The Cincinnati Reds (66-74, 36-32 home) host the New York Mets (66-74, 33-39 road) for three games this weekend. The two teams enter play on Friday with identical 66-74 records. Both teams are all but eliminated from their division races (Reds 11.0 GB, Mets 12.0 GB). The two teams are 7.0 back from the second wild card spot with 22 games left.

The Reds were just swept in Baltimore, and are 6-16 in their last 22 games. The Mets are coming off back-to-back wins against the Marlins, and are 20-24 in the second half.

**Projected Lineup**

The Mets have got surprisingly good seasons from Juan Lagares (102 wRC+, 4.0 WAR), and Daniel Murphy (119 wRC+, 3.1 WAR). Murphy is however on the 15-day DL right now. The Mets have got an unbelievably incredible offensive season from Lucas Duda (.251/.343/.478, 26 HR, 77 RBI, 132 wRC+), which prompted the article Is Lucas Duda A Star Now from Mike Petriello on Fangraphs. The Mets have however had very disappointing seasons from former All-Stars David Wright (96 wRC+, 1.8 WAR), and their big 2013-14 off-season signing, Curtis Granderson (95 wRC+, 0.5 WAR).

**Pitching Matchups**

Friday’s matchup of Simon vs Colon is a good old ERA vs FIP debate. Traditional measures say that Simon (3.28 ERA) has been much better than Colon (4.01 ERA). But, if you are a regular reader of this site you know by now that FIP is a much better way of measuring what a pitcher can actually control (K, BB, HR). That means Colon (3.39 FIP) has actually been much better, and much less lucky than Simon (4.41 FIP).

Saturday’s matchup of Cueto vs Gee gives the Reds a significant advantage. I think the numbers below speak for themselves, no matter what you value in a pitcher.

Sunday’s matchup should be a fun one, as Latos and Wheeler have nearly identical numbers (with Latos a much better BB rate, and Wheeler a much better K rate). Zack Wheeler was the sixth overall pick back in 2009 for the Giants, and was traded for Carlos Beltran in 2011. Wheeler’s fastball averages an impressive 95 MPH.

**Defense**

The Mets are pretty good defensively. They rank 9th in the MLB in DRS (27), and 7th in UZR/150 (3.0). The Reds are 3rd in DRS (43), and 4th in UZR/150 (5.6).

Billy Hamilton has been very good defensively, with 9 DRS (8th in the MLB among CF), and 19.5 UZR/150 (3rd among MLB CF). As hard is it is to believe, Mets CF Juan Lagares has been around three times better. Lagares has 29 DRS (most among MLB CF), and 34.6 UZR/150 (most among MLB CF). I hope Steve Smith has seen footage of Lagares’ cannon of an arm.

Well, not all contact is created equal.

http://www.fangraphs.com/blogs/limiting-hard-contact-nl-leaders-and-a-laggard/

Seems the assumptions underlying FIP are in question. Science marches on.

Simon is giving up a lower percentage of line drives than Colon while getting more infield pop ups and grounders. This has resulted in Colon giving up 28 more hits than Simon in only 5 more IP.

The “traditional measures” have it right; Simon has pitched better than Colon this year.

AS a measure of how you need to look closely at the numbers, consider the HR/9 numbers. It looks like Colon is better at avoiding home runs. But that is solely due to the differences in their home parks; in 90 IP on the road, Colon has allowed 10 HRs while Simon in 92 IP away has allowed only 9. WE’ll see how Bartolo’s meatballs fare at GABP; he got rocked here last year and survived only 2.2 IP.

We saw.

According to FIP, Colon pitched badly last night; somewhere around a 5.00.

6-16 is a .273 winning rate. That would extrapolate to a 45 win season across a 162 game season. Obviously nothing that bad ever happens but to have that poor of a winning % over approaching 15% of the season pretty much says this team isn’t really winning any games. It just not losing on those rare occasions when the other team plays even worse.

Recall that one of Sparky Anderson’s tenets of managing was that regardless of anything a manager might do or not do a team would win a third and lose a third and that his job was to help win as many of the remaining third as possible while not causing the team to lose any of them. The Reds currently can’t even get to this threshold.

The following NL pitchers are all better than Johnny Cueto according to FIP:

Santana, Eovaldi, A. Wood, Lynn, Ross, Kennedy, Greinke, Hamels, Strasburg, Wainwright, Bumgarner, Zimmerman, Ryu and Kershaw.

Do you agree with that assessment i.e. that Cueto is the 15th best starter in the NL?

I guess we could go with your anecdotes about the value of ERA and the ERA estimators like FIP, xFIP, and SIERA. Or we could go with the studies like this one of ALL pitchers over EIGHT seasons that shows the fielding independent estimators do a better job of predicting future ERA than past ERA.

http://www.beyondtheboxscore.com/2012/1/9/2690405/what-starting-pitcher-metrics-correlate-year-to-year

Maybe you should stick with the earth is flat.

From the article:

Scrolling left to right we can quickly see what metrics correlate strongly with, say, next year’s Earned Run Average (ERA). ERA itself has a Y2Y correlation of .38. True ERA (tERA) came in at .47, the highest of all the ERA estimators. Fielding Independent Pitching (FIP) had a correlation of .46, followed by SIERA .45 and xFIP .43.

Yawn. That a minor difference at best and a weak correlation. And I certainly don’t claim that ERA, in and of itself, is a great predictor. You actually have to do some work and look at a bunch of information including lifetime performance.

True ERA is a defense-independent metric that is similar to SIERA that takes batted balls into play.

Given this conclusive statistical proof, and that you seem to accept it other than to say you think the difference is small, why would you keep making your poorly substantiated anecdotal assertions that the stats are flawed compared to ERA?

It’s perfectly fair to say that a statistic, like FIP, doesn’t accurately measure a specific pitcher, say Cueto. But to imply a broader point about the stats in relation to each other based on that is flying in the face of research.

And the difference in correlation of .38 to .47 is pretty large. For an ERA of 4.00 that’s a difference of about 0.40 or 3.60 and 4.00. You may yawn over that, but that’s a big gap when describing how well a pitcher does.

You’re arguing something different than Nick did in his article. He claimed that Colon’s better FIP this year indicates that he pitched better than Simon. That’s completely different from what FIP’s supposed “predictive value” is. For reasons given (and not for the Strawman ones you’ve created), I disagree with Nick’s claim.

Nice try. If a statistic is a better predictor of the future is it not necessarily also a better description of how the pitcher has actually pitched this year? Nick’s point is that FIP better isolates the factors a pitcher can control.

Let’s not pretend that your attack on FIP is limited to this one issue (as your broadside at “sabermetric guys” on the game thread proves.

Seems like any time you can’t answer someone’s point you just call it a strawman.

That is based on the assumption that pitchers have utterly no influence on how well a batter will strike a ball. That assumption has been shown to be flawed as I have already pointed by linking to the “contact management” studies. So why should we keep relying on a “stat” that has been shown to be based on an incorrect and contrary to common sense assumption? tERA, for example, utterly rejects that assumption.

And whether FIP does, in general, a “good job” in predicting future ERA isn’t all that persuasive when we look at a specific pitcher and it has consistently failed to do so. Johnny Cueto is an obvious example; why should we rely on FIP in his case? Answer, we shouldn’t.

I already said that. Individual pitchers might be exceptions. But there has to be solid, specific reasoning for it – ex. why you might expect a pitcher to give up more (or fewer) home runs than average (although FIP uses actual home runs). You can’t just look at a pitcher (Alfredo Simon the first few months of this year, for example) and see a gap between FIP and ERA and say there must be something about the pitcher the stat doesn’t capture. There has to be a specific explanation why the pitcher isn’t normal. But that doesn’t make the stat bad. That’s the point I’ve been trying to make with you the past few days.

It predicts slightly better in general on one year to one year; I have not seen data that it is more predictive than say lifetime ERA even as a general rule.

That study I cited before covered EIGHT seasons. That’s an awful lot of careers.

Can we not make the test of whether something is proven or not whether *you* have seen it, especially when you obviously poorly informed on the topic. Nothing wrong with that. Most people haven’t read this stuff (and you’ve read a little more than most). That’s why we write about it, so that it gets discussed.

But instead of dismissing it because *you haven’t seen it* how about allow for the possibility that you have a lot to learn about this topic. And if you want to grasp it, you really should try to be open minded when you read it. Right now, you are so locked into proving your narrow “earth is flat” point of view, you aren’t even giving it a fair read.

That is incorrect. FIP doesn’t rely on measuring factors that a pitcher can “best control”; since it is entirely based on HRs allowed, strikeouts and walks it claims that a pitcher has NO control over anything else. And that assumption is now known to be false.

The point of FIP is precisely that pitchers primarily control strikeouts, walks and home runs, but have little to no control over balls put in play. The fact that FIP better predicts future runs-allowed than ERA does, confirms that. xFIP takes it a step further and normalizes HR based on fly balls rates. SIERA modifies it a bit more, giving credit for ground-ball percentages. But it’s all premised on what pitchers can and can’t control.

You’re kind of showing that you just don’t yet understand FIP and what this debate is really about.

Your condescending post aside, is there a specific study comparing the predictive results of lifetime ERA on the next year’s ERA or not?

Like I said, that study I’ve already cited correlates future ERA with EIGHT years of pitching. That’s a career for a lot of players. You could search ERA and ERA estimation. ERA fluctuates wildly from year to year for many pitchers.

By saying what is exactly in the FIP formula, I’m showing that “I don’t understand FIP”?

Sure Steve.

The figures given were r or the linear correlation coefficient. I was trying to remember my long ago college Statistics course when I described the r’s given as “weak correlations”. Turns out I was right:

A correlation greater than 0.8 is generally described as strong, whereas a correlation less than 0.5 is generally described as weak.

Further:

The coefficient of determination represents the percent of the data that is the closest to the line of best fit. For example, if r = 0.922, then r 2 = 0.850, which means that 85% of the total variation in y can be explained by the linear relationship between x and y (as described by the regression equation). The other 15% of the total variation in y remains unexplained.

http://mathbits.com/MathBits/TISection/Statistics2/correlation.htm

Given the maximum figure cited .47 that is a “weak correlation”. Further, the coefficient of determination would be only 22% meaning 78% of the total variation remains unexplained. And since using even as crude a measure as last year’s ERA to predict this year’s ERA yields a r of .38 according to the study you cited, the difference between them is a trivial one of about 7.5% explanatory power.

The bottom line is that nothing so far devised by sabermetricians does a very good or even minimally adequate job of predicting future ERA.

Who disagrees with this? You’re the one who has been defending in great repetition that ERA is what we need to look at. Your own statistics textbook proves how wrong you’ve been about that. Do you think sabermetricians can’t read the co-efficient data and realize exactly the size of the correlation? You’re the only one? Just because statistics don’t perfectly predict an outcome doesn’t mean we shouldn’t give more weight to those that do better.

Of course *of course* FIP predicts future ERA only so much. But ERA predicts *even less*. And, as I said yesterday, 7.5* explanatory power isn’t trivial. For a 4.00 ERA, that’s the difference between 3.70 and 4.00.

And, of course:

Why? Because the structure of FIP is in no way meant to predict future ERA.

FIP is commonly used in that fashion because it does a fairly good job of predicting future ERA, but that is not the statistic’s purpose. FIP is meant to be a describer of a pitcher’s performance that is scaled to look like ERA. It’s best described as a what a pitcher’s ERA should have been. That type of description may make FIP sound similar to a true talent evaluator, but it is in no way correlated or meant to describe future performance.

http://www.hardballtimes.com/standard-deviation-and-era-estimators/

Of course that’s not what FIP was designed to do. It was designed to evaluate a pitcher on the factors he can best control (just like Nick described it). Nonetheless, these stats do predict future runs-allowed better than ERA.

Lots of things are initially designed for one purpose but end up having another one as well.

Kevin, we disagree on many things but I think the one thing we probably agree on is that blind reliance on ratios and derivatives and derivatives of derivatives as evaluative tools can work out just about as awkwardly as they did in the financial markets in the late 1990’s/ early 2000’s and again in 2008.