The Year Sabermetrics Win

http://www.thechiefly.com/features/the-year-sabermetrics-win/

3132014sabermetrics

The annual Sloan Analytics Conference took place at the Massachusetts Institute of Technology earlier this month, as advanced statisticians across every sport gathered in Boston to share their respective strategies for finding new ways to evaluate players. The existence of this conference primarily disproves one of the basic misconceptions about advanced statistics: that they purport to know all the answers. If they did, sabermetricians wouldn’t seek help from one another each year.

“Advanced statistics” are only “advanced” because many of the ones we have used for decades are so laughably flawed. Sabermetrics statistics are more accurate and predictive than traditional stats because they incorporate external factors into every action on the field, while many metrics like RBI, runs, and pitcher wins remain devoid of all context. Baseball is a better example of this conflict between old and new than any other sport because of its extreme reliance on stats for over a century now, as fans everywhere know what these numbers mean: 755; 56; 73; .407.

We have centered so many of our arguments around statistics which presume to say a lot, but in reality say very little, and it’s time for that to change. It’s 2014 and the world is very different. Now we can print guns from home, find people to have sex with on our phones while waiting for the bus, and even Skynet actually kind of exists. Life can answer almost any general inquiry, so why shouldn’t the metrics we use to evaluate our pastime evolve with the times?

Spring training is here. We’re all in the best shape of our lives and primed for a career year. Let’s make this the season that we finally move forward and abandon many of these mostly meaningless statistics in favor of better, more descriptive “advanced” statistics.

Note: Every advanced stat listed here could be irrelevant in 5 years with the advent of MLBAM’s new tracking system launching this summer, which will basically measure every movement of everything on a baseball field.

Eliminate: RBI

Baseball got it right the first time when it refused to recognize this stat prior to 1920. Many will tell you that this is one of the best statistics to measure a hitter’s skill because, after all, it shows the amount of Runs a player Bats In.

Yes, technically RBI assigns someone at the plate credit for runs that other players score, but it does not mean that one’s number of RBI reflects their relative skill of batting runs in. In sports, sometimes shit happens, and people get credit for an outcome they had little to do with. We can’t just blindly say “guy A scored with guy B at the plate so guy B is a run producer!” Correlation does not imply causation.

The central flaw of RBI is that the basis of it relies on everyone but the hitter. Whether someone steps into the batters box with a runner on base depends on a ton of factors – all of which are out of the hitters control: where they’re slotted in the lineup, the players directly in front of them in the batting order, the stadium they play in, the teams they face, etc.

A solo home run and a grand slam are essentially the same action; both at bats produced the best possible outcome for a hitter (unless you live in Tim McCarver’s world, where 0 > 1), but the guy who hit the grand slam is supposedly a better “run producer” because his bomb came with people on base? C’mon, we’re better than this. If anything, it’s easier to hit a grand slam than a solo shot since a three ball count with the bases juiced is the most favorable situation a hitter can find himself in.

Embrace: Isolated Power

The formula for ISO% is simple: Slugging Percentage – Batting Average

By essentially eliminating singles, ISO attempts to measure how good someone is at hitting doubles, triples, and home runs (since every hit counts as 1 in the eyes of batting average, while singles are the only hit that count for 1 in the equation for slugging).

Want to know who is the best “run producer?” Almost by definition, it must be the person hitting for the most power. This is the most efficient stat we have to display a player’s prowess for “driving in runs.” Use it until we find something better.

Eliminate: Runs

Same exact argument here as RBI.

For example, players A and B both draw leadoff walks, and both steal 2nd. Player A scores when his #3 hitter hits a double. Player B does not after watching the next three hitters all strike out. How is A’s run any more exceptional than B’s non-run?

Sum stats are dumb.

Embrace: Out Percentage

This is a statistic that I just made up because we have a perfectly fine alternative and don’t give it the respect it deserves. The equation for O% is:

1 – OBP% = Out %

Many broadcasts opt to show batting average instead of on base percentage in their player snapshots, which represents a fundamental misunderstanding of what both stats express (or how baseball works, but I’m hoping for their sake that it’s the former).

If you turn the equation for OBP upside down, it gives you the percentage of times a player makes an out, which has a snappier ring to it. Hopefully a simpler concept than “percentage of the time a player reaches base” will catch on with many of our broadcasts across the country, which is sad to say since the goal of any at-bat is simply to not make an out.

Eliminate: OPS (On Base % + Slugging %)

It’s astounding that we accept and use this statistic, considering that it’s a mathematical abomination.

Does 3/4 + 5/8 = 8/12?

No? Because that’s the model for OPS. These are the formulas for the two percentages that makeup this “stat.”

(H + BB + HBP)/Plate Appearances = On Base %

((1B)+(2B*2)+(3B*3)+(HR*4))/At Bats = Slugging %

At Bats ≠ Plate Appearances

Oops.

Embrace: Nothing, because this ridiculous statistic doesn’t mean anything. OPB and SLG are fine as is.

Eliminate: Errors and Fielding Percentage

These stats are entirely dependent on the home scorer (low home error totals for fan favorites in MLB are like high home assist totals for fan favorites in the NBA), contain a moronic rule that a player must almost always touch the ball for it to be an error, give zero credit to run saving defensive efforts, and dole out no penalty to lazy plays on balls that a player should have reached.

If someone cites this stat (or fielding percentage) to argue on behalf of or against a player’s defensive credentials, the only proper response to their “argument” is “yes, that is a number.” Errors and their decimaled buddy are the most meaningless statistics in sports.

Embrace: MLBAM’s new tracking data will tell us in a few years. Until then, trust your eyeballs and defensive WAR (known as “The Andrelton Simmons Rule”).

Eliminate: WHIP

Another mathematical blunder, but not quite on the scale of OPS.

(Walks + Hits)/Innings Pitched = WHIP

In the numerator, we measure events that can only occur in a singular plate appearance. So, why are we using innings in the denominator? One inning can be three plate appearances, or nine, or, well, you get it. It removes a lot of context from what could have been a pretty simple and useful statistic.

Embrace: WHPA 

(Walks + Hits)/Plate Appearances = Duh

Eliminate: Batting Average

The simplistic formula behind batting average is enough proof of its crudeness and relative uselessness in trying to compare players, so there’s no need to mount an argument against it, we’ll just let it do the talking.

Hits/At Bats = AVG

Hits = 1B = 2B = 3B = HR

Is drawing a walk not a part of “batting?” Is a home run the exact same thing as a single? If your answer to either of these questions is yes, then this is a great statistic for you to use.

Embrace: wOBA

Wins Over Replacement is the trendy sabermetric stat adopted by outlets like ESPN and others who are glancing at Fangraphs and Baseball Reference’s pages for the first time. But WAR should not be the metric used as the hallmark of player evaluation that many assume it to be (at least not in-season, since it is a cumulative statistic).

Why? For starters, no one agrees on the formula for WAR. Fangraphs and Baseball Reference do not have the same formula, and 30 teams are most likely have different opinions, as well. Needless to say, this stat isn’t settled, especially on the defensive side. WAR also favors guys who play up the middle (for good reason defensively, but it seems to go too far).

Weighted On Base Average is the best all-encompassing hitting stat that we currently have. Per Fangraphs,

wOBA is based on a simple concept: Not all hits are created equal…Weighted On-Base Average combines all the different aspects of hitting into one metric, weighting each of them in proportion to their actual run value. While batting average, on-base percentage, and slugging percentage fall short in accuracy and scope, wOBA measures and captures offensive value more accurately and comprehensively.

If it sounds complex, that’s because it is. Just look at the formula in that link above, and just because the math is complicated doesn’t mean the logic is. Of course a triple is worth more than a single, but is it exactly three times as valuable as a single like slugging percentage says it is? Doubtful, baseball is never that simple. wOBA has more savvy to it than just counting the number of times one has to run 90 feet and multiplying it by the total.

Check out the leaders in wOBA from last season. That’s pretty much what you’d expect the leaderboard for best hitters to look like if you took a poll of attentive MLB fans. Plus, wOBA is already set to OBP’s scale, so there is no need to feel like you’re in math class while arguing about sports.

Eliminate: Pitcher Wins and Losses

Ban Johnson, the founder of the American League, was ahead of his time on ditching the stat that embodies the absurdity of so many of baseball’s traditional statistics.

Baseball is a game that will rely on the efforts of at least 12 or 13 people on most nights, but one guy (albeit the one with the most individual responsibility) gets credit for the whole team’s effort? Give me a break. Few serious baseball fans bring this stat up in Cy Young arguments anymore, but that does not mean the sport is completely rid of it. Let’s just finish the job once and for all.

Embrace: Nothing, because expressing the totality of any pitcher’s efforts on a binary scale was a stupid idea in the first place.

Eliminate: ERA

We won’t kill it, we just want to redefine it. We can sacrifice wins in ERA’s place. The problem with ERA is that it credits the pitcher with every single defensive effort, so we should look at it as a metric that evaluates a pitcher plus his defense, as opposed to just the pitcher.

For example, even though this play brought the pitcher’s ERA down, no pitching coach on the planet will tell you that this is an out.

Embrace: FIP

Fielding Independent Pitching is not a sabermetric stat. Well, technically it is, but the basic philosophy behind it is not.

Growing up, I only pitched one season out of the 13 that I played, but I spent plenty of time next to pitchers and their pitching coaches, and all I ever heard from our coaches was some variation of:

“Trust your defense and just throw strikes.” 

“Whatever you do, keep the ball down in the zone.” 

“Pick your spots to really ramp it up.”

Or to put it in brute statistical terms: BB and HR bad, K good. Don’t sweat anything else.

It’s crude, but that really is the basic philosophy behind pitching. There’s a reason there are eight other players on the field. Given the compartmentalized nature of baseball, the primary pitching statistic we use should express a player’s talents based on what they can most control, and that’s exactly what FIP does.

Plus, in the same vein as wOBA, it’s already adjusted to ERA’s scale, so you don’t need to relearn calculus in order to compare the best pitchers and hitters.

This will be an easy, seamless transition to a higher quality product. We can do this America. I believe in us. The summer of 2014 will be ours. THEEND

Jacob Weindling
Pure bred Coloradan with a dash of Masshole (go UMass). Sports and politics junkie. If I've learned one thing in life to this point, it's that stupid loses more games than smart wins.
Jacob Weindling
Jacob Weindling

Latest posts by Jacob Weindling (see all)