How are we supposed to interpret all the new stats in baseball, and which ones have proven the best?
Let's say you really want to understand all the various new metrics that have been applied to baseball in past years (yes, yes, I know, plenty of people are just fine sticking to batting average and RBIs), but all the plusses and weighted numbers start to make your head spin.
They make mine spin, too. So I tried to make some sense of it with Curt Hogg, our Brewers beat reporter.
JR: Let's start as broadly as possible. If I'm trying to evaluate a player, especially a position player, what stats should I be using, and why?
CURT: The short answer? WAR is the best metric. While it's not without its flaws, it is the best catch-all statistic in baseball. Different sites (FanGraphs, Baseball Reference, Baseball Prospectus, etc.) all use their own specific formula to calculate a single number that measures, well, how many wins above a "replacement-level" player a specific player was. It takes into account all aspects of the game: hitting, fielding and base running.
JR: I remember last year, Jace Peterson was the highest-regarded Brewers position player by that metric (easily) midway through 2022 and Rowdy Tellez barely above 0, and I'm questioning how that can be possible. Is my eye test really that far off? Then, throw in the fact that there are multiple versions of WAR. Peterson's WAR tallies according to Baseball Reference and Fangraphs were very different. How can you trust a metric that yields different numbers depending on whom you ask?
CURT: Not only does WAR factor in things like defense and baserunning (both of which favored Peterson over Tellez) but defensive position factors in, too. Positions like center field, shortstop and third base are considered more valuable defensively because they are harder to play. Peterson had not only been excellent at third but added value with his versatility, as well. It also doesn't hurt Peterson that he had been just as good as Tellez on offense.
The multiplicity of types of WAR can be seen as a good thing; it allows for a handful of different calculations of what is most valuable and how to best showcase that. But it's also a reminder that WAR is not exact. Typically, especially for position players, it's the different defensive metrics used by different sites that lead to most discrepancies. That's part of the difficulty of evaluation defense. Each major site gives an overview of their process.
JR: If we're taking defense out of the equation and comparing just the bats, what are the stats you're looking at?
CURT: I'll always say I think it's best to look at multiple different stats or metrics, but the one catch-all stat that I use most is wRC+ (weighted runs created plus). It measures a player's overall hitting production and adjusts it for the league and home ballpark factors so that a player in hitter-friendly Colorado can be weighed on the same scale as one in a pitcher-friendly park like Seattle. Then, it's set up so 100 is average across MLB and anything above is better than league average. A hitter with a 120 wRC+, then, is roughly 20 percent better than an average hitter.
The "underlying" stats that help paint the picture of who a hitter is include walk and strikeout percentage (BB% and K%); average exit velocity and hard-hit rate (as found on Baseball Savant); chase rate and swinging strike rate, which help show a player's discipline; and batting average on balls in play (BABIP) to see if a player may be a bit lucky or unlucky.
More: Can we better predict future hitting performance for MLB using a new stat? We tried to find out.
JR: Is there a layman's way to explain how a stat can be weighted for different ballparks? Does that essentially mean homers at Coors only count for, like, three bases? And homers in Seattle count for five?
CURT: Here's a great site that lays out how different types of hits are weighed at different parks. To simplify what can be a somewhat overwhelming layout, just look at the "park factor" column and then everything from "1B" over to "SO." It's on a similar scale as wRC+ where 100 is average and any number above that means the ballpark is more conducive for that type of offensive outcome than average.
You'll see Cincinnati is far more conducive to homers than any other park. While a Reds player will still get credit for the raw homers he hits, having these park factors allows us to equate that player's production with someone in Detroit or Seattle.
JR: Now I need to know the difference between wRC+ and OPS+ (on-base plus slugging percentage). My understanding is that they're basically read the same way — OPS+ is also scaled so an average player has 100 and each point above that is a percentage point above average. So what does one tell me that the other doesn't? Also, forgive my math stupidity, but why scale it to 100 in the first place? Why not just have the average player be zero?
CURT: The difference is the initial metric that is being weighed. OPS+, as you could assume, takes a player's OPS and weighs it against the league average. wRC+ is an improved version of famous sabermatrician Bill James' "runs created" stat and uses a player's weighted on-base average (wOBA). Both OPS and wOBA are generally good measures of offense and correlate well with runs, but wOBA is a bit more complex and is more closely related to scoring runs. Whereas OPS simply combines on-base and slugging percentage, wOBA weighs each event specifically; in other words, it assigns values to each event to give a truer sense of how much a single, double, triple, etc. is worth. That gives it an advantage over OPS, which overvalues singles compared with walks, and doubles compared with singles, for instance.
As for why they are weighed at 100 ... I have no clue. Great question!
JR: Whoa, we took a hard turn from "wRC+" to "wOBA" and you lost me.
So wOBA is basically looking at the things that happen in the batter's box (on-base percentage, slugging), weighting it against parks and providing a value, right? It's a fancier OPS? Which, seems counterintuitive, because I would presume "on-base average" is an equivalent to "on-base percentage" but I guess not.
And wRC+ is throwing in base running, too?
Does either stat factor in RBIs? Or clutch situations?
CURT: Well, wRC+ is based in wOBA so they go hand in hand. wOBA itself isn't park or league-adjusted, which is why wRC+ is a slightly better measure. And because wRC+ is on a scale where 100 is always average, it's easier to eyeball and get an idea of how that hitter has done.
wOBA is calculated to be a similar number to OBP (league average is somewhere around .320 usually) but is obviously meant to be different than what OBP measures.
No, wRC+ doesn't measure base running. That (as well as defense) are factored into WAR. Neither wOBA nor wRC+ factors in game situations like clutch or RBIs, which allows you to evaluate players apples-to-apples on a much more context-free basis. There are certain stats, like Win Probability Added that show you how impactful a player's hits were within the context of the games they played.
More: Former Brewers players (and other familiar names) on opening-day MLB rosters
JR: WPA is sort of a weird measure, I've always found, because it seems like it favors the guys who've had more "clutch" opportunities, but I suppose that fits the concept of "clutch" anyway. Does every stat need to begin with a W?
Anyway, my eyes are starting to glaze over. What about pitchers? I know there's ERA (earned-run average) and WHIP (walks and hits per innings pitched), which are pretty easy to understand, and things like K% (the percentage of batters you strike out, also easy to understand) that shows how dominant a pitcher is. Then there's FIP (fielding independent pitching), which I've always interpreted as "What your ERA should be," since it tries to judge a pitcher independently of his defense's contributions. How does that work exactly?
CURT: FIP essentially takes on the premise that only three outcomes truly are controlled by the pitcher: strikeouts, walks and homers. The idea behind it is that the pitcher has very little control over balls in play. The stat is calculated as to be similar to ERA where, for example, 3.00 is good and 6.00 is bad. That makes it more familiar to use.
That said, there are some flaws with FIP. We now know that a pitcher does exert some control over how hard the ball is hit and whether it’s in the air or on the ground. There’s a reason Corbin Burnes doesn’t allow as many hard hits as a rookie just up from the minors. FIP is generally a decent predictor of future ERA, but tread with caution.
JR: Break this down for me just a little bit more. You're saying FIP only cares if the pitcher strikes out a hitter, issues a walk or gives up a homer. So, let's say a guy retires 27 guys for a complete game with no walks but it's all groundouts and flyouts. Does FIP evaluate that game as something like a 3.00 because it figures he should have given up some runs and would be roughly expected to give up one hit per inning? Or does it give him no value at all because it only cares if he gets one of the three true outcomes? I'm just trying to get a handle on how FIP is taking those three values and coming up with an ERA equivalent.
CURT: A pitcher who throws a perfect game but with no strikeouts would have a FIP of 3.11 for that game. In other words, they would be expected to allow an average of 3.11 runs in a game where they went nine innings and every hitter put the ball in play.
The pitcher would still get credit for that performance (a 3.11 FIP is good!) but the stat would also tell you that, because they had a 3-run difference between their ERA and FIP from the game, they were a bit lucky.
JR: If I look back in time through the lens of these stats, there are some surprises.
Jim Gantner only had one above-average season out of 17 according to OPS+ and wRC+, though he did have a quality 4.3 WAR season in 1983 ... but a career average of 2.0 per season (basically like a Tyrone Taylor last year).
The season tied for the fourth-best in franchise history by wRC+? Sixto Lezcano in 1979! (Slightly) better than any Robin Yount season and tied with Paul Molitor's best!
The best FIP seasons aren't surprises. Devin Williams in 2020, Corbin Burnes in 2021, Josh Hader in 2021 ... Jim Henderson in 30 innings in 2012? Granted, it wasn't a big sample, but we did not appreciate how good he could be, apparently.
The best WAR seasons? Robin Yount in 1982, Ryan Braun in 2011, Carlos Gómez in 2013 and ... Tommy Harper in 1970! Dang. Another player we don't appreciate enough.
Christian Yelich had a franchise-record 174 wRC+ in 2019. Ryan Braun's 2011 (171) is the only thing close, and Yelich had 167 in 2018. The 30th-best season in Brewers history is 140, and Yelich was more than 30% better than that. Whatever high esteem we hold those first two Yelich years with Milwaukee, we should hold them higher. And these stats are weighted so they can be compared across eras, right?
CURT: They're weighted for the specific run environment of the season in which they happened. So, for instance, we can compare hitters' wRC+ from the height of the steroid era to before the mound was lowered.
JR: Last but not least: Are we nerds?
Curt: Yes.
Our subscribers make this reporting possible. Please consider supporting local journalism by subscribing to the Journal Sentinel at jsonline.com/deal.
DOWNLOAD THE APP: Get the latest news, sports and more
This article originally appeared on Milwaukee Journal Sentinel: How do we interpret the new stats in baseball? Which ones are best?
