Why ERA is becoming meaningless to me

[TRF]

So, great game by Latos today so I took a casual look at the Reds starter's stats.

Bailey and Latos have had nearly identical years, with Bailey actually being better in regards to BB's and hits. They have just about the same OPS against .680 for Latos, .675 for Bailey. BAA .245 for Latos, .244 for Bailey.

Latos has given up 62 runs 54 ER. Bailey 69 and 67.

Both have been very, very good this year. All these numbers show it.

ERA? pfft. I mean I'll still look, but I no longer take it as the gospel that a pitcher is pitching well like I would have when I was younger. I barely grasp advanced metrics. But I like the top level ones like OPS, WHIP H/9 etc. Easy enough to grasp, simple components that allow me to look a little deeper without breaking out a calculator.

[swaisuc]

A lot of this thread will depend on whether you believe stranding runners is a skill or if it's essentially random.

[dougdirt]

A lot of this thread will depend on whether you believe stranding runners is a skill or if it's essentially random.

Which goes back to whether or not you believe in BABIP.

Bailey strikes out 24% of batters when runners are on. Latos is at 23%.
Bailey walks 5.3% of batters with runners on. Latos is at 8.2%.
Bailey has a BABIP with men on of .385. Latos is at .279.

Bailey shows better control and misses more bats in those situations, but Latos somehow reduces hits much better. Logically, that doesn't make much sense.

Stats are just for 2013, coming into today.

[Kc61]

So, great game by Latos today so I took a casual look at the Reds starter's stats.

Bailey and Latos have had nearly identical years, with Bailey actually being better in regards to BB's and hits. They have just about the same OPS against .680 for Latos, .675 for Bailey. BAA .245 for Latos, .244 for Bailey.

Latos has given up 62 runs 54 ER. Bailey 69 and 67.

Both have been very, very good this year. All these numbers show it.

ERA? pfft. I mean I'll still look, but I no longer take it as the gospel that a pitcher is pitching well like I would have when I was younger. I barely grasp advanced metrics. But I like the top level ones like OPS, WHIP H/9 etc. Easy enough to grasp, simple components that allow me to look a little deeper without breaking out a calculator.

But take a step back....

The object of pitching is to hold the other team without many runs. Runs.

You can have the best strikeout rate, the best OPS against rate, the best WHIP, the best of all these.

But if you allow a bunch of runs, the impact on games is adverse.

Now, obviously, these other stats are very important. Including the fielding independent stats, and league/park adjusted stats. They are all important.

And sometimes, as Doug notes, ERA is impacted by BABIP which may make it unreliable.

But at the end of the day, if you give up too many earned runs, you and your team have a problem. So ERA needs to be considered IMO.

[VR]

Would it help to understand their peripherals in the stretch vs. windup?

[RadfordVA]

What stat says they are a lot different? ERA?

Even that doesn't seem to different to me. If you drop the last number off it is 3.6 to 3.0. Not a huge difference. Now it is 20% more. But if a guy had a walk rate of 3.0 and one had a 3.6 per 9 innings I don't think we would think much of it. Also I don't think we should make to much of a difference here. As others have pointed out they have been similar just Latos has had slightly better results so far.

Batting average is another number that people put too big of an emphasis on a small difference in numbers. If a guy has a .320 average and next year he hits .296 you think big drop off. But if it was percentage based the difference is 32% to 30%.

[Patrick Bateman]

Jeff Locke has a great ERA. I therefore look past ERA every chance I get.

Kc is right, peripherals aren't everything. Great ERA's year in, year out can't be ignored. Everything needs to be looked at in conjunction. Every stat has outliers that look bad come year end. ERA has a lot of them though, and with some pretty basic analysis a lot of them can be picked out easy, and over time they regress.

[Kc61]

These somewhat small differences in ERA do have meaning. Bailey and Latos have not been equally good this year. Of course, ERA alone doesn't tell you why and upon a quick glance all I can come up with is Homer's lack of consistency. And Latos' rather superb consistency.

Latos has allowed four earned runs or more five times this season. His ERA is now 2.93 after today's game.

Bailey has allowed four earned runs or more ten times. His ERA is now 3.68.

Homer has allowed 7, 7, 6 and 5 earned runs in four of his games. Latos has allowed 6 earned runs twice, nothing else above four.

If Homer had avoided five of his "bad" outings and also, like Latos, had only five of them, I think his numbers and Latos' numbers would probably be very close.

Why has Homer had these poor games? I'd have to dig deeper to figure out what went wrong those days, or to try.

So the OP is correct that ERA doesn't tell the whole story, but it's not meaningless, it does clue you in to the pitcher's effectiveness. Then you have to go deeper.

[Falls City Beer]

Latos has pretty easily been the superior pitcher this season and for each pitcher's career. How that's much debatable is kind of silly. ERA alone doesn't confirm it, but it's an enormous piece of the picture.

[REDREAD]

Yep, as KC's post above illustrates, Latos has been better at consistently keeping his team in the game.

Quality start percentage is another way to track this..

The pitcher's job is to ultimately stop the other team from scoring.
ERA is not a perfect stat, but it's a bit of a stretch to call it meaningless when evaluating the actual results of a season.

I'm not saying ERA is predictative or repeatable, but it's a useful measure to tell us what actually happened.

Homer has had a good year, but Latos has had a better year (IMO)

Stats like BAA are also not necessarily useful for determining the results, since some hits end up being harmless.
Again, some stat analysis assume that the pitcher and the batter approach every plate appearance exactly the same way.
Real pitchers have said differently. Some starting pitchers have said that they will pitch to contact with no one on base and let the defense work for them, but when runners are on base, they "bear down" or try for the strikeout, etc.
Can I prove this? Can I even prove that it makes a difference in results if the pitcher thinks he's bearing down? Of course not. But the opposite extreme of assuming the results of a baseball game are akin to a random number generator, and everything unexplained is luck is also not right either.
To me, it's clear that Latos has had a better year than Homer. No disrespect to anyone that disagrees though.

[wheels]

ERA isn't as awful as wins and losses, or RBI. I think it's fun trying to figure out why a pitcher has that particular number. What lead to that 3.60 ERA? Was he hit lucky? Etc. I get a kick out almost every pitching stat (except wins and losses).

Pitchers are a fun puzzle.

[Kc61]

Wins and losses can be interesting. For a reliever they are fairly telling stats. For a starter, they are useful if very good or very bad. Less so in the middle.

If you look at starting pitchers won/lost together with ERA and other stats you can tell something about the pitcher's luck. If a guy is 5-15 with a 3.25, we know something about his season.

Really, every data point is helpful, even with their limitations.

[RedsManRick]

But at the end of the day, if you give up too many earned runs, you and your team have a problem. So ERA needs to be considered IMO.

There are two issues with this:

1) Just like with run scoring, we can too cute about assigning specific team-level outcomes, namely runs, to individual people. Runs being scored or allowed are, in most cases, a function of the team, not an individual player. Assigning runs to individuals hides this reality and can cause us to lose sight of the role of teammates in producing those outcomes.

While it's a helpful shortcut to use simplistic rules on a day-to-day basis to keep tabs on things (as opposed to doing the more advanced math of divvying up responsibility to all players involved), it too often results in us creating over-simplified, false narratives about how well certain players have performed.

2) Even when we can attribute outcomes to individual players as fairly as possible, we have to be careful regarding our assumptions regarding the predictability of that performance.

If a pitcher has a really high strand rate because of his great K rate with men on base (or say, a hitter's production with RISP....), I'm all for celebrating that he had a great year. But that doesn't mean I'm going to say he's necessarily "skilled at stranding runners" without accounting for the bigger picture of what we know about the ability of pitchers to affect strand rates and what that guy has done outside of the sample we're looking at.

When it comes to ERA, if the question is simply: Which pitcher has been on the mound when the most earned runs scored, sure, it's a useful shortcut. But do we really care about that?

To your comment, ERA is really a team stat. Latos may actually be getting off the hook a little bit because of the defensive struggles behind him showing up as errors whereas Bailey is being giving "credit" for runs scoring when the poor defensive play was counted as hits.

For example, let's say they allow the exact same batted ball: a hard grounder in the hole. For Bailey, Cozart has a slow first step and the ball roles in to LF for a single. For Latos, Cozart had a great first step, fields the ball, and air-mails it to Votto. Both then give up a HR. Bailey has 2 ER, Latos has 1 ER.

Ultimately, the runs scored and that's what matters. But in trying to account for defense, we've created a weird, subjective, incomplete thing called an error that can can be just as much a hindrance as a help to understanding how well a pitcher truly contributed to his team's runs allowed.

[Brutus]

ERA is important because it is a bottom line measurement of production in the thing that matters most (runs scoring).

It's obviously not worth a ton in predicting future performance. There are some better measures for estimating that.

Those two things being said, I think ERA is important depending on the context of what's being discussed.

[blumj]

Unearned runs are still runs, some earned runs are every bit as much the result of fielding as some unearned runs.