Yes, relievers are pitchers. You are essentially saying they are not.

No, he's saying the best RP are much less talented and valuable than the best SP.

This is obviously true if you look at contracts closers get vs. SP, or trades. No one ever trades an above average SP for a closer.

Dropping the cutoff to 50ip, Mo falls to a tie for 3rd.

The top 5:

Kimbrel - 160ip, 269
Al Alburquerque - 56.2ip, 264
Mo - 1219.2ip, 206
Robby Ross - 65ip, 206

and

Mike Crudale - 73ip, 197

The idea of snipping out a chunk of a player's career that's equal in innings to another player's entire career and using that to draw any conclusion at all is absolutely bizarre to me. That's not just cherrypicking, it's like a defining example of cherrypicking.

I'm fine with someone saying that comparing Rivera's ability to prevent runs to a starting pitcher is basically a pointless exercise, that's a reasonable view. That just leaves us concluding that he's the best reliever ever by a wide margin, but can't be compared to starting to starting pitchers.

I don't know if there's another person alive that buys into Ray's argument that the 140 innings pitched by Rivera in the playoffs is both irrelevant and insufficient to draw any conclusions about true talent level. I guess I could go run the statistics, but I'd bet quite a bit that Rivera's postseason performance is likely to be a statistically significant difference from his regular season performance at about a 0.001 P value.

If you're drafting players with 100% foresight, and 20 y.o. Rivera and Pettitte are available, who do you take?

I take Pettitte every time. Hell, I take 5 of him to fill my rotation before I draft Rivera. It's just not that hard to find very good relievers.

I think this was an excellent point. I think Pettitte was a better pitcher than Rivera and had a better career. I don't think Pettitte is a Hall-of-Famer. So therefore....

No, he's saying the best RP are much less talented and valuable than the best SP.

And I am saying there is no data whatsoever supporting the "less talented" part when it comes to preventing the other team from scoring on a rate basis. Value no one here is arguing, it's not relevant beyond setting an IP cut off point. You can debate where that point should be, bbref has arbitrarily picked 1k, but if you purposely set out to exclude relievers with the cut off point you are essentially saying they are not pitchers as the should never be included, even when discussing rate basis performance.

Then you purposely NOT figuring out who was the best "pitcher" you are figuring out who was the best starter. Yes, relievers are pitchers. You are essentially saying they are not.

No, I'm saying that relievers are not among the very best pitchers. If I ask, "who was the greatest strikeout pitcher in baseball history?" are you really going to answer "Brad Lidge"? (tied with Billy Wagner for top K/9) I guess you can if you want. But most people will say Randy Johnson or Nolan Ryan (or maybe Walter Johnson if you believe in making aggressive era adjustments). At a career level, I'm just not going to consider anyone with less than 2,000 IP in that kind of discussion. And even if you want to discuss peak, you still have to have a separate discussion for relievers. One simple example: John Smoltz was 7.9 K/9 as a starter, but 9.6 K/9 in relief (as an old man!).

Is this really controversial?

but to date I have not seen a single one of them actually penalize a player for poor postseason play. It's intellectually dishonest.

That's primarily becuase you only hear what you want to hear. I think Posada's poor postseason record damages his HoF case, which is definitely borderline.

As I said, repeatedly, postseason play is not extra credit, which is the mistake you keep making. It's more information about a player's career.

Now, with pitchers, there's some need to account for the fact that the extra innings do count against their shoulders and elbows. Considering Pettitte's postseason work as part of an overall assessment of his career, I'd tend to just add the innings to his body of work. He wasn't any better in the playoffs than he was in the regular season, but the playoffs give his career a little more heft because those pitches were thrown in anger. They shouldn't simply be ignored because Roy Halladay was on some beach someplace where his right shoulder was safe from harm.

...but if you purposely set out to exclude relievers with the cut off point you are essentially saying they are not pitchers as the should never be included, even when discussing rate basis performance.

I think this is a little strong. Relievers are purposely excluded by the IP threshold when deciding rate state "leaders" for individual seasons.

However, this does not mean that people claim things like Kimbrel WASN'T better than Tim Hudson on a rate basis in 2012.

(2) Mariano was an unimpressive starter in 1995, but (as pointed out previously) he learned the pitch that turned him into a HoFer in 1997.

As of that point, Mariano regularly threw 3 pitches (depending on your definition of pitches): a cutter, a 2-seam fastball and a 4-seam fastball. And it stands to reason that he could have included an off-speed pitch if he had been pushed back to starting, but that was not a decision that the Yankees were willing (or chose) to make.

So the argument is that he could have succeeded in a much more difficult role (starting) that he had already failed at, and he could have succeeded using a different style -- including actually learning another pitch. I find that wildly absurd to believe.

The idea of snipping out a chunk of a player's career that's equal in innings to another player's entire career and using that to draw any conclusion at all is absolutely bizarre to me. That's not just cherrypicking, it's like a defining example of cherrypicking.

No, it's pointing out that other pitchers had stretches the same or better than Rivera's entire career -- while pitching in a more difficult role.

But they are lower than him in career ERA+ because they pitched thousands more innings.

I don't see the point in arguing about whether Rivera or Wagner or whomever could have succeeded as a starter. Teams need relievers, so the role/position has value. The important point is that replacement level for relievers is much higher, i.e. it's much easier to get good performance from relievers. Therefore, the same rate performance from a reliever is much less valuable.

By a rough calculation Mariano is 46th all time in runs saved versus average
Whilhelm 27th
Eckersley 95th
Wagner 107th
Gossage 123rd
Tekulve 135th
Franco 141st
Quiz 155th
Hoffman 163rd
Hiller 165th
Lee Smith 168th
Lyle 174th
Fingers 190th
Sutter 194
Orosco 201st

No, it's pointing out that other pitchers had stretches the same or better than Rivera's entire career

Right. Cherrypicking. Choosing to look at what someone did when they were at their pinnacle and comparing it to the entirety of a separate sample is either a stupid thing to do a willfully dishonest thing to do.

By a rough calculation Mariano is 46th all time in runs saved versus average
Whilhelm 27th

Not coincidentally, Mo and Hoyt are the only relievers who belong in the Hall.

Right. Cherrypicking. Choosing to look at what someone did when they were at their pinnacle and comparing it to the entirety of a separate sample is either a stupid thing to do a willfully dishonest thing to do.

Why? I'm not seeing it.

Try applying your logic in any other statistical setting where there's two different sample sizes, let me know if anyone buys it.

? The sample size is the same. ~1200 innings.

"Cherrypicking" is not always bad, so it's bizarre that you think it is. It depends what the exercise is, what the purpose is, and what one is trying to show. Cherrypicking to show that there are samples available for cherrypicking is perfectly valid -- indeed, it's precisely the point of the exercise. The cherrypicking is to show that other pitchers had a 200 ERA+ in the same sample. It is the entire point.

I really hope you don't do anything that involves thinking scientifically for a living. Ignoring massive amounts of data in order to match sample sizes, and picking a set of data that fits a desired belief is exactly what cherrypicking is as a logical fallacy. Your approach to this question is completely invalid, and that's not a matter of opinion.

I think Pettitte was a better pitcher than Rivera.

Why? He wasn't as good at pitching. He was not as good over many more innings however and while that is valuable does that really make him "better"? I don't see anyone questioning Pettitte's humanity in awed tones. This seems to be coming down to the old quantity vs quality debate that really has nothing to do with baseball. If there was some magic pitcher that was able to pitch 100 years and was mediocre but he got the playing time to shatter the counting stats records including WAR would he then be the "best" pitcher ever despite never being what anyone would describe as even good(much less god like Rivera) at any particular moment? Or really when it comes down to it we can divorce this debate from baseball altogether. You don't consider the "best" thing you've eaten to be the thing you've eaten the most that you enjoy at least somewhat. That's not how the word is generally used. We've come down to throwing statistics at what is in the end a question of semantics.

You don't consider the "best" thing you've eaten to be the thing you've eaten the most that you enjoy at least somewhat that's really not how the word is generally used.

Best thing I've ever eaten was probably the Babi Guling (suckling pig) at Ibu Oka in Bali. Worth moving there for that alone.

If there was some magic pitcher that was able to pitch 100 years and was slightly above average but he got the playing time to shatter the counting stats records including WAR would he then be the "best" pitcher ever despite never being what anyone would describe as great at any particular moment?

You didn't get the memo that "greatest" really just means "compiled the most WAR"? It was in the materials packet right next to "playoff games are basically exhibitions".

In terms of the discussion, I guess this is going to be laughed at by some, but I certainly feel it was more special watching Mariano pitch than Pettitte.

I really hope you don't do anything that involves thinking scientifically for a living. Ignoring massive amounts of data in order to match sample sizes, and picking a set of data that fits a desired belief is exactly what cherrypicking is as a logical fallacy. Your approach to this question is completely invalid, and that's not a matter of opinion.

I have a degree in Electrical Engineering.

Taking a small sample of IP and drawing a conclusion based on it -- "Halladay gave up 7 runs in 5 innings - he sucks" -- is flawed.

However, saying "There are other pitchers who posted a 200 ERA+ over a 1200-IP span - here they are" is not.

I have no idea why you would think there is a problem with the latter. I'm not allowed to point out that there were other pitchers who posted a 200 ERA+ over a 1200-IP span? What the hell?

[222] I don't think that's laughable at all. "More valuable" and "Better pitcher" are distinct (though related) concepts.

So the argument is that he could have succeeded in a much more difficult role (starting) that he had already failed at

Remarkably presumptuous. The way you think is impressively geared to finding anything you can throw at the other person without even thinking it through for a second. Why don't you go through the top pitchers of all time in your own estimation and see if their rookie seasons are all over the map in terms of quality. Hell, for many of them you need to give them years before their stats give any indication that they are particularly special or even good.

I'm not allowed to point out that there were other pitchers who posted a 200 ERA+ over a 1200-IP span? What the hell?

Feel free to point it out, as long as you realize it's utterly meaningless to select endpoints out of a sample and then compare it to a complete data set for a different pitcher. It's an interesting factoid, but a useless way of comparing two things.

If you're unable to comprehend why cherrypicking isn't a valid analytical technique, I don't think I can help you, and I have to conclude that you're either willfully obtuse or just dishonest. I suppose I could list a ton of analogies, but you seem to be fundamentally not grasping why it's not analytically valid; even if you can't understand why it's not, you should understand that it's a settled matter that cherrypicking is not a good way to reach conclusions. If you submitted any scientific paper or financial analysis that used this technique, it'd be rejected and you'd be scorned for thinking it was good analysis.

edit - Screw it, here's an an illustrative analogy anyway. If I had a two groups of vaccinated patients, and my readout was antibody titers, but the groups were of different sample sizes, I would not be obliged to simply choose the best consecutive sample of titers from the larger group. If I did so, every reader would say, "what the ####, why would anyone do that?".

The way you think is impressively geared to finding anything you can throw at the other person without even thinking it through for a second.

He simply has a reaction against people caring about the postseason, and everything else flows from that.

Nah, Rivera vs Pedro is likely a debate that hits home for Ray much more so than caring about the postseason .

Rivera vs Pedro is likely a debate

What debate? Is there anyone, even the most delusional Yankee fanboy, that wouldn't say Pedro was much, much, much better than Mo?

Nah, Rivera vs Pedro is likely a debate that hits home for Ray much more so than caring about the postseason .

I'm not really even sure what can be cogently debated about the two. I guess you could argue about aesthetic differences, but there's not really any empirical questions that are interesting to ask about Pedro vs. Rivera.

Coke to snapper. Yeah, there's just not much to say there. Peak Pedro's work is vastly more impressive, more valuable, and higher quality than Rivera's - and I'm a Rivera fanboy.

Feel free to point it out, as long as you realize it's utterly meaningless to select endpoints out of a sample and then compare it to a complete data set for a different pitcher. It's an interesting factoid, but an utterly useless way of comparing two things.

Sorry, but no. It is quite useful to show that other pitchers put up this ERA+ over a 1200-IP sample.

It doesn't mean that Rivera sucks. It just means that he looks far better using b-r's 1,000 IP cutoff list than he actually was -- and that's without even getting into the difference in closing vs. starting.

Yes, I have more IP from Pedro to choose from. But I am not saying Pedro's ERA+ was as good as Rivera's over Pedro's entire career; I'm just saying he had stretches as good.

even if you can't understand why it's not, you should understand that it's a settled matter that cherrypicking is not a good way to reach conclusions.

Once more: the cherrypicking _is_ the purpose. "Here. I can show you other pitchers who pitched this well over that same IP span."

If you're unable to comprehend why cherrypicking isn't a valid analytical technique, I don't think I can help you, and I have to conclude that you're either willfully obtuse or just dishonest.

While this is often/usually true, the difference here is that the size of the sample itself is important. The fact that Mo's entire career fits within those other pitchers' peaks (in terms of innings pitched), makes Mo's performance less impressive. Walt was simply citing some examples to illustrate that point.

Maddux and Koufax have the same ERA+. Are they equally great pitchers? Of course not. To illustrate that, isn't it legitimate for me to point out that Maddux's best 12-season peak (equal to Koufax's whole career) exceeded Koufax's career? Yes, I'm comparing peak to career, but in this case the peak is as long as the career.

The fact that Mo's entire career fits within those other pitchers' peaks (in terms of innings pitched), makes Mo's performance less impressive. Walt was simply citing some examples to illustrate that point.

No, Walt was disputing that "Rivera has been the best pitcher of all time at preventing the other team from scoring on a rate basis vs average" and that was certainly not an appropriate use of cherry-picking.

It is factual to say over their careers Rivera was better at throwing a baseball past hitters on average than Pedro. If you think that is a laughable standard to define who has been the best pitcher I'm just going to laugh back at you as that is the only response that is deserving. Yes you do need a playing time cut off point or the question becomes meaningless but if you think the entirety of Rivera's career is just junk that needs to be filtered out when parsing who is the "best pitcher ever" I do sincerely think you are doing things incorrectly.

No, Walt was disputing that "Rivera has been the best pitcher of all time at preventing the other team from scoring on a rate basis vs average" and that was certainly not an appropriate use of cherry-picking.

It was a good set of examples (except Clemens) to illustrate the point being made. You're really badly tangled up with this "cherry-picking" thing here. You're making Ray look good, which can't be a good thing. :)

You're mixing me up with Spectral!

The examples were good at illustrating the point that other guys could fit Rivera's career (in innings and excellence) within their peak. It was also interesting that it seemed like the only ones who did that were the utmost elite SPs.

But the examples were useless at trying to disprove the claim "Rivera has been the best pitcher of all time at preventing the other team from scoring on a rate basis vs average," which is explicitly what Walt's post #174 was about.

So the argument is that he could have succeeded in a much more difficult role (starting) that he had already failed at, and he could have succeeded using a different style -- including actually learning another pitch. I find that wildly absurd to believe.

209, No Ray, I am not making that argument. I AM making the argument that Mariano pre-cutter is a different animal than Mariano post-cutter. Would he have been an elite starter? Who knows. But the fact of the matter is that the post-Cutter Mariano never got a chance to start, so we simply don't know what the results would have been.

We DO know that he pitched for 15+ years at an elite level after learning the cutter, and even the most anti-Mariano types would agree that Mariano's learning the cutter was crucial for his (Mariano's) career.

It is factual to say over their careers Rivera was better at throwing a baseball past hitters on average than Pedro.

No, it's not.

Rivera got to pitch in the most favorable circumstances. Most often one IP at a time, and only 60-70 IP per season.

That's like saying someone who ran a 9 sec 100M dash on flat ground is faster than a guy who ran a 12 sec 100M up a 30 degree slope.

Maddux and Koufax have the same ERA+. Are they equally great pitchers? Of course not. To illustrate that, isn't it legitimate for me to point out that Maddux's best 12-season peak (equal to Koufax's whole career) exceeded Koufax's career?

That his entire career works as someone else's peak... Yeah, I'm going to say I think the opposite. If your whole career including pre-prime and decline is at the same performance level as someone else's "peak" then you were a "better" pitcher. More valuable? Probably not. Better? Yeah.

That's like saying someone who ran a 9 sec 100M dash on flat ground is faster than a guy who ran a 12 sec 100M up a 30 degree slope.

It is a fact that the runner on flat ground ran faster.

No, it's not.

Rivera got to pitch in the most favorable circumstances. Most often one IP at a time, and only 60-70 IP per season.

That's like saying someone who ran a 9 sec 100M dash on flat ground is faster than a guy who ran a 12 sec 100M up a 30 degree slope.

As I have pointed out repeatedly reliever advantage is not some unfathomable mystery. You are surely aware of the studies, why are you simply ignoring their conclusions to make statements like the above that directly contradict them? You believe they are flawed in some way? Not based on enough data perhaps?

As I have pointed out repeatedly reliever advantage is not some unfathomable mystery. You are surely aware of the studies, why are you simply ignoring their conclusions to make statements like the above that directly contradict them? You believe they are flawed in some way? Not based on enough data perhaps?

They're flawed b/c there have been almost no conversions of elite SP to closer. Smoltz is the only one I can think of.

My point is any elite SP would be an excellent closer. The role is easy in its very nature.

It was a good set of examples (except Clemens) to illustrate the point being made.

Sure, it's a good set of examples to illustrate something that's kind of cute and interesting, which is that Rivera's career is basically of equal quality and size to the absolute top pitching peaks that the best pitchers ever have performed at. That's a fun thing to know, and puts both Rivera and the great starters into perspective. In terms of analyzing their abilities over the course of their careers? Essentially useless.

You're really badly tangled up with this "cherry-picking" thing here. You're making Ray look good, which can't be a good thing. :)

I disagree that someone standing by their use of a formal logical fallacy to make a point looks good.

My point is any elite SP would be an excellent closer.

Why? Care to respond to my points about Pedro's usage pattern in post 200?

Why? Care to respond to my points about Pedro's usage pattern in post 200.

Ok, how do you know Pedro couldn't have handled being a closer? That's pure speculation.

Can you name a really good starting pitcher who failed as a closer?

Yes, pure speculation, just like your conclusion that Rivera or elite SPs would defy the standard reliever advantage, which appears to be the entire basis of what you are saying. Thing is that is not the basis of what I am saying, but rather a response to an argument against what I am saying in kind.

That his entire career works as someone else's peak... Yeah, I'm going to say I think the opposite. If your whole career including pre-prime and decline is at the same performance level as someone else's "peak" then you were a "better" pitcher.

Maddux thru age 30: 139 ERA+, 2365 IP
Koufax thru age 30 (or age 70): 131 ERA+, 2324 IP

Maddux was a better pitcher for a period as long as Koufax's entire career. But in your view they were equally effective at preventing runs, because Maddux's ability to pitch for another 11 seasons reduced his career rate to match Koufax? Seriously?

The Maddux/Koufax comparison is quite different, because it's explicitly not cherrypicked. You're comparing the same chunk of Maddux's career against Koufax's. Maddux then went to on to tack on a bunch of very effective work.

It's using selective endpoints that are selected specifically for peak that I object to. Using end points that make sense isn't a problem.

Yes, pure speculation, just like your conclusion that Rivera or elite SPs would defy the standard reliever advantage,

What do you mean by "standard" advantage? The general estimate is that a reliever's ERA would rise by about 1 run as a starter. That would make Rivera's ERA+ 142 -- still excellent but below Pedro. But that's an estimate for all relievers. My guess is that Rivera's ERA would have been much more than +1.00. But who knows? All you can really do is rate Rivera against other relievers, and replacement level at his "position."

It's using selective endpoints that are selected specifically for peak that I object to. Using end points that make sense isn't a problem.

OK, good so far. But can I also point out that from age 25-36, another 12-season stretch, Maddux put up an ERA+ of around 170? That is, would it be OK to mention in a discussion of these pitchers that for a period as long as Koufax's entire career, Maddux was MUCH better? Or must I then be shunned as methodologically impure?