The top nine teams in WAR wins includes all eight playoff teams.
• If Boston had made the playoffs, the top eight teams in total WAR would have been the eight playoff teams.

What causes teams like Boston and KC to differentiate so much?

[Jordan Tuwiner]

My only problem with WAR is this: it tells me Alex Gordon was more valuable than Adrian Gonzalez this season. It tells me Howie Kendrick and Brandon Phillips were more valuable than Prince Fielder this year. It tells me JJ hardy was more valuable than Mike Stanton. I could go on and on but does that make sense? and if so how?

That's mostly due to the positional adjustment. It's much harder to find a good hitting 2B than it is to find a good hitting 1B, for example.

That makes sense. So it's not as much value, but helps adjust to the fact that Kendrick is a better 2B compared to the rest of the 2Bs in the league?

That's exactly right.

[Jordan Tuwiner]

My post on WAR will be up at 12:00 AM. Here's some food for thought:

• 18 of the 30 teams were within one standard deviation.
• 28 of the 30 teams were within two standard deviations (9.32). Only two teams -- Boston (11.9) and Kansas City (10.1) -- were more than two standard deviations away.
• WAR standings would give us the same order of standings for two divisions: the AL West and NL West.
• The top nine teams in WAR wins includes all eight playoff teams.
• If Boston had made the playoffs, the top eight teams in total WAR would have been the eight playoff teams.

What causes teams like Boston and KC to differentiate so much?

Hard to say. I'll dig into the stats and get back to you on that tomorrow.

[Jordan Tuwiner]

Here's the article: http://orioles-nation.com/2011/10/03/is ... rate-2011/

[A_K]

That was a great analysis, Jordan.

The thing that most strikes me when traditionalists discuss statistics like WAR, FIP, UZR and wOBA is their determination to isolate the newer statistics' flaws, all while defending a group of statistics that, if the defenders are even remotely honest, are full of a set of flaws all their own. Sure, there are weaknesses in UZR, and therefore WAR since it uses UZR, and maybe SABR defenders aren't altogether honest about those weaknesses. But are those weaknesses really more egregious than those of fielding percentage? Are the issues with HR/FB% and the way they relate to xFIP really more troubling than those surrounding a pitcher's W-L record?

I think analyses like yours show that statistics are, almost invariably, getting us closer and closer to a complete understanding of the game. There will always be exceptions and imperfections, but the direction of our understanding is clearly right.

[Jordan Tuwiner]

That was a great analysis, Jordan.

The thing that most strikes me when traditionalists discuss statistics like WAR, FIP, UZR and wOBA is their determination to isolate the newer statistics' flaws, all while defending a group of statistics that, if the defenders are even remotely honest, are full of a set of flaws all their own. Sure, there are weaknesses in UZR, and therefore WAR since it uses UZR, and maybe SABR defenders aren't altogether honest about those weaknesses. But are those weaknesses really more egregious than those of fielding percentage? Are the issues with HR/FB% and the way they relate to xFIP really more troubling than those surrounding a pitcher's W-L record?

I think analyses like yours show that statistics are, almost invariably, getting us closer and closer to a complete understanding of the game. There will always be exceptions and imperfections, but the direction of our understanding is clearly right.

Thanks, A_K! It does amaze me when I have people telling me that wOBA, UZR, BABIP, etc. are useless. As you said they are quick to point out flaws but fail to see the flaws in the statistics like SLG, AVG, etc. At least I haven't seen anyone argue with me when saying RBI is useless.

[BuckMagic]

Does anyone know if organizations use WAR, or something similar?

I recently read something about use of statistics within MLB front offices and it appears 10 or 12 use a ton of statistics when evaluating players. Wondering if WAR is among them.

[A_K]

Does anyone know if organizations use WAR, or something similar?

I recently read something about use of statistics within MLB front offices and it appears 10 or 12 use a ton of statistics when evaluating players. Wondering if WAR is among them.

My understanding is that virtually every team uses WAR as a primary statistical indicator.

The figure I think you're mentioning is the 10-12 teams that are considerably more statistically inclined than other teams, having developed their own proprietary statistics and making decisions based upon those moreso than would be traditionally expected.