I was on Sean’s case like crazy regarding ERA+, as many of you know. Basically, while every other index stat in the world did the value of the metric divided by the “average”, ERA+ did the league average divided by the ERA of the player. In effect, instead of ER per IP, it was doing IP per ER. What made it worse is when people started to use this in calculations, using it for simple averages etc. The math did not work out.
I had proposed that he do it the consistent way, which would mean someone who gives up runs at half the league average show as 50, rather than 200. Sean was rightfully concerned that people are used to “bigger is better”, and so, that would look like a sticker shock.
Guy proposed something very simple: 2 - ERA/lgERA then times 100. This way, what would look like 50 for me would show up as 150. And the top end is 200 in the Guy method (or 0 in my method). And Sean did just that.
I was relentless in such a seemingly small thing. But it was important to show the symmetry of 50 and 150 to hold. Either my method or Guy’s method would have done that (and the original version of ERA+ did not hold to that). Kudos to Sean for being good enough for taking the brunt of my esotericness (esotericality?). For all the crap I gave him about it, I deserve to give him his kudos just as much.
Repoz
Posted: March 24, 2010 at 10:15 PM |
249 comment(s)
Login to Bookmark
Tags:
general
Reader Comments and Retorts
Go to end of page
Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.
Yes, Bonds' 02 year added more runs to his team than Pedro's 99. But it is not true that each run contributes the same # of wins. We're used to the semi-universality of 'Y runs = 1 win', but if Pedro saves 3 runs in a game by limiting a team to 2 per 9 instead of typical 5 per 9, that is a greater portion of a win than Bonds going W-W-HR-W-W and helping his team score 8 instead of 5.
Probably not, they just think "sligtly better than 1/6th of the league average". That's what I do anyway, and it's really all the granularity you need at the extremes. And for numbers within normal range, when you've been using them for years, you simply develop a sense for the scale. And with the new version, that sense is going to get completely scewed over. As I've said, I'll get over this, but it does make me worry about the casual fan, who does not have such a firmly rooted understanding of the metrics.
To me, this proposal seems like the worst of both worlds. You either care about the casual fan, in which case you don't screw them over by pulling the rug out from under their feat, for what is mostly an aesthetic change. Or you don't care about them, in which case the whole "higher must equal better" conundrum seems completely misplaced.
But why not do what I say and think that 160 means that it's 60% below the league average?
Honestly habit. I'm just not used to thinking about ERA+ that way. For every new ERA+ number I've seen in this thread, I have simply converted it back to the form "x% of the league average" instead of seeing it as "x% above/below league average". I mean it's not exactly rocket science, but I'll admit that it's mostly personal prefernce.
To read that chart: a newERA+ of 140 corresponds to a win percent of .700. And if you give that guy 162 IP (18 decisions), then he's +3.6 wins above average (+.200 * 18).
An OPS+ of 140 corresponds to 118 runs created (given 700 PA, and the average is 84 RC). That's +34 runs above average.
Really? You're going to run with a comparison of min IP for qualifying for the ERA title, to a number of PA's that is close to the league leaders in that category. 2009 PA leader: 734, about 5% more than the 700 you used. IP leader 240 almost 50% more than the 162. I don't call that comparable. More comparable than the old one sure. I also don't see why this should be desireable in the first lace, ERA and OPS aren't comparable either.
That he was really awesome?
Honestly, the new calc makes sense, you can now say he is 75% better then the league, as opposed to the "He was really really awesome". Just that "bigger" numbers look more impressive
It adjusts for park.
It adjusts for era.
It's centered at 100.
That was true before and this new version doesn't change that. Does it have some improvements in some other areas? I really don't care -- other areas are of negligable importance, and ain't worth any ensuing confusion that comes with changing the numbers around. ERA+ is already out there and works fine at what it's supposed to do. Arguing against that is like calling for Betamax in a VHS world.
Are the super-huge ERA+s of the old version weird looking? Sure - then again so are any negative ERA+s by the newly proposed version.
***
The point I was trying to make is that 162 IP ~= 700 PA (I should have noted that). So, "pound-for-pound", ERA+ and OPS+ is now on a similar scale.
***
Anything outside of the 80-140 ERA+ will give you something that is off-scale between the two ERA+ versions. The oldERA+ would go to 600 or 700 or infinity on the good side, while it would go to 0 on the bad side. The newERA+ will go to 180 or 200 (max) on the good side and in the negatives (or infinity) on the bad side.
I'm not sure that this is a "problem". It's the result of rescaling.
Right.
Chris, you are saying it works fine, but how many times have you seen even the brightest people average out oldERA+? You qualify it for "what it's supposed to do", but you're putting your personal use experience to what it's supposed to do.
To me, when I see two guys with 100 IP and an ERA+ of 200 and 67, would most people understand that the average is 100?
What Sean was presenting was IP per ER (relative to league), not ER per IP (relative to league).
But under the old ERA+, 280 wasn't twice as good as 140 either. In fact, depending on how good the two pitchers were on an absolute basis, "twice as good" could have been represented by very different numbers.
For example, in a league with a 5.00 ERA:
Pitcher A - 4.00 ERA, 125 old ERA+
Pitcher B - 3.00 ERA, 167 old ERA+
Pitcher C - 1.00 ERA, 500 old ERA+
Pitcher B was "twice as good" as Pitcher A, but his ERA+ is much less than twice as good.
Pitcher C was "twice as good" as Pitcher B, but his ERA+ is much more than twice as good.
Under the new system, the relative relationships are clear -- Pitcher A has a 120 ERA+, Pitcher B has a 140, and Pitcher C has a 180. It may take some getting used to, but it is consistent.
--------------------------
FYI, I think the aesthetic complaint about the new calculation is due to the fact that it doesn't capture the rarity of extreme performances the way the old one does. The scale exaggerated the extremes, but that seemed to capture the difficulty of the accomplishment if not actual value of it.
Not saying I agree with this complaint, but that's likely what's driving it.
Last season, Wil Ledezma had a 9.53 ERA with the Nationals. The Nationals had a pitcher park factor of 100, which should simplify things. The league ERA last year was 4.19.
So:
(2-9.53/4.19)*100 = -27.4463007
Yet Ledezma's ERA+ is listed as -13.
The problem isn't the rescaling. It's that BB-Ref has the wrong numbers for players with ERAs of 7 or higher, compared to the results I (and others) get using the new formula. And this is ONLY true for players with an ERA of 7 or higher.
The problem with averaging ERA+ doesn't seem like a real problem to me. All of that stuff is done with a calculator, so complicated formulas aren't as much of a problem. In contrast, if you're like me and you like to do mental math to estimate things, simple ratios are much nicer to work with than the formula for ERA#.
You've also destroyed the symmetry of the definitions for ERA+ and OPS+, which makes both statistics less appealing.
Bottom line is that while the new statistic might be more useful for the specific application it's intended for, I find it less intuitive and harder to think about than the old one.
1) Each point on the scale has the same value. The difference between a 120 pitcher and a 100 pitcher (in value to the team) is the same as the difference between a 160 pitcher and a 140 pitcher, or 100 and 80.
2) It tells you the same thing as OPS+, which is what many people have always assumed it did: how much better or worse a pitcher/hitter is at preventing/creating runs, compared to the average pitcher/hitter.
When people say "why mess with it?" I think they are giving too little weight to the fact that a lot of people misunderstood the old metric. That's been revealed even in comments in this thread. For example, people often said that a 130 ERA+ meant a pitcher was 30% better than average. That wasn't true. But now, it is.
As for the "weirdness" of negative values, I would just say 1) giving a guy with a 9.00 ERA and ERA+ of 48 is a much less accurate estimate of his suckitude than -9, and 2) how much do we care about these pitchers in the first place?
After all, the objective of a pitcher is to prevent runs, just like the objective of a batter is to score runs. League average does not enter into it. An ERA+ of 200 should mean that the pitcher gives up half as many runs as the league.
Extensive quantities should not be normalized. They should have units like runs, wins or "Vorpies" (as Posnanski called them). What are the units of "ERAplussies"?
In what way? Seems to me that symmetry has been newly established.
You must not be a Royals fan.
I would just like to say that I appreciated this addition when I stumbled upon it in the course of browsing.
(OBP/lgOBP) + (SLG/lgSLG) - 1
and
lgERA/ERA
How? The OPS+ formula is (OBP/lgOBP + SLG/lgSLG - 1)*100
The newERA+ formula is (2 - ERA/lgERA)*100
The only difference is where the "1" or the "2" appears in the formula: which reflects that higher is better for OBP/SLG while lower is better for ERA. (Edited because I had the OPS+ formula wrong initially)
More importantly, as Guy pointed out on the previous page (post #95), OPS+ and newERA+ are now scaled the same: in terms of runs above/below average. I think that peoples' confusion is stemming from the fact that the old ERA+ numbers seemed to have the same range as OPS+ (150 is very good, 200 is historic, 70 is replacement-level). But that was a distortion. The spread of talent is wider for hitters. Bill James observed this many years ago: no pitcher allows home runs as frequently as Mark McGwire hit them (or as infrequently as Jason Tyner hits them), no pitcher strikes batters out as frequently as Mark Reynolds strikes out, no pitcher walks batters as frequently as Barry Bonds walked.
I tend to be pretty conservative and agree with the suggestion that perhaps replacing ERA+ with ERA% or ERA# or even ERA++ would be preferable. But newERA+ seems to pretty clearly be more "correct".
Fair enough, but I don't think it's that big a deal. Or rather, I don't think it's that big a deal now that B-ref lets you click on a couple years and get the cumulative results for that period.
Chris, you are saying it works fine, but how many times have you seen even the brightest people average out oldERA+?
Heck, I'm one of those people - well, aside from the brightest qualifier. I haven't needed to do that since b-ref let me click on 1991 and 1996 and let me see what Greg Maddux's cumulative numbers were in that period (to take a totally random example).
Yeah I missed when that was first unleashed, and absoutly love it. Possibly the best addition I have seen to the site(which has had many many great additions from the begining)
In your case, in order to get a newERA+ of -13 with an ERA of 9.53 would mean a leagueparkERA of 4.47 or 4.48.
On the Nationals page, Villone has an ERA+ of 99, meaning that his ERA is close to the leagueparkERA. His ERA is 4.25, so taht makes the leagueparkERA as 4.19 to 4.22.
Hmmmm... looks like you got something.
Also on that team is Garate with a 22.50 ERA. In order to get a -253 ERA+, the leagueparkERA needs to be 4.97.
If I can guess, Sean probably did something to avoid the problem of division by zero. He might have added say 1 ER and 2 IP to everyone's totals before doing the calculation. Something like that.
For example, here is the implied baseline for each pitcher on the NAtionals team, sorted by IP:
impliedBaseline IP
4.22 206.1 John Lannan*
4.22 105.2 Craig Stammen
4.21 91.1 Jordan Zimmermann
4.23 91.1 Garrett Mock
4.20 85.2 Shairon Martis
4.23 77 J.D. Martin
4.24 75.2 Ross Detwiler*
4.22 62.2 Scott Olsen*
4.20 60.1 Tyler Clippard
4.24 50 Mike MacDougal
4.21 48.2 Ron Villone*
4.22 48.2 Livan Hernandez
4.25 48 Jason Bergmann
4.24 40 Daniel Cabrera
4.25 39.2 Joe Beimel*
4.24 38.1 Saul Rivera
4.25 35 Julian Tavarez
4.24 32.2 Joel Hanrahan
4.24 30.1 Collin Balester
4.25 28 Logan Kensing
4.24 26.1 Kip Wells
4.27 25.1 Sean Burnett*
4.27 22.1 Jorge Sosa
4.31 15 Jesus Colome
4.36 10.1 Zack Segovia
4.36 9.2 Mike Hinckley*
4.42 7.1 Marco Estrada
4.47 5.2 Wil Ledezma*
4.50 5 Steven Shell
4.97 2 Victor Garate*
If I am correct here, then the issue you describe is not based on the high ERA, but on the low IP.
(2-7.24/4.35)*100 = 33.5632184
BB-Ref is 35.
That's close. Maybe it IS the IP issue.
OPS+
new ERA+
old ERA+ and new ERA+ on the same chart
The x-axis on all 3 is player's sorted by last name.
I still would have preferred ERA# or some other new name, with a gradual phase-out of the old stat perhaps. That would have been less controversial than what was actually done. I think Basketball-Reference did something similar with one of their uber-stats a couple years ago, and now I can't even remember what it was called.
(post edited)
(2-7.14/4.49)*100 = 40.9799555
Listed is 40.
Pitcher A - 4.00 ERA, 125 old ERA+
Pitcher B - 3.00 ERA, 167 old ERA+
Pitcher C - 1.00 ERA, 500 old ERA+
Pitcher B was "twice as good" as Pitcher A, but his ERA+ is much less than twice as good.
Pitcher C was "twice as good" as Pitcher B, but his ERA+ is much more than twice as good.
Under the new system, the relative relationships are clear -- Pitcher A has a 120 ERA+, Pitcher B has a 140, and Pitcher C has a 180. It may take some getting used to, but it is consistent.
You do realize that by that logic somebody with an ERA of 4.98, was "twice as good" as a pitcher with an ERA of 4.99. And both were infenitely better than a pitcher with an ERA of 5.00. In other words: This is complete horsesh1t.
There's no such thing as "twice as good" in a vacuum. Twice as above-average from 140 is 180, though. That's intuitive, and it wasn't true before.
impliedBaseline
4.22 John Lannan*
4.22 Craig Stammen
4.21 Jordan Zimmermann
4.22 Garrett Mock
4.20 Shairon Martis
4.23 J.D. Martin
4.23 Ross Detwiler*
4.21 Scott Olsen*
4.22 Tyler Clippard
4.25 Mike MacDougal
4.21 Ron Villone*
4.22 Livan Hernandez
4.25 Jason Bergmann
4.23 Daniel Cabrera
4.27 Joe Beimel*
4.22 Saul Rivera
4.24 Julian Tavarez
4.21 Joel Hanrahan
4.22 Collin Balester
4.22 Logan Kensing
4.22 Kip Wells
4.30 Sean Burnett*
4.24 Jorge Sosa
4.24 Jesus Colome
4.27 Zack Segovia
4.34 Mike Hinckley*
4.34 Marco Estrada
4.28 Wil Ledezma*
4.43 Steven Shell
4.17 Victor Garate*
So, I would say that's what Sean is doing, that to deal with the division by zero issue he was getting with the oldERA+ (ER=0), he had to do a little fudge. And it looks like he kept this fudge in the newERA+, when it's much less of an issue with IP=0.
There's no such thing as "twice as good" in a vacuum. Twice as above-average from 140 is 180, though. That's intuitive, and it wasn't true before.
Wasn't my wording, hence the quotes...
I guess I just didn't quite realize all of the issues with old ERA+.
The issues seem to be:
1. inertia
2. name confusion
That doesn't seem to be a good enough reason to stop progress, but just good enough to be more careful.
But what I don't get is why those people care about ERA+ in the first place. It's a rough stat in either circumstance.
Which is why I prefer the symmetry old system. It lets you know that a guy with a 200 ERA+ is world-class - in the same way that someone with a 200 OPS+ would. Neither of them are perfect stats or anything close. But they give the average fan a slightly more sophisticated way of making a quick estimate.
In short: I'm sure the new version is slightly better. It just doesn't seem worth it to change things to gain that small advantage.
I would, yes. Because I'm familiar with the scale and what it generally means, it's very easy for me to immediately picture what sorts of performances those are. I would think of it as 'the average is 100' but I would think that the 67 guy was just about erasing all the benefits provided by the 200 guy.
But it's not like there's been a breakthrough here and that the old way of calculating ERA+ was incorrect (e.g., for calculating park factors wrong or something); it was just being presented on a different scale.
As I've said, I have no real problem with addition, it's the complete replacement I'm against, especially without a acclimitazation period. Although I still vote for dropping the "two minus" part and calling it ERA-minus. I think the minus signifies the "less is better" part of the equation nicely. And I think, what it actually means is more intuitive in that form. But that's just my oppinion...
The new stat can be ERA ♯
ERAs normalized to a standard baseline can be ERA ♭
And unadjusted ERAs, of course, can be ERA ♮
Baldrick: you are obviously not most people.
***
I said "progress", not "fix".
It's a progress when people will see 150 and 50 and that averages to 100 (newERA+), rather than averaging to 75 (oldERA+).
It's a feature (progress?) that the newERA+ has a direct linear relationship to win%.
It's a feature (progress?) that newERA+ has the same scale as OPS+, on a per PA basis.
***
If Sean had a blank slate, and knowing what he knows today, he'd have done ERA/lgERA or 2 - ERA/lgERA. He would not have done lgERA/ERA.
So, the argument always goes back to inertia and marketplace confusion.
Sean, my Total Baseball from 1993 lists ERA+.
What we think it should be called is kind of irrelevant, we'd be happy using stats called §äµµ??, but the media clearly has a problem with esoteric names.
To err is human.
To really screw things up takes a computer.
You, sir, are a gentleman of the first order.
Yes, and I seem to recall that 10 or 12 years ago, Total Baseball had a website that listed ERA+ and OPS+ (though I think they called OPS+ "PRO+" or something).
"Your brain is so minute, Baldrick, that if a hungry cannibal cracked your head open, there wouldn't be enough to cover a small water biscuit."
I'd be more than happy with the ERA# solution, BTW. If the new version is better in a meaningful way, it will gain support over time. And having both side by side will allow some transition period for people to get used to the new scale.
I have to agree with those who say it should have a new name, since it is a new thing.
Whenever ERA+ was invented, it became a "standard", and people got used to it. There is no "intuitive" meaning to ERA+. If you knew nothing about sabermetrics but were told that there was a stat that normalized ERA over eras and ball parks, they could call it whatever they wanted to. What they got was what became known as ERA+
Now we have something that is maybe better, but it is not ERA+ as that term was originally derived. It is a new way of looking at it, and it should have a new name.
If we're voting, +1 for this. ERA/lgERA. Simple as possible. Scaling lower as better is intuitive just like ERA itself. And we intuitively understand the "pop" quality of low numbers. Pedro's 2.00 ERA is substantially more impressive than a 2.50 ERA, and we'll know that ERA- (or ERA# or ERA%) of 40 is substantially better than 50.
There already exists something of a movement, especially around the Blyleven/Morris debate, to assess pitchers as "allowing runs at 75% of league average" or whatever number. At least a small segment of sabermetricia already thinks that way so it's perfectly natural to have a stat capturing it.
I would add a third component, which is that at least some of us (myself included) didn't entirely understand what the old ERA+ was actually telling us.
And I'll add yet another vote for "ERA#" -- as well as note that Sean has (unsurprisingly) been very forthright in saying that he simply goofed in releasing it the way he did (which seems to be by far the biggest complaint).
When I want a quick overview of a pitcher, the first thing I want to see is his ERA. The next thing I want to do is compare it to the league and correct for the park. It's not that ERA+ is super accurate, it just answers exactly the question I want answered.
ERA# sounds like it will be about as good at answering the question as ERA+, so it will probably catch on pretty easily. Plus, I kind of like the idea that you just divide ERA# by two to get the neutral winning percentage, and I could probably think of uses for averaging or adding and subtracting.
I guess I agree with Tango. There are at least a few advantages of the new stat, so the goal should be to avoid confusion.
You can frame it as inertia, and make the opposition sound like Luddites. I would call it "familiarity." Thousands of fans have a general idea of what a 100 ERA+ represents, and what a 150 ERA+ represents, and what a 200 ERA+ represents. They've been seeing these numbers, as Harold points out, for more than 15 years. And now when they look at BB-ref this morning, the numbers will no longer carry that familiar meaning.
Sean posted that he takes the interests of the casual fan more seriously than that of the sabermetric community; this switchover seems to me to be directly at odds with that. 95 percent of bb-ref's readers couldn't tell you how ERA+ is figured, but they can tell you what it represents. Or what it represented.
The people around here have much more knowledge of sabermetric constructs than the public at large, and there's even a significant amount of blowback here. Among the general public, there's going to be tremendous confusion on the part of people who don't care how ERA+ is computed, for months if not for years.
But, it's his site. If he wants to use it to make Tango happy, that's his business.
Full Story here:
http://www.baseball-reference.com/blog/archives/5159
I'm going to assume you mean the part about cracking his head open and eating his brains...
And I suppose OPS+ is still the simple OBP (or OBA) plus SLG, even though Tango and others have demonstrated that OBP should be weighted more than SLG? If ERA+ is changing, might as well change OPS+ too!
Based on the posts last year on my blog, Sean's position seemed to be that going to lower-is-better would be too confusing for the marketplace. This is why Guy proposed what he did. It set the scale appropriately, and gave Sean the bigger-is-better he needed.
I still think Sean is just wrong in this case. Everyone understands that a lower ERA is better. If they also understand that ERA+ is based on ERA, I don't see why it would be confusing that a lower ERA+ was better. As I mentioned previously, the current situation seems like it would cause more confusion than one where ERA and ERA+ were both lower = better.
I think this would be a very useful statistic, but since OPS+ and ERA+ have started to gain some mainstream popularity, it would be best to keep them simpler. For instance, you're not going to have Jon Miller explaining standard deviations on Sunday Night Baseball while Joe Morgan snickers in the background, but you might hear Jon Miller cite a player's ERA+/OPS+ of 110 and indicate that it means he performed 10% better than the league average.
I have to agree with this. It's been several years now and so, of course, now it all seems obvious, but I think it actually took me a little while to get used to the idea that bigger was better for ERA+ precisely for this reason. Baseball fans grow up understanding that fewer runs is better for a pitcher and worse for a hitter (or, at least, I did).
There is nothing wrong with "inertia" per se; you need a better argument than that. Right now the argument for co-opting ERA+ instead of just calling this something else is "Because it makes everyone notice that Guy and Tango are so smart."
Call it ERA-GTT for "ERA-GuyTangoTiger" if you like. But we should really leave ERA+ alone.
I don't see the justification for changing the formula so radically for an established stat that is 20 years old. I agree with the calls to rename the new formula and use the simpler ERA/lgERA formula to distinguish it from ERA+.
I never thought ERA+ was that hard to understand, and the averaging problem was not hard to solve, especially for the hardcore sabermetricians who are advocating the change.
The change will confuse less hardcore folks, and the new formula's cap at 200 is inelegant, not to mention all of the negative numbers that will be produced.
I don't think it was broke, so don't "fix" it.
I've already said:
Did Pete Palmer actually do it as lgERA/ERA?
Jon Weisman's 100 Things Dodgers Fans Should Know & Do Before They Die uses ERA+ throughout the book.
Should I return it?....
As I remember, the early Pete Palmer Total Baseball books used the format of ERA+ that has been in use up until now, with bigger numbers being better.
I don't actually have those early books. I saw them in my local library years ago.
Yes, but you also said, in that same post (#133):
Later, in #139, you said:
So it seemed to me you were still hung up on re-defining ERA+ instead of simply using a different name. If you've now changed your position on that -- a cause you've been in "relentless" pursuit of, see below -- that's fine, but you should probably clarify that the comments and high-fiving in your original blog entry (sorry, glove-slapping) no longer apply:
1. Sean has already said that he didn't mean to roll it out (which was clear by the fact that the old formula remained on the hover-glossary) and that he had intended on giving it a new name.
2. Tango clearly wanted the new formula implemented and is happy to see the new formula presented.
3. Who gives a rat's ass. Honestly, what in the hell are you arguing?
I don't care, at all, what something is called. As long as Sean gives us the ERA/lgERA in some form, be it just like that, or with 2 minus in front of it, I'm good. If he calls it ERA+, ERA#, or quatloos, I don't care. If he keeps the original and puts it side-by-side, fine. All that stuff is extraneous to getting the core down right.
I always used to do it by multiplying a guy's ERA by his "decimalized" ERA+ (i.e. 1.5 instead of 150).
Fair enough. I of course agree that the new calculation is useful and should be presented.
Is there actually anything to adjust there, or do they just use league ERA? I mean I assume that offensive parks have more PA's, and you could normalize that out if you wanted to. But I doubt that has a very large effect.
179: It's adjusting the league average ERA for the pitcher in question's park. EG, "Felix plays in Safeco. An average pitcher should have an ERA of ___ in Safeco."
Things move with lightning speed at bb-ref. You can't afford to blink....
Actually, though, the place where this adjustment would make the most sense would be in the batter vs. pitcher stats, where Sean already calculates BA/OBP/SLG/OPS for hitters vs. pitchers. It'd be a relatively simple add-on there to put in OPS+ for pitchers, which would be exactly symmetrical to hitters' OPS+.
It would even be the place where the "less is better" construct could start out as a way to slowly educate the mainstream audience. You can get a look at what that would look like at the link below (for Pedro), where the figure in question is shown at the far right as "sOPS+":
Pedro's batter vs. OPS+ sample
(Hopefully that worked.) The one thing that Sean needs to add there is the lifetime OPS+ figure, which I estimate (by doing that awful "averaging" thing) to be about 60.
As for the new "lgrelERA," which is really a "runs adjusted lgrelERA+" it needs to be broken in slowly. Adding it alongside the commonly-known stat shouldn't be a problem.
I do wonder, though, why folks who find these stats to be so inadequate and imprecise are spending so much time trying to improve them in ways that can never actually remedy their overall deficiencies. At any rate, this is certainly not the "shadowy" Tom Tango as described by the Toronto press...
Because they're heavily used. So if we can get some of the more fundamental flaws fixed, it's still improvement.
In my opinion, yes. I think it's great that B-R has published OPS+ and ERA+ for the last ten years, as that's been vital in their uptake (first with casual statheads, and then the MSM).
The problem that I see is that in terms of WINS a 3.00 ERA pitcher is further away from .500 (the 4.00 ERA pitcher) than a 5.00 ERA pitcher is, and old ERA+ more accurately reflects THAT:
ERA OLD NEW Pythag1.00 400 175 .926
1.50 267 163 .856
2.00 200 150 .779
2.50 160 138 .702
3.00 133 125 .628
3.50 114 113 .560
4.00 100 100 .500
4.50 89 88 .447
5.00 80 75 .400
5.50 73 63 .359
6.00 67 50 .323
6.50 62 38 .292
7.00 57 25 .265
Established convention would be to call ERA#, ERA sharp, in conversation.
Not according to Harold's chart (link in post 51). Both versions of ERA+ introduce distortions, but the ones in newERA+ are considerably smaller.
I see. I somehow read that as league ERA adjusted for all Ballparks, and not for each ballpark individually, which didn't seem to make much sense.
"ERA sharp" certainly sounds better than "ERA number sign," but of course the # sign is not the same symbol as the sharp sign (sharp has diagonal lines), so the name could be a little confusing.
On a computer, the # and the sharp symbol are one and the same and used that way. C# would be the obvious example. It is always called C sharp.
In music for example, the # refers to "sharp." So, E sharp as opposed to E flat or E. So, it works.
I think this tangent is becoming silly even by internet dialogue standards, which is impressive, and it's my fault... I agree with Randy that people tend to use # as the music sharp in practice because # is on the keyboard, so maybe my initial thought that # would be confusing isn't valid. However, technically, the music sharp symbol is unicode U+266F and the number symbol is unicode U+0023, and the symbols do look different on the screen. For the programming language "C Sharp," Microsoft uses the music symbol in their graphic art and not the number sign.
Obviously this is silly and if the consensus wants ERA# to be the designated name and called "ERA Sharp," I'll go along. I still do like the name "ERA Star" or "ERA Percent" better in any case.
Just a wild guess here, but I think if you ask an average fan the question: "if the league average ERA is 4.00, what ERA is twice as good as that league average one", they are not going to say 0.00. Just a guess on my part.
You must be Registered and Logged In to post comments.
<< Back to main