Read, watch TV, listen to the radio or just talk baseball around the water cooler, and someone will inevitably start mentioning things like OPS, WAR, VORP or any of a host of other acronyms that seemingly bring value to the discussion. But do they really?
It’s a question that brings about wild enthusiasm on both sides of the issue. The diehard “sabers” tend to act as though advanced metrics form the beginning, middle and end to any discussion regarding a player’s talent, and only those of us lacking intellect or the capacity for scientific thought would deny this. The “traditionalists” think the “sabers” are a bunch of basement-dwelling dweebs who simply can’t appreciate the intangible aspects of the game. The standard line of BA, RBI, etc... is more than enough information for them, and the rest boils down to visual observation, baseball IQ and just plain gut instinct.
So, who’s right?
I don’t know. But wouldn’t it be great if we could put them to a simple test?
I thought it might be fun to look at one of the metrics designed to more directly represent a player’s contribution to his team’s success and see if it actually holds up.
At its basic core, WAR attempts to quantify the ability of a player to add wins to his team’s record. Specifically, it is measured as the number of wins a player can create for his team vs. what the average replacement player can. It’s a rather complicated equation that factors in a number of other metrics and weights them accordingly.
To keep things as constant as possible, I will only use data from www.FanGraphs.com. I also wanted to avoid using a team that performed well outside of expectations (traditional or metric based), as that would certainly lead to accusations of bias from one side or the other. For example, the Red Sox performed significantly below expectations no matter which camp you belong to.
However, the Toronto Blue Jays were almost universally picked to finish mid/bottom in the AL East. They finished fourth. So, let’s start with them.
According to FanGraphs, the total offensive WAR for the 2011 Blue Jays roster was 21.3. On the pitching side, it was 12.5. After adding in the 2011 expected replacement level wins of 42.2 (as per FanGraphs), we get 76 wins (42.2+21.3+12.5). In fact, the Blue Jays finished 81-81. Their actual performance was five wins above what the math would indicate.
Now, let’s look at the Houston Astros, another team picked near the bottom of their respective division. Their offensive WAR was 17.3. Pitching WAR, 4.1. This totals 63.6 with the replacement line added. They ended up winning 56 games, so their actual performance was eight wins below the metric.
How about the Cleveland Indians? Offensive WAR: 14.6. Pitching WAR: 15.5. Here we come up with 72 wins. They won 80. This time, eight wins above WAR.
I’m not sure what to make of this quite yet, but it’s not looking good for the “sabers." This is where I started scratching my head and decided to take on the dreaded task of crunching the numbers for all teams.
Here’s what I came up with:
A full 21 teams’ actual records were within five wins of their total WAR calculation. Thirteen of them were within three wins, and five were pretty much dead-on. Only two teams were more than eight wins off: the aforementioned Red Sox and the Kansas City Royals.
I decided to look at it another way. What if I totaled each team’s WAR values and then sorted them in order from highest to lowest within their division? Would the metric match up as effectively here?
This was really eye-opening. The team WAR calculation predicted positions within each division remarkably well. Only the Marlins, Royals, Red Sox and Cubs were out of position. In the case of the Red Sox and Cubs, the difference in position was determined by a variation of only one win. So WAR was able to compute a team’s divisional standing to within one or fewer wins in 28 out of 30 cases.
It’s important to remember here that since we only looked at the total WAR for each team, it doesn’t say a whole lot about its value at an individual player level. It also doesn’t answer the question about whether or not it is a good predictor of future performance. But then, no traditional statistic is a foolproof method for predicting the future. All stats are merely a measure of performance within certain parameters.
I also want to be clear that as I understand it, the concepts of “moneyball” and “sabermetrics” are not one and the same. Sabermetrics simply attempt to objectively measure a player’s effectiveness. Moneyball attempts to use sabermetrics to quantify a player’s monetary value within a set ideology. Perhaps confusion between the two contributes to the disagreement between “sabers” and “traditionalists." Perhaps it’s just representative of one group’s willingness to apply mathematical analysis to a subjective topic, and the other group’s opposition to having their beloved game reduced to a bunch of numbers.
I honestly didn’t know what to expect as I worked out these rather simple tests. One year is a very small sampling to work with, but I clearly failed to disprove its credibility.
Now, I don’t believe that it’s possible for any metric to account for every single aspect of the game and accurately quantify a player’s value to his team. Derek Jeter’s ridiculously low UZR is a prime example of how things can get thrown off when intangibles are missed. But I can’t deny how closely WAR mirrored actual performance at the team level.
So, my conclusion is that the various metrics can’t form the entire determination of how good a particular player is. But like most other stats, it can and should serve as a valid component to the discussion.
However, I still doubt that either side will give in one bit. And, in retrospect, it’s probably a good thing we didn’t replace the water cooler with a keg like I suggested at the last office party.