Over the course of my own lifetime, I’ve watched hundreds of games and even though I’m into stats, the eye test is often my first parameter before I even begin to analyze a player. There are many players whom you don’t need numbers to tell you who are pretty darn good: Mike F’ing Trout, for one. Kole Calhoun. Garrett Richards. Cam Bedrosian circa 2016-present. Likewise, there are players whom you don’t need numbers to tell you who aren’t: 2013 Blanton, 2015 Joyce, for example.
In these cases, traditional stats like AVG, OPS, and ERA would confirm what the eye test already tells you. But what if the players were a lot closer in caliber? It would be preferable to have one all-encompassing stat that gives you a rough holistic estimate of a player’s value, wouldn’t it?
That’s where WAR comes in.
The replacement in Wins Above Replacement refers to a standard AAAA guy. In other words, if your starter got hurt and you had to sign a free agent that were to step into a starting role midseason, who would it be?
Because each ballpark’s different dimensions, it is easier to hit for power in the AL East parks (Fenway, Yankee Stadium, Camden Yards) as opposed to AL West parks (the big A, Oakland’s ginormous grounds, Safeco, Minute Maid Park). In addition, because the American League has a DH and the National League doesn’t, run scoring environments are lower in the NL. As a result, WAR is both park-adjusted and league-adjusted so you can compare any one player’s WAR to another, regardless of stadium or league.
Here are Fangraphs’ general guidelines of WAR, and below is the chart of where position players fall.
Is it even a surprise that Mike Trout breaks this scale? [WAR accumulates over the season just like walks or strikeouts, but it can go up or down with each game, by the way]
Since being created, Wins Above Replacement, or WAR, has been pointed to as the pinnacle of sabermetrics, but it’s crucial to keep in mind that there is a margin of error with every statistic. If position player X’s WAR is 2.2 and that of position player Y is 2.0, is Player X more valuable? In other words, Player X is 10% more valuable than average but is he definitively better? Not necessarily. Depending on who you talk to, each person has their own opinion of what this margin of error should be. I say it’s 25% of the 2 WAR average at the absolute highest, or 0.5 wins. (make sure you don’t make margin of error so big that it’s meaningless, there’s a reason polls/surveys have a margin of error usually of 10% or less).
Have a Mike Trout GIF before we continue.
Point being, WAR is an excellent way of ballparking a player’s value without looking at anything else, but it’s not the be-all end-all like some traditionalists or statheads proclaim and it shouldn’t be treated as such. We’ll get deeper into WAR as we trot along into this series, so stay tuned!
Lastly, there are two main ways of viewing WAR. One is via Fangraphs, called fWAR and the other is through Baseball-Reference, or bWAR (aka rWAR). The difference between the two is in the way that defensive calculations are made, as well as that Fangraphs uses a FIP-based WAR for pitchers (more on this later in the series).
There is also Baseball Prospectus’s WARP which takes pitch framing into account — it is mainly used used for catchers solely for this reason. Statcast is also supposedly coming out with WAR soon: it will be interesting to see how that unfolds.
This might seem overwhelming and hopefully I’ve done a good job of explaining this in layman’s terms, but if there’s one thing you should take away, it’s this.
It would be preferable to have one all-encompassing stat that gives you a rough holistic estimate of a player’s value, wouldn’t it?
That’s where WAR comes in.
WAR is an excellent way of ballparking a player’s value without looking at anything else, but it’s not the be-all end-all like some traditionalists or statheads proclaim and it shouldn’t be treated as such.