FanPost

Alphabet Soup: FIP, OBA, UZR, and the rest of the keyboard

Just like any other topic, there are going to be debates in baseball. Who was better, Babe Ruth or Barry Bonds? Should Pete Rose be allowed into the hall of fame? Does instant replay help or hurt baseball? Who shot first, Han or Greedo? (oops, wrong forum) One of the things I have recently noticed here on AN is that there are a lot of new names and contributors. And with new faces talking baseball comes new debates.

Among the most commonly (and contentiously) debated topics here on AN seems to be what statistics we should use when evaluating a player. You have the traditional statistics, like AVG, RBI’s and ERA. Then you have the modern statistics that look like a jumble of letters randomly thrown together. Just about every single person on this forum understands traditional statistics. They are what we all grew up with. I remember memorizing the back of Rickey Henderson’s baseball card when I was 8 years old. But not everyone here knows much, if anything, about some of these newer stats. At first glance they don’t look too appealing. After all, a .405 wOBA doesn’t sound nearly as sexy as 40 HR’s.

What I am going to attempt to do here is simply explain some of the most commonly used modern stats on AN without using any math. I’m going to try to keep the explanations short and put them in actual, easy to read English. The goal of this is not to try and convince anyone that these statistics are superior to traditional statistics. Instead I want to give people an idea of what these statistics attempt to measure and why some people may choose to use these stats over traditional stats. These modern stats may not be for everybody, and that’s totally fine! But I feel that if we can understand one another then this will lead to less arguing in game threads and more enjoyment. I hope what follows lives up to that expectation.

wOBA

wOBA stands for "Weighted On Base Average." Put simply, what this stat does is apply what is known as linear weights to an individual player’s hitting line. As we all know, not all hits are created equal. A double is a better outcome than a single, a homerun is a better outcome than a double, all three of those are generally better than a walk, and any of those are better than an out (and an out is what we call Kurt Suzuki). One of the big problems with AVG or OBP is that they treat each of those outcomes as if they were the same. Linear weights does not.

What linear weights does do is it assigns a value to each possible outcome in an at bat. It does this by looking at how many times the event in question (such as a HR) occurred and then averages the number of runs that were scored during and after that event in the same inning. For instance, between 1999 and 2002 there were 21,026 HR’s hit. From the moment that the HR was hit to the end of the inning, 40,838 runs scored. That gives you an average of 1.942 runs scored per HR. Now we have the average value of a HR. Linear weights does this sort of calculation for each possible outcome in an at bat. Keep in mind that these values change every year, as a HR may be worth more in some years than others (think "Deadball Era" verses "Steroid Era").

So wOBA essentially applies these linear weights to what a batter has actually done in the batters box. Each walk will be given its appropriate value (in 2011 it was 0.69 runs) as will each HR (in 2011 a HR was worth 2.08 runs). If you want, you can sort of think about wOBA as simply being OBP that gives each result a different value. In an effort to keep the math out of this post, you can find the formula for wOBA at Fangraphs here. Danmerquery also did a nice piece on wOBA last year that goes into more detail. It can be found here. One important thing to note is that this stat looks very similar to OBP. A .320 wOBA is about average, a .290 wOBA is just terrible, and a .400 wOBA is amazing. Applying 2011 numbers, .330 is Brett Gardner, .285 is Vernon Wells, and .406 is Adrian Gonzalez.

FIP

FIP stands for "Fielding Independent Pitching," and is what we call an ERA estimator. That is, FIP has a very different purpose from ERA. ERA tells you what actually happened on the field and the actual results of a pitching performance. At the same time, it is subject to all kinds of random chance events, such as fielding miscues (like a botched double play or mis-reading a fly ball), bloop singles, running mistakes, screaming line drives hit directly at a fielder, etc. And while these events are all part of the game, they are not necessarily the result of what the pitcher did.

FIP attempts to remove the defensive component of ERA and measure only those events that exclude a teams’ defense. That is, it measures strikeouts, walks, and homeruns. Once again, FIP isn’t designed to explain what has already happened. It is used to estimate future performance. FIP basically says that if a pitcher continues doing what he has been doing, his ERA will look like X. And the reason so many people like run estimators such as FIP is because every study done has shown that FIP does a far better job of estimating the ERA of a pitcher for the next year than ERA itself does. If you want to know what a pitcher’s ERA will look like in the future, looking at ERA is going to be less accurate than looking at FIP.

A couple little side notes. FIP doesn’t completely ignore all balls in play. It’s complicated. If you want to know how that is actually included in the analysis, Dan explains it well here. You can also do some good reading on FIP here if you are interested.

UZR

UZR stands for "Ultimate Zone Rating" and is used as a defensive metric in place of fielding percentage. What fielding percentage does is it simply calculates the total put outs and assists with respect to the total chances a fielder has. What it doesn’t do is account for a fielder’s range (among other things). If Jemile Weeks can get to a greater number of balls hit in his area than Dan Uggla, he can take more hits away from the hitters. But he also has more opportunities to make errors. A ball that Weeks dives for and gets his glove on while not making the play (allowing the runner to be safe at first) gets counted as an error. Yet Uggla would not have the ability to get his glove on the ball in the first place and the runner would be safe at first anyway. Uggla would not be penalized for that runner being on first, but Weeks would be.

So that’s one of the problems many see with fielding percentage. UZR attempts to fix that by including range along with a bunch of other factors. To keep this short and sweet, UZR looks at components such as "Outfield Arm Runs" (ability to prevent runners from advancing based on an OF’ers arm), "Double-Play Runs" (how well a fielder can turn the double play), "Range Runs" (how well the fielder gets to balls in their area), and "Error Runs" (this is sort of like fielding percentage).

UZR is far from perfect. Measuring defense is by far the hardest thing to measure in baseball. But I think you can see why a lot of people find UZR to be more attractive than fielding percentage. For a more in depth look at this, visit Dan’s article here or you could read a very detailed primer here. Also note that this is a counting stat, much in the same way that HR's and RBI's are counting stats.

WAR

I’m going to keep this nice and simple. If you want to do further reading, go here or here.

Put simply, WAR combines the above stats and attempts to put an overall value on the player you are looking at. For hitters, WAR uses wOBA (but expressed in a different manner), UZR, and a base running component (BsR). For pitchers, WAR uses FIP. The result will give you a positive or a negative number. Positive numbers are good, negatives are bad.

What WAR stands for is "Wins Above Replacement." What does this mean? I don’t think I could say it any better than they do over at Fangraphs, so:

WAR basically looks at a player and asks the question, "If this player got injured and their team had to replace them with a minor leaguer or someone from their bench, how much value would the team be losing?" This value is expressed in a wins format, so we could say that Player X is worth +6.3 wins to their team while Player Y is only worth +3.5 wins.

To understand what a good WAR is, a 0 WAR means that a player is essentially a replacement player, or a scrub. 2 wins is actually a lot (a team consisting of nothing but 0 WAR players would win between 40 to 50 games in a year). A 2-3 WAR player is your average MLB player at that position. Anything above 3 WAR is good, and 6 WAR or above is MVP level.

To put that last part into perspective, in Giambi’s last year with the A’s in 2001 he was considered to be a 9.3 WAR player. That means if he were on a team with nothing but 0 WAR scrubs (guys like Matt Murton, Eric Sogard, etc.) he would have added 9 wins to that rather talent-less team. His presence alone would have turned a 45 win team into a 54 win team. That is HUGE. WAR gives you a nice, easy, simple to use number for evaluating a player.

One last small note. There are two different calculations for WAR. Fangraphs (fWAR) does it somewhat different from Baseball-Reference (bWAR). They both comprise of the same idea, but calculate the number differently. So far my links have been to Fangraphs. If you would like to see the way bWAR is created check it out.

Notable Mentions

There are a few other stats that you will see thrown around from time to time. Here are the more common ones and a quick explanation:

· wRC+ - Stands for "Weighted Runs Created Plus," and is an offensive stat similar to wOBA. It is park and league adjusted. 100 wRC+ is always league average, with every point above or below 100 representing a percentage point above or below league average. 120 wRC+ means a player created 20% more runs than the league average.

· SIERA – Another run estimator like FIP. It is also scaled to look like ERA.

· BABIP – This stands for "Batting Average for Balls in Play." Basically, it gives you the batting average for every at bat that does not end in a K, BB, or a HR. League average is generally around .300. Hitters can control their BABIP better than pitchers can, as a pitcher’s BABIP (the good pitchers as well as the bad ones) are usually right around league average.

I hope this has been helpful. This is also my first fanpost, so I hope everything shows up the way I intend. If anyone has any questions, I'd be more than happy to answer them. And if any of the "stat guys" notice a mistake in the way I presented the info, please let me know. I tried to keep this post simple, but I don't want to be misleading as well. Thanks :-)