I'd also gone to bed before the game was over (I fell asleep after Mussina finished the bottom half of the 6th).

Being pissed about this loss (my first of the year!), especially after reading Cliff Corcoran over at Bronx banter, has lead me to propose propose a new stat for Primer Yankee Fans: the JTBPFUL, or the Joe Torre Bull Pen Fck Up Loss.

This stat would apply to any Yankee loss which is directly the consequence of Torre's "unique" way of managing his bullpen, such as yesterday's.

In order to prevent abuse of this stat, its use must be proposed in the Game Chat of the next Yankee game (it cannot be proposed just after the Yankees loss), and it must be confirmed by at least 5 Yankee chatter, non-troll regulars (I wouldn't qualify, but I think we pretty much know who would qualify and who - Kevin - would not).

If nothing else, this stat should give us Yankee fans something to rail at.

I encourage any other Primer Yankee fans to develop the stat concept a little bit more (i.e., I haven't gone into the specifics of what would constitute a JTBPFUL, just the general outline therein).

I supposed I'd be more upset if, with the Harden vs. Mussina match-up, I hadn't mentally chalked this game up as a loss. Annoying to lose that way, of course (annoying to lose any way), but less so than it might be under different circumstances

Well, what Joe Torre cost was the difference in the chance of winning with Rivera pitching (probably just north of 50%) and with Proctor in the game (considerably less than 50%).

Now, I'm a pretty big Torre supporter most of the time, but this move was absolutely astounding. I laid in bed afterward trying to come up with what his rationale for using Scott Proctor could possibly be. Nothing. Farnsworth had only thrown like 10 pitches. The best reliever in the history of the game ended the game in the bullpen. Infuriating. I like Scott Proctor, and think he's better than he gets credit for, it just makes no sense to use him when there are two better options available.

On the Oakland broadcast they had a clip of Huston Street saying how awesome it would be to go head-to-head with Mariano in the ninth inning of a tie game. Obviously he took for granted that a team's best reliever should be pitching in that spot.

Nothing about having one of the 15-20 best hitters in the league bunt in the 6th (or 7th?) with no one out. Great, he's just going to keep doing it.

D'ya think the umps might call Jeter out on strikes once in a while? That was embarrassing and ridiculous and criminal for all concerned.

one of the 15-20 best hitters in the league

Uh, not really. He's a good hitter, a great hitter for a SS, but not top 15-20.

Yet another reason why I would have preferred that the Yankees go with 12 pitchers instead of three catchers, but I guess Torre would have found a way to blow that too.

Torre set up Proctor to fail.

"He's a good hitter, a great hitter for a SS, but not top 15-20."

Two out of the last three years he's ranked in the top 20 in EQA in the AL (16th last year, 19th in 2003).

He projects exactly 20th in the majors RC by ZiPS, though only 91st in RC/27 outs.

This initially surprised me, but after I thought about it I realized that his durability is a very good asset to have. So I guess it depends on whether you would define his durability has one of his "hitting" attributes.

PECOTA projects him for the 34th best EQA in the AL this year. Don't forget about Thome, Pena, Glaus, and the other new additions to the AL. Also, EQA includes SB/CS (which Jeter obviously excels at), which isn't the best way to judge how good a hitter a player is.

"PECOTA projects him for the 34th best EQA in the AL this year."

I bet he posts another .300 EQA this year. But it's not really a serious point for me, more a throwaway comment to see if anyone would respond.

"Don't forget about Thome, Pena, Glaus, and the other new additions to the AL."

Pena won't out hit Jeter this year, none of those guys outhit him last year (per EQA). I would guess one of those three will outhit him this year.

"Also, EQA includes SB/CS (which Jeter obviously excels at), which isn't the best way to judge how good a hitter a player is."

Good point.

I think we can still agree, even if we disagree about his exact ranking or who's better then him, that having a hitter of Jeter's quality bunt in that situation is foolish.

I'll agree with that. Kendall bunting, though? Brilliant.

The best reliever in the history of the game ended the game in the bullpen.

When did the Yankees sign Hoyt Wilhelm?

He is a little long in the tooth, but no more than Bernie... and he is likely still better than proctor.

"Kendall bunting, though? Brilliant."

Definately a good move. Hopefully someone explains to Joe (or he wakes the #### up again and regains his interest in the game) that the context is what made that a good move.

What, the 20 managers on the Yankee bench couldn't make the right call?

Tough loss. Someday, perhaps, the Closer Reign of Terror will end and people will look back and say how silly it was to use your best reliever with a three-run lead but not in a tie game. Some sweet day.

I'd like to see Larry Bowa and Joe Torre get into an arguement just to watch.

Someday, perhaps, the Closer Reign of Terror will end and people will look back and say how silly it was to use your best reliever with a three-run lead but not in a tie game. Some sweet day.

It just simply has to happen, someday. The sheer towering idiocy of the choice Torre made last night -- the choice managers have routinely been making for about 15 years now -- will someday penetrate the thick fog of Save Worship. Won't it?

The dumb move wasn't in not bringing in Rivera, it was in removing Farsworth for Proctor after Farnsworth had only thrown 10 pitches.

"The dumb move wasn't in not bringing in Rivera"

You left of the "est". They were all dumb moves.

The dumb move wasn't in not bringing in Rivera, it was in removing Farsworth for Proctor after Farnsworth had only thrown 10 pitches.

Bringing in Proctor to replace Farnsworth was obviously daft. But the bottom of the ninth of a tie game is an extremely high-leverage situation; having anyone in the game other than your best reliever (particularly when he's warmed up and well-rested) is even more daft. Bringing Farnsworth into the game in the eighth inning ahead of Rivera is of highly debatable wisdom.

When did the Yankees sign Hoyt Wilhelm?

Hmmm... At least two guys I can think of stake a better claim to that title than Wilhelm... but I think we can dispense with Goose's credentials and talk about Mo.

ERA: Hoyt - 2.52 v. Mo - 2.33 EDGE: Rivera
ERA+: Hoyt - 146 v. Mo - 197 EDGE: Rivera
K/9: Hoyt - 6.43 v. Mo - 8.12 EDGE: Rivera
WHIP: Hoyt - 1.13 v. Mo - 1.05 EDGE: Rivera
Decisions(W+Sv-L) Hoyt - 370-122 v. Mo - 433-35 EDGE: Rivera

Of course Wilhelm did it forever seemingly - 20 seasons, 2254 innings, and Mo's only got 11 seasons under his belt and just over 800 innings, but based on rate stats not counting numbers he crushes Wilhelm.

Take your snarky bitterness elsewhere Mr. Warped Standards.

McNally - quantity matters my friend. Not to mention the decisions stat is wrong, since they played under different save rules.

If you don't want to count quantity that River is no Huston Street.

ERA: Street - 1.72 v. Mo - 2.33 EDGE: Rivera
ERA+: Street - 260 v. Mo - 197 EDGE: Rivera
K/9: Street - 8.3 v. Mo - 8.12 EDGE: Rivera
WHIP: Street - 1.01 v. Mo - 1.05 EDGE: Rivera

Rivera has pitched half as many innings, and only been marginally better. Another 5 year season without falling off a cliff, and I'll agree with you.

Take your poor analysis elsewhere, we would be better off if your standards were as high as mine. Then again you wouldn't be a yankee fan.

McNally - quantity matters my friend. Not to mention the decisions stat is wrong, since they played under different save rules.

No they didn't, all "saves" are applied retroactively under Holtzman's definition after 1969.

I wouldn't call being 50% better vs. the league "marginally better."

You say quantity matters, but your strawman Street argument is based on the premise that we can compare a guy based on 10% of the work - so that's OK, but about 40% isn't? Which is it?

Mo's peak his higher, he's sustained that level longer than Wilhelm did - considering since he became a fulltime player it's essentially been all peak.

Mo's worst year (2000) would fit pretty neatly with a random collection of Wilhelm's best years.

ERA: Hoyt - 2.52 v. Mo - 2.33 EDGE: Rivera
ERA+: Hoyt - 146 v. Mo - 197 EDGE: Rivera
K/9: Hoyt - 6.43 v. Mo - 8.12 EDGE: Rivera
WHIP: Hoyt - 1.13 v. Mo - 1.05 EDGE: Rivera

Are you seriously presenting a comparison of relievers from 40 years apart, using only rate stats?

And not even making mention of the fact that Wilhelm made 52 starts in his career, and the one season in which he was primarily a starter he blew both leagues away in ERA and ERA+?

You are joking, right?

Mo's worst year (2000) would fit pretty neatly with a random collection of Wilhelm's best years.

Next time Rivera puts up back-back seasons of 73 and 66 games and 131 and 144 innings with ERA+ figures of 173 and 176, you will let us know, right? Thanks.

Comparing Rivera and Wilhelm is a fairly ridiculous apples-and-oranges situation, the way they were deployed in the eras they pitched are so fantastically different. That being said, if one was presented with Torre's situation yesterday and had both Wilhelm and Rivera in the pen, it seems like Rivera would be the obvious choice

That being said, if one was presented with Torre's situation yesterday and had both Wilhelm and Rivera in the pen, it seems like Rivera would be the obvious choice

Rivera might well be a more effective pitcher than Wilhelm if they're both limited to essentially 1-inning stints. But that very limitation is what makes the comparison so difficult and interesting, and why it isn't obvious in the least that Rivera is a better, or certainly not a more valuable, pitcher than Wilhelm. Because with Wilhelm in his bullpen, no manager in his right mind would have bothered to wait until the ninth inning to face the question of bringing him in. Wilhelm's standard stint was two innings. He'd have already pitched the eighth and been more than ready for the ninth. This capacity allowed his team to not even need to carry that Scott Proctor-style waste of roster space.

Appear, blast you!

Keep banging that drum Treder. We'd all love to here some more about this, for the 9,000 time.

Of course Wilhelm did it forever seemingly - 20 seasons, 2254 innings, and Mo's only got 11 seasons under his belt and just over 800 innings, but based on rate stats not counting numbers he crushes Wilhelm.

And based on rate stats, Barry Bonds wishes he hit like John Paciorek. Nevertheless, quantity <u>does</u> matter. Would Rivera's rate stats be as good as Wilhelm's if he'd been asked to pitch as much as Wilhelm pitched? I suppose it's possible, but I certainly wouldn't bet on it (if there were an alternate universe in which we could see the result). Rivera's certainly been awesome in the innings he's pitched. Perhaps he's the best pitcher per appearance ever, but that doesn't necessarily make him the most valuable (in terms of overall value added).

Let's play with numbers for a second:

Wilhelm pitched in 2254 innings over 21 seasons. That's an average of a little over 107 innings per season. He had a career ERA+ of 146.

Rivera has pitched in 806 innings over 11 seasons. That's an average of a little over 73 innings per season. He has a career ERA+ of 197.

In order for Mariano to match Wilhelm's production (on a seasonal basis), he'd have to average roughly 34 more innings per season... but with an ERA+ in those innings of about 37.

Using this (flawed) means of comparison, Rivera's still more valuable than Wilhelm, on average, per season. Of course, he needs to keep it up for another 10 years...

And of course this rough analysis ignores the fact that Wilhelm hurts himself by pitching 20 innings at age 48 and 25 innings at age 49, lowering his ERA+ slightly and adding two "seasons" onto his career total without adding a significant number of innings. Let's subtract those innings and seasons (but for the sake of ease of calculation, we'll keep the same career ERA+) and recalculate for a second...

Wilhelm would then average 116 inning per season. In order to match him on a seasonal basis, Rivera would need to throw another 43 innings per season with an ERA+ of 59. To put that in context, the average Wilhelm season was roughly the average Rivera season + the 2005 Yankee versions of Alan Embree and Paul Quantrill. Again, by this rough measurement, Rivera seems better to me on a seasonal basis, but Wilhelm managed to pitch that well for 8 more seasons than Rivera has thus far.

I suspect Rivera has not yet caught Wilhelm in terms of value-contributed over the course of his career, despite the quality of the innings he's thrown thus far.

We'd all love to here some more about this, for the 9,000 time.

It will still be true, 9,000 more times. And because it's still true, and because otherwise intelligent people like Sean fail to comprehend its significance, and compare 2000s style closers to 1960s style ace relievers with no consideration for usage pattern, it remains a drum that deserves banging.

I suspect Rivera has not yet caught Wilhelm in terms of value-contributed over the course of his career, despite the quality of the innings he's thrown thus far.

Bingo. And as another element of how quantity matters, Rivera's tightly limited usage pattern requires his teams to carry an extra pitcher on their roster, at the sacrifice of a position player.

9,001 and counting!

Rivera's tightly limited usage pattern requires his teams to carry an extra pitcher on their roster, at the sacrifice of a position player.

Yeah, but if that pitcher manages to pitch better than Paul Quantrill and Alan Ambree combined did for the Yankees in 2005, then the Yankees will have gotten better pitching from Rivera + X than a team would from Wilhelm. It's not THAT hard to outpitch Quantrill/Embree. The loss of a roster spot is bad, but is it as bad in a league with a DH where PHing for the pitcher isn't as important?

Yeah, but if that pitcher manages to pitch better than Paul Quantrill and Alan Ambree combined did for the Yankees in 2005, then the Yankees will have gotten better pitching from Rivera + X than a team would from Wilhelm. It's not THAT hard to outpitch Quantrill/Embree.

No, it isn't very hard to outpitch two pitchers we get to hand-pick from the stat sheet, at the conclusion of the season, with the worst ERA+ performances. But in real life that isn't how it works, of course. In real life we have to choose, before the season, which pitchers we will keep, and nobody realistically expected Quantrill or Embree to be anywhere near as bad as they were.

So, yes, in real life under most circumstances, it IS hard to outpitch Quantrill/Embree.

The loss of a roster spot is bad, but is it as bad in a league with a DH where PHing for the pitcher isn't as important?

It isn't quite as bad, but the DH himself is occupying one of the spots, and to gain the platoon advantage, often it's a good idea to pinch-hit for him.

The loss of a position player from the roster basically makes it harder for a team to regularly platoon at any one position, including DH.

Steve your arguments in general not specific to Hoyt and the Fruitbat would be more convincing if you could show:

1) Pitchers are not more effective in the current usage. (they are)
2) Pitchers are not healthier in the current usage. (they are)
3) Getting the platoon advantage on batting side is more beneficial than on the pitching side. (Which seems to be the current though among managers).

You never answer these questions remotely convincingly.

1) Pitchers are not more effective in the current usage. (they are)

Closers are more effective on an ERA+ basis than ace relievers, yes. But that is not the same thing at all as saying that "pitchers" are more effective. Every pitching staff nowdays contains one or two (and sometimes even three) extra pitchers -- you know, the very worst relievers on every staff -- who are almost universally not[/e] effective. Their wretchedness is one of the things that helps to drive up league ERA, and thus make the ERA+ figure of the modern closer look better than it otherwise would.

2) Pitchers are not healthier in the current usage. (they are)

If pitchers are healthier today, the evidence that it is their usage pattern, and not the vast improvements in modern conditioning and sports medicine that are creating that health increase, is simply non-existent. You don't have it, I don't have it, and no one else does either.

3) Getting the platoon advantage on batting side is more beneficial than on the pitching side. (Which seems to be the current though among managers).

It isn't clear that, once gotten, the platoon advantage is better for the offensive team than the defensive. But it is abundantly clear that getting the advantage in the first place is vastly easier to do for the offensive team, for the simple reason that every pitcher brought into a game must face at least one batter, while teams can pinch-hit at will.

Thus even the most extreme LOOGYs never face a proportion of platoon-advantaged PAs that approaches the percentage teams routinely get with platooned left-handed hitters. It isn't close. The data that records this is abundant, and has been for decades.

A team can get far more platoon-advantaged PAs for its batters by carrying an extra bat than it can for its pitchers by carrying an extra arm.

Oh, crap, can somebody fix the italics? Sorry!

Fixed. Now.</I>

I don't know if pitchers are healthier, but it seems apparent that they can be returned to health a lot more readily than was once the case. Rotator cuff and ligament tears once ended careers. Now, far less so.

Thanks, Sean!

I don't know if pitchers are healthier, but it seems apparent that they can be returned to health a lot more readily than was once the case. Rotator cuff and ligament tears once ended careers. Now, far less so.

That's certainly my subjective assessment in watching the difference between pitchers of today and those of 30-40 years ago. They don't get hurt any less often today -- in fact, with the great improvements in diagnostic methods, they very likely appear to get hurt more often today. But they recover from injuries far quicker and better than ever before, as a result of the improved diagnostics, the improved nutrition/conditioning, and definitely the improved surgical procedures.

Steve - You just repeated what you have said for years. I inform you, your answers to these questions aren't compelling or convincing - you reply with the same answers. That doesn't help your case. The issue won't be resolved by who can scream the loudest.

I inform you

No, Matt, you don't. Simply asserting that modern usage patterns increase the health of pitchers, and improve the effectiveness of pitching staffs, with no evidence in support of such assertions, is not "informing."

your answers to these questions aren't compelling or convincing

They're the conclusions I draw from the evidence available. If you want to persuade me that they aren't compelling or convincing, I invite you to present compelling and convincing refutations based on evidence.

The issue won't be resolved by who can scream the loudest.

I'm not screaming. I'm banging a drum, remember? ;-)

No, it isn't very hard to outpitch two pitchers we get to hand-pick from the stat sheet, at the conclusion of the season, with the worst ERA+ performances. But in real life that isn't how it works, of course. In real life we have to choose, before the season, which pitchers we will keep, and nobody realistically expected Quantrill or Embree to be anywhere near as bad as they were.

So, yes, in real life under most circumstances, it IS hard to outpitch Quantrill/Embree.

#p = number of pitchers in 2005
#bp = number of pitchers who had an ERA+ of 59 or lower in 2005



Yes, after-the-fact, it's easy to choose pitchers who will outpitch a 59 ERA+, but there just aren't that many pitchers who pitch THAT badly in the majors. Pick a pitcher at random from among those good enough to pitch in the majors last year and the chances are that he outpitched a 59 ERA+. Heck, toss out 154 pitchers (that's 11 per team), pretend that the rest are the ones who would be getting the extra "Wilhelm innings" -- you STILL have only a 27% chance of picking a 59 ERA+ (or worse) pitcher if you pick one at random.

I'd say Wilhelm had the better career. It's entirely possible he was a more valuable pitcher than Rivera in a number of seasons (we don't know what would happen, for instance, if one were to look more closely at the leverage of the outings pitched by Wilhelm, as compared to those pitched by Rivera. It's also possible that the roster flexibility that Wilhelm gives you through his sheer durabiity is worth the lesser quality of inning that he gives you. It's possible that Wilhelm could have been as dominant as Rivera if he'd only had to pitch as many innings as Rivera or that Rivera would not have been as dominant as Wilhelm if he'd been asked to pitch as much as Wilhelm. And so forth.

But looking at seasonal averages (for what little that's worth), the average Rivera season looks more valuable than the average Wilhelm season because the average Wilhelm season was basically a Rivera season plus 40 innings of pitching with a 59 ERA+. Picking pitchers at random, it's just hard to find ones who are that bad, even if we arbitrarily reduce the set of pitchers to pick from to ones who aren't "the top 11 pitchers" (by ERA+) in an organization. Since it's preferable to have pitchers who are better than an ERA+ of 59, and since picking at random suggests we will find ones who are better than that, I can't think that Wilhelm's additional 40 innings of crappy pitching is better than the lower number of innings that Rivera threw at a much higher quality. Again, without getting into roster shenanigans.

Shoot. I coulda sworn I got the columns right. (*sigh*) My bad.

re: #42... I don't even know what I ment by that series of words that is supposed to be a sentence... I confuse myself so much sometimes.

Since it's preferable to have pitchers who are better than an ERA+ of 59, and since picking at random suggests we will find ones who are better than that, I can't think that Wilhelm's additional 40 innings of crappy pitching is better than the lower number of innings that Rivera threw at a much higher quality.

Good point. (Excellent post!) Understanding your it properly now, I stand corrected.

But, of course:

It's also possible that the roster flexibility that Wilhelm gives you through his sheer durabiity is worth the lesser quality of inning that he gives you. It's possible that Wilhelm could have been as dominant as Rivera if he'd only had to pitch as many innings as Rivera or that Rivera would not have been as dominant as Wilhelm if he'd been asked to pitch as much as Wilhelm. And so forth.

is all true. And the larger issue is that the tradeoff that teams really make isn't between the greatest ace reliever of all time and the greatest closer of all time, but is rather how to make the best use of the bullpen talent available to them year-to-year. And looking at the situation at the macro level, I remain highly skeptical that the gains one generally gets from the shorter stints/tighter role definitions of the modern closer-driven expanded bullpen outweigh the costs.

Understanding your it properly now

Actually your it and I aren't acquainted. But I do understand your point properly now.

Wilhelm would then average 116 inning per season. In order to match him on a seasonal basis, Rivera would need to throw another 43 innings per season with an ERA+ of 59.

This is flat-out incorrect. ERA+ can not be averaged linearly over innings; the denominator is (adjusted) runs allowed. In order to achieve the same ERA+ in Wilhelm's innings, you'd have to add 43 innings at an ERA+ of 102.

Rivera's awesome, but Wilhelm was way more valuable.

Shoot. I coulda sworn I got the columns right. (*sigh*) My bad.

No, it's not your bad. It's a software bug. An inexcusably stupid bug at that.

ERA+ can not be averaged linearly over innings; the denominator is (adjusted) runs allowed. In order to achieve the same ERA+ in Wilhelm's innings, you'd have to add 43 innings at an ERA+ of 102.

You state this authoritatively enough that I don't at all doubt that you're right (and that my numbers are wrong), but would you mind walking me through the math?

Specifically, I'd like to see how you got the 102, when you get a chance (so I'll know better for next time).

To test my wrongness, I just tried to 'predict' Joe Blanton's 2004 from his 2005 and career numbers. By my wrong method, I got an ERA+ of 75 in 2004, so I am totally convinced I'm doing it wrong. (I was as soon as you said something, actually, but it's nice to have the numbers confirm it).

I mean, I suppose I <u>could</u> actually sit down with a paper and pencil and try to work out the right way of doing it...

Ok, I've worked the numbers out (by pushing paper and pencil) with Joe Blanton, so I don't think I need to have it explained for Wilhelm/Rivera.

I will say that it's amazing (well, okay, maybe not amazing) how much rounding the numbers early on can throw off the end numbers (or at least how much it did with Joe Blanton and his short career). For instance, by using his 'official' ERA+, I got his 'predicted' 2004 ERA+ to be 93 (rather than the 'correct' answer of 83). By using ERA+s rounded to 5 decimals (by using his 'official ERA' and 'official lg ERA,' I was able to get down to a predicted 2004 ERA+ of 85. I'm sure if I had a calculator handy with a memory and/or bothered to plug the numbers into a spreadsheet, I could get a lot closer to the real answer of 83, but it's a pain to try to get there without doing either of those things.

Oi, I've now fiddled with several different variations and I've yet to get your 110. What numbers did you use to get 110? (I'm not complaining, just curious -- different choices concerning where and when I round end up yielding very different results -- I've gotten as low as 95 and as high as 119).

Specifically, I'd like to see how you got the 102, when you get a chance (so I'll know better for next time).

Yeah; rather than using ERA+, use it's inverse; this way, innings are the denominator. So Wilhelm's 146 and Rivera's 197 become .685 and .508 respectively. Then do the same thing you did using weighted averages; in this case, I get .98. I then wrote that .98 as a 102 ERA+.

I didn't mean to knock you for this, btw, though I realize my post came out like that. It's just something that annoys me about ERA+, and I shouldn't have sounded annoyed with you.

Yeah; rather than using ERA+, use it's inverse; this way, innings are the denominator. So Wilhelm's 146 and Rivera's 197 become .685 and .508 respectively. Then do the same thing you did using weighted averages; in this case, I get .98. I then wrote that .98 as a 102 ERA+.

Oh man, now I'm really confused how you got the numbers you did. I ended up not using inverses at all. I suppose (again) I could figure it out from what you've said, but at this point, I'd rather just be told.

Some math to follow, feel free to skip past it:

-----

ERA = 9(ER/IP)
lgERA = 9("lgER"/IP)
And ERA+ = 100(lgERA/ERA)

So ERA+ = ...

I'm skipping writing out the math, but I'm pretty sure it's right.

... = 100("lgER"/ER)

And/or "lgER" = .01(ER * ERA+)

(though I end up ignoring the .01 in most of the calculations I do because I'm just as comfortable with an ERA+ of "1.27" as one of 127).

-----

Example (which will also show the danger of rounding):

<u>Joe Blanton</u>:

<u>career</u>: 84 ER and an ERA+ of 125.
That gives him 105 lgER.

<u>2005</u>: 79 ER and an ERA+ of 127.
That gives him 100.33 lgER.

Thus, we can figure out what his ERA+ in 2004 was by subtracting his 2005 lgER from his career lgER and then dividing that by the difference in his ER.

105 - 100.33 = 4.67,
84 - 79 = 5, and...
4.67 / 5 = 0.93

Blanton should have had an ERA+ of 93 in 2004. Ta-da! (Unfortunately, he actually had an ERA+ of 83 in 2004 -- oops!)

-----

The problem is not with my arithmetic (I think) but with rounding errors. I used the ERA+ numbers given on Blanton's bb-ref page, but those ERA+ numbers are rounded.

For instance, if you divide his career lgERA by his career ERA, instead of a nice round 1.25, you actually get 1.246537396121... which bb-ref rounds up (and multiplies by 100) to get an ERA+ of 125. Similarly we'd get 1.2719546742209... which gets multiplied by 100 and rounded down to 127 for his 2005 ERA+. If we use these numbers (rounded to the nearest hundred-thousandth, we get...

<u>career</u>: ER = 84, ERA+ = 1.24654
so, lgER = 104.70936
<u>2005</u>: ER = 79, ERA+ = 1.27195
so, lgER = 100.48405

104.70936 - 100.48405 = 4.22531

And 4.22531 / 5 = .845062

Thus, his 'projected' 2004 ERA+ gets rounded up to 85, which still isn't the 83 we want it to be, but is a <u>heckuva</u> lot closer than 93 was.

It wouldn't surprise me if we used the ERA components (rather than the rounded ERA) and/or knew how to compute lgERA ourselves, we could get the 83 we're looking for. Of course, doing all that is even more a pain in the petushky than all this was (and all this was hardly painless arithmetic).

-----

Which brings me back to an earlier question: what numbers did you use for the Rivera/Wilhelm problem? Could you walk me through your math? I now understand you solved the problem a different way -- I'd love to see how you did it. I'd also like to know if you ignored Wilhelm's last two years or not (and what you used for his career ERA+).

I didn't mean to knock you for this, btw, though I realize my post came out like that. It's just something that annoys me about ERA+, and I shouldn't have sounded annoyed with you.

No worries, I wasn't offended. And I'd rather get the math right than wrong, so thank you for speaking up in the first place. I'll know better for next time.

And the reason I did any of the math in the first place was that I expected to find Wilhelm's quantity to be a lot more valuable than Rivera's quality. When my (wrong) calculations suggested otherwise, I was surprised, but kept my surprise to myself and shared the numbers. I'm just glad you were interested enough in the truth as to not simply accept my (wrong) numbers and to help me get on the right track.

I will stand by my earlier conclusion from the (wrong) numbers, however: if Rivera = Embree/Quantrill = Wilhelm, then I'd rather just have Rivera, I think. Of course, that's not the case. Instead, it's turning out to be more along the lines of Rivera + (40 innings of 2005 Tim Wakefield) = Wilhelm, in which case, I'm pretty sure I'd like to have Wilhelm.

It wouldn't surprise me if we used the ERA components (rather than the rounded ERA) and/or knew how to compute lgERA ourselves, we could get the 83 we're looking for.

No way. There's no way that rounding error could lead to such a huge discrepancy. I'll take a look at your math and go through mine in more detail tomorrow.

While waiting for your reply, Harold, I ended up setting up a spreadsheet (one which won't round unless I tell it to round).

-----

I used ER, IP, and lgERA as found on bb-ref.

ERA = 9 * ER/IP
ERA+ = 100 * lgERA / ERA
"lgER" = .01 * ER * ERA+

Working backwards from his career and 2005 ER, IP, and lgERA numbers (and without rounding), I got Joe Blanton's predicted 2004 ERA+ to be 84.47, which is still closer to his actual 2004 ERA+ of 83 than I got before.

Moreover, tweaking his lgERA numbers can get me down below 83: if I change his career lgERA to 4.495 (which would round up to the same 4.50 which shows up on his page), I get his predicted 2004 ERA+ to be 82.15.

Trial and error (well, ok, I just made one more trial) suggests that a career lgERA of 4.497 (again, which would round up to 4.50) gets Blanton a predicted 2004 ERA+ of 83.08 (which rounds to 83).

Anyway, it seems very likely to me that the discrepancy between my predicted ERA+s and the actual ERA+s earlier in this thread is the result of my using the rounded lgERA numbers provided on bb-ref. (Not that I really have a choice, mind you).

-----

Comments?

Anyway, using Wilhelm and Rivera's career ER, career IP, and career lgERA's (i.e., without dropping the last two years of Wilhelm's career or anything), I determine the difference between Rivera and Wilhelm to be an average of 34 innings pitched per year and an ERA+ of...

wait for it...

120.

In other words, for the average Rivera season to be as valuable as the average Wilhelm season, he'd have to average 34 more innings per year with an ERA+ of 120 in those innings. (That's 34 innings of slightly better than the 2005 version of Randy Johnson).

First reaction: that can't be right, can it? It's too big an ERA+.

Second reaction: no seriously, that can't be right. It's greatly different from what you got. There must be something wrong with my math.

Fiddling with the lgERA numbers (as I did with the Joe Blanton numbers) doesn't yield vastly different results. Which makes sense -- we're dealing with much larger numbers of innings pitched and earned runs allowed here, so small differences in lgERA are going to be dwarfed by the other numbers in the calculation.

Maybe my means of doing the math is wrong. I checked my math by using it to predict Joe Blanton's 2004 ERA+. Maybe there's something unusual about his numbers that's hiding my bad math. Let's check with Scott Proctor, another pitcher with two years of pitching in the majors.

Using (44 + 2/3) as his 2005 IP, (69 + 2/3) as his career IP, (30) and (45) for his ER for 2005 and career respectively, and 4.50 and 4.47 for his lgERA for 2005 and career respectively, I get a predicted 2004 ERA+ of 81.79 (which rounds up to 82). His actual ERA+ was 83.

I dunno. I think I must be doing bad math somewhere here, but the numbers seem pretty close. (On the other hand, I'm not getting the numbers exactly right either -- so maybe that's the indication that I am doing something wrong).

In other words, for the average Rivera season to be as valuable as the average Wilhelm season, he'd have to average 34 more innings per year with an ERA+ of 120 in those innings. (That's 34 innings of slightly better than the 2005 version of Randy Johnson).

First reaction: that can't be right, can it? It's too big an ERA+.

Nah, it's right. Wilhelm was even studlier than I thought! ;-)

I used the technique to predict other two-year players' 2004 ERA+s.

Lenny DiNardo:
Predicted 2004 ERA+ = 115.26... (rounds to 115)
Actual ERA+ = 115

Brian Bruney:
Predicted 2004 ERA+ = 103.27...
Actual ERA+ = 103

Jeff Francis:
Predicted 2004 ERA+ = 97.78...
Actual ERA+ = 98

Gustavo Chacin:
Predicted 2004 ERA+ = 191.14...
Actual ERA+ = 189

Chacin is like Blanton in that he pitched a lot more innings in 2005 than he did in 2004. Chacin is also like Blanton in that his predicted 2004 ERA+ is (at a glance) the farthest "off" from his actual 2004 ERA+.

Anyway, if my technique is wrong, it's managing to get some pretty close results.

And if it's right, well, Steve Treder is right: Wilhelm was even more of a stud than we've heretofore been giving him credit.

Anyway, if my technique is wrong, it's managing to get some pretty close results.

And if it's right, well, Steve Treder is right: Wilhelm was even more of a stud than we've heretofore been giving him credit.

Ok, I think I spotted what I may be doing wrong in looking at the career numbers. Or not.

Let's look at Bob Stanley vs. Mariano Rivera. If we plug in their career numbers, we get the difference to be about 901 innings of 83 ERA+.

But does that mean the average Stanley season is worth the average Rivera season plus another 69 innings of 83 ERA+ per year. Or to know that, do we need to actually plug in the seasonal averages?

Using (roughly) 73.3 IP and 19 ER per season for Rivera (with a lgERA of 4.60 still), and (roughly) 131.3 IP and 53 ER per season for Stanley (with a lg ERA of 4.29), I get the difference to be 69 innings of 74 ERA+.

I used the career numbers before for Wilhelm vs. Rivera. If I use the average seasonal numbers instead, I get that Wilhelm's average season = Rivera's average season plus 34 extra innings at an ERA+ of 58. (Which is obviously a far cry from the 120 of before).

Which is the right approach?

I don't know. The second approach seems right, but now I'm back to a number which seems surprisingly low to me (like when I first crunched the numbers); moreover, my ability to be surprised by the numbers is, at this point, nowhere near being a reasonable test for whether or not the numbers are actually right (or not).

Hmm, too bad I didn't save my spreadsheet from the other day. It looks like your technique is correct if you're getting the results you are in post 62. And I take back what I said in 58, as I didn't realize Blanton had only 8 innings in '04. Yeah, rounding error can be that large when the number of innings is so few; what that means is that any '04 ERA+ within a fairly large range would still yield the same career ERA+. I should have paid more attention to the specifics there.

Let me play with the numbers a bit and see if we're doing the same thing, or what.

Hmm, looks like my way is incorrect as well. I was basically using the inverse of ERA+, which does have the right denominator, but I think it introduces some distortion when combining different run environments. IOW, it might work when the two seasons or two pitchers pitched in identical environments, but not when they didn't. To handle the general case, we need to go the long way like Robert did.

I used the career numbers before for Wilhelm vs. Rivera. If I use the average seasonal numbers instead, I get that Wilhelm's average season = Rivera's average season plus 34 extra innings at an ERA+ of 58. (Which is obviously a far cry from the 120 of before).

Which is the right approach?

I don't think one way is clearly the "right" approach. Taking the seasonal averages and then comparing is one way, and comparing careers is another. When you do the former, you're not giving Wilhelm credit for having a significantly longer career.

Interesting little discussion. I'm not sure that any method gives you value. However, if Wilhelm performed exactly the same comparable to the league as he did in his time in the same era that Rivera pitched, I presume he would have a 3.15 ERA. (that is 4.60/3.15 *100 = 146). If Mariano Rivera were to end the year with a 3.15 ERA and match Wilhelms IP, he would need on average to pitch 34 more innings and give up a total of 37.5 runs or 18.5 more runs during those innings. The means he needs a 4.89 ERA during those 34 innings. If the lg ERA is 4.60 that looks like an ERA+ of 94.

Its rather silly to say "those innings go to the worst pitcher". I don't think they do. Specifically, those innings are pitched by the setup man. The setup man's mop up innings, which have low leverage go to the worst pitchers.

I think if you have a setup man that pitches with a 94 ERA+ or a 4.89 ERA if Vinay doesn't like the ERA+ math, you are looking for another setup man.

if Vinay doesn't like the ERA+ math,

It's not a question of me not "liking" the math. The math flat-out doesn't work if you just use ERA+. You need to convert to innings and runs like BL does in 67.

Its rather silly to say "those innings go to the worst pitcher".

It sure looked like they did in the game that this thread started out with, not that I'm complaining as an A's fan!

More seriously, it depends how you define "those". Over the season as a whole, those innings do go to the worst pitchers, who would not need to be on the roster with the sort of usage patterns Steve is pining for. You are right to say that the higher leverage innings <u>should</u> go to the next best pitchers rather than the worst pitchers. But of course even low-leverage innings have some leverage and many managers do not do a good job of maximising leverage. There are plenty that want to put better pitchers in the game when winning by 5 runs after 6 innings than they do when losing by three at the same stage of the game, for example.

Ok, the way to figure out exactly what we want is to make up two hypothetical pitchers, put them in the same context, and then figure out which questions we want answered (and answer those questions) for them.

-----



(1) For pitcher A's average year to match pitcher B's average year, he would have to have thrown 45 more innings each year and yielded 40 more earned runs in those innings. Give those innings and runs to a third pitcher C...

Pitcher C's average season:
45 IP, 40 ER, 8.00 ERA, (4.00 lgERA) 50 ERA+

Unless it's hard to find pitchers with ERA+s of 50, pitcher A has probably had the better seasons, on average, despite throwing fewer innings per season.

(2) If both A and B retired today, pitcher B's career would equal A's career plus the career of an imaginary pitcher D...

Pitcher D's career:
495 IP, 260 ER, 4.73 ERA, (4.00 lgERA) 85 ERA+

I don't know how hard it is to find pitchers with 85 ERA+s, but it's certainly harder than it is to find pitchers with ERA+s of 50.

Let's look at that another way: if pitcher B were retired, and if pitcher A could throw 495 more innings, he'd need an ERA+ of 85 (or better) in those 495 more innings in order to have a career as good (or better) than B's career. It's doable, of course (and A has certainly been dominant so far) but he'd still have to go out and accomplish it.

-----

What I think I'm figuring out (in my slow way) is that both of the ways of approaching the Wilhelm vs Rivera problem are correct (depending on what question you're asking)...

Wilhelm's average season = Rivera's average season plus 34 innings of a 58 ERA+. Although it's possible to find pitchers worse than that, I'd guess that that makes Rivera's average season more valuable (ignoring roster spots and such).

On the other hand, Wilhelm did that for 21 years. Rivera has only done it for 11. Wilhelm's career = Rivera's career + (the career of a pitcher who throws about 1450 innings with an ERA+ of 120). That's pretty damned good. If Rivera could throw another 1447.2 innings, he'd have to average (or do better than) an ERA+ of 120 over those 1447 innings in order to match Wilhelm's career.

Putting that another way...

Mariano Rivera, so far (~806 IP, 197 ERA+)
plus
Mark Mulder, so far (1208 IP, 116 ERA+)
is slightly less than
Hoyt Wilhelm (~2254 IP, 146 ERA+)

or

Mariano Rivera, so far (~806 IP, 197 ERA+)
plus
Derek Lowe, so far (1312 IP. 120 ERA+)
is slightly less than
Hoyt Wilhelm (~2254 IP, 146 ERA+)

In that context, I'd say Wilhelm's career is pretty damned impressive and that Rivera still has a ways to go to match him (in terms of career value). Of course, if Rivera keeps pitching like he has...

What I think I'm figuring out (in my slow way) is that both of the ways of approaching the Wilhelm vs Rivera problem are correct (depending on what question you're asking)...

I agree with this.

Wilhelm's average season = Rivera's average season plus 34 innings of a 58 ERA+. Although it's possible to find pitchers worse than that, I'd guess that that makes Rivera's average season more valuable (ignoring roster spots and such).

I think there are a lot more confounding factors which make it difficult to answer this. There's usage patterns and leverage, the roster spots do have value, etc. I guess what I'm saying is that two pitchers who allow the same number of runs in the same number of innings can still have greatly different values; it's even harder to compare Rivera and Wilhelm, whose numbers are different but arguably close.

Wilhelm's average season = Rivera's average season plus 34 innings of a 58 ERA+. Although it's possible to find pitchers worse than that, I'd guess that that makes Rivera's average season more valuable (ignoring roster spots and such).

So here's a neat comparison (ok, it's neat to me, perhaps not to anyone else):



Of course, I don't know what that means about whose average season is better. I'd probably want Rivera first out of the three, but Stanley's average season vs. Wilhelm's average season depends on how hard it is to fill the extra Wilhelm-less innings (that Stanley would give you if you had him) with pitching at 81 ERA+ (or better).

I think there are a lot more confounding factors which make it difficult to answer this. There's usage patterns and leverage, the roster spots do have value, etc. I guess what I'm saying is that two pitchers who allow the same number of runs in the same number of innings can still have greatly different values; it's even harder to compare Rivera and Wilhelm, whose numbers are different but arguably close.

Totally agree -- I was specifically ignoring context and roster management and salary and all that. If you wanna pore through Wilhelm's box scores to figure out his leverage (in order to compare it to Rivera's), by all means, be my guest... ;)

Rivera's the best closer ever.

Rivera's the best postseason reliever ever.

Wilhelm's the best overall reliever ever.

And Wilhelm won a Purple Heart in the Battle of the Bulge, eight years before he first came up with the Giants.

I hate to admit it, but Advantage, Wilhelm.

Here are two (possibly related) things that I never see mentioned in these cross-era closer vs relief ace arguments:

1) In 1967, Wilhelm's best ERA+ season as a pure reliever, he appeared in 49 games and threw 89 innings. He pitched on consecutive days 5 times, and both games of a double header twice. Rivera's best ERA+ season was 2005, which Retrosheet doesn't have up yet, so I randomly selected 1999 as a representative Rivera season. He made 66 apearances and threw 69 innings. He pitched on consecutive days 14 times, and pitched on the third day five of those times. (Note: I'd love to look at Wilhelm's rookie year, when he pitched 71 games and 159 innings, but Retrosheet doesn't have dailies for 1952). Point: The certainty of having your ace available almost every day is obviously worth a lot to modern managers, and they are willing to trade some leverage and some innings for that. Rather than dismissing current bullpen management as slavish devotion to the save stat, wouldn't it make more sense to try to find the balancing point between maximizing IP + leverage on the one hand and maximizing availability on the other?

2) A huge part of the reason that managers crave the security alluded to in (1) is that nobody ever throws a comlete game any more. When Leo Durocher was running Wilhelm out there every 71 times in 1952, he was also getting 49 complete games in a 154 game season. Every time he called the pen, he knew there was about a 1 in 3 chance that he wouldn't have to pick up that phone again the next day. I'm not saying that modern closer usage is anything close to optimal, but I am saying that I think it is dictated to a very large extent by modern starter usage.

Hmmm... looking at other relief pitchers' numbers...



which means, career-wise...

Stanley = Sutter + 665 innings of a 102 ERA+
Stanley = Fingers + 6 innings of a 112 ERA+
and Gossage is simply better than Stanley ;)



...which means, average season-wise...

Stanley = Sutter + 46 innings of 99 ERA+
Stanley = Gossage + 49 innings of 110 ERA+
Stanley = Fingers + 31 innings of 116 ERA+

So what does this mean? Given the above information, does Stanley still rank 4th-best among Sutter, Gossage, and Fingers? How would you rank these four pitchers?

: The certainty of having your ace available almost every day is obviously worth a lot to modern managers, and they are willing to trade some leverage and some innings for that.

Except that leverage is maximized in the current reliever deployment model.

That is what is astounding by the person(s) that keeps preaching "ace reliever".

LI is better using the current method.

Availability is better in the current method.

Health is better in the current method.

Roster maximization is better in the current method.

Bullpen effectiveness is better.

There is no tradeoff whatsoever. Everything measurable is better than before. We have had countless long threads that show each of these elements. We have shown the increase in career length. We have shown the improvement in ERA+ for the top relievers. We have shown the longer careers. We have shown the leverage of the 25th spot of a LOOGY compared to a third catcher. We have shown the LI combined of the relievers. We have shown how diversication leads to better bullpen effectiveness than concentration. We have shown the bullpen models are more repeatable than ACE RELIEVER and platoon.There is no argument left for the ACE RELIEVER crowd.

Instead, they will first just make a list of players that you have never heard of like Dick Goatshit, Billy Grabinass, or Zach Lacy. Then there supposed killer argument is listing out other teams. That was hysterical. After researching the issue for years, they came up with about six teams over forty years. Systematically we showed were random teams were outperforming them using current methods. The only team left standing was ONE, AND ONLY ONE, year of the Orioles. And after that, the team went to hell in a handbasket, and a bunch of Hall of Famers and promising youngsters had their arms shredded.

But instead of dealing with that, or even attempting to bring new evidence, they just keep finding a thread where they can just post the same erroneous conclusions over and over. If anyone calls them on it, they will just give you a

"Well, you don't know and I don't know...and Goose Goosage...and that is my conclusion."

Yep, that's the point. It falls in the same bucket of conclusions as:

"People's liability for getting hit with chairs should be diminished because of who they marry.";

"Roid em up and test the survivors."

"We should increase the number of innings teenagers throw by 10%"

or inferences that are not supportable, valid in only some solipstic sense, and generally daffy.

"People's liability for getting hit with chairs should be diminished because of who they marry."

Well of course it should.

"Roid em up and test the survivors."

Pass.

"We should increase the number of innings teenagers throw by 10%"

Umm... no.

Well, you don't know and I don't know...and Goose Goosage...and that is my conclusion."

You might know and I might know... but Goose Gossage rocks.

There is no tradeoff whatsoever. Everything measurable is better than before. We have had countless long threads that show each of these elements. We have shown the increase in career length. We have shown the improvement in ERA+ for the top relievers. We have shown the longer careers. We have shown the leverage of the 25th spot of a LOOGY compared to a third catcher. We have shown the LI combined of the relievers. We have shown how diversication leads to better bullpen effectiveness than concentration. We have shown the bullpen models are more repeatable than ACE RELIEVER and platoon.There is no argument left for the ACE RELIEVER crowd.

Could you point me to these threads? Post a link or something? I've not seen them. I can totally believe that all of those things have been shown, but would prefer to see the arguments being presented (with the supporting evidence) rather than merely take your word for it.

Could you point me to these threads? Post a link or something? I've not seen them. I can totally believe that all of those things have been shown, but would prefer to see the arguments being presented (with the supporting evidence) rather than merely take your word for it.

That's a wise approach, Robert, but I wouldn't hold my breath waiting if I were you. BL delights in the theatrics of internet battle, but demonstrates no mettle for genuine evidence-oriented, content-focused debate and civil interaction. He'll throw up clouds of invective, and proclaim himself a persecuted victim, but it's all show and no dough. If he does link to any threads, they will be ones like this, in which he presents unsubstantiated, unresearched opinons and passes them off as "findings" that "we have shown." You might find it entertaining, but don't expect to find it substantive.

These issues are interesting, certainly debatable, and by no means resolved in any direction. If you're unfamiliar with the research I've performed and the articles I've written on the subject, I'd be happy to share them with you. Email me at streder@hardballtimes.com if you're interested.

I wouldn't hold my breath waiting if I were you. BL delights in the theatrics of internet battle, but demonstrates no mettle for genuine evidence-oriented, content-focused debate and civil interaction. He'll throw up clouds of invective, and proclaim himself a persecuted victim, but it's all show and no dough. If he does link to any threads, they will be ones like this, in which he presents unsubstantiated, unresearched opinons and passes them off as "findings" that "we have shown." You might find it entertaining, but don't expect to find it substantive.

Although I really do appreciate how I believe you meant these words (in terms of your regard for helping me avoid tilting at windmills), I worry that if Backlasher is as you say he is, your words will provide an ample jumping-off point to ignore my query and instead deal with whether or not you have accurately represented his character. I'd rather just hear the evidence (or be directed to it) and form my own opinions on his evidence (or possibly lack thereof), rather than have to wade through a dozen back-and-forth posts about whether or not he's won his previous arguments with you.

I don't mean that in a bad way -- again, I appreciate what you're trying to do -- but I I'd rather stick to the actual evidence and arguments, and keep the character issues separate for as long as it is possible to ignore them.

So, back to you, Backlasher... which threads/studies can you point me to?

I worry that if Backlasher is as you say he is, your words will provide an ample jumping-off point to ignore my query and instead deal with whether or not you have accurately represented his character.

Your worry is well-founded, because that is of course his extremely predictable behavior. Here's hoping he proves us both wrong.

Here is one thread and one of the articles Steve wrote.

Could you point me to these threads? Post a link or something? I've not seen them.

Click on Search, and type in keywords like "pitcher utilization". or try Emeigh who has really torn that posturing apart.

You can also start with one in 83.

This one will link a few other studies and threads.

You can also check Tango's web site for the LI studies.

I worry that if Backlasher is as you say he is, your words will provide an ample jumping-off point to ignore my query and instead deal with whether or not you have accurately represented his character.

You can judge for yourself, who ignores whom and who cuts and run. Its not even an issue that should require you to see data.

And the thought that I am interested in winning an internet debate with the likes of Treder is gut-busting funny. There is no debate; there isn't even a contest of issue. There is just a loon screaming everytime the subject comes up and young impressionable people listening to him.

But by all means ask him for his "research". That is even funnier. His research is getting somebody else to do some arithmetic operations on numbers, which will conclusively show point one, relievers are more effective per outing. He himself couldn't divide a slice of pie, as evidenced by this thread when he attaches one of his "well said's" to Vinay on a calculation that Vinay later admits was erroneous. I was half expecting a "Well you don't know and I don't know that the calculation is wrong, that is just one opinion."

BL delights in the theatrics of internet battle, but demonstrates no mettle for genuine evidence-oriented, content-focused debate and civil interaction.

ROFLMAO. Peaches, I'm happy to debate, but that takes someone on the other side. You have no training in this endevor, no experience in this endevor, no education in this endevor, and no capacity in this endevor. And this endevor spans the ability of discourse itself, its underpinnings of logic, or its underpinnings in mathematics.

You still haven't said anything that isn't a parrot of what Bill James said in the 1980s and has since backed away from. I guess that's your modus operendi. You'll probably start claiming the 102 ERA+ on the Wilhelm issue, which you immediately endorsed even though it is erroneous, retracted by the author, and you couldn't even begin to figure out how to do a calculation yourself. Heck, I'm not sure if you could calculate ERA, much less an ERA+ delta.

But that is what you do. You read a book, proclaim yourself an expert, then surround those thoughts with Hank Kimballesque sophistry and solipism.

You want to talk about a debate, that's hilarious. You never present any evidence. Its usually an authorative declaration that's wrong, like Vida Blue's name change, or Conte's role in Tower of Power. Your evidence is usually making a list of players, like Lee Lacy, Champ Summers, etc. which gets eviserated after everyone stops laughing.

I recall you using somebody else's numbers and Tango's Pitch Count Estimator to show that pitchers throw less now. Guess what, nobody disputes that. The other thing you did was use somebody else's numbers to show THAT CLOSERS HAVE HIGHER ERA+ THAN YOUR PRECIOUS ACE RELIEVERS.

I would have thought Rauseo giving you a spanking for raving and yelling would have slowed you down. I would have thought that the new poster who showed how comical the "Well, you don't know and I don't know" reductionism into solipism would have slowed you down.

Instead, you want to go big game hunting. Well go ahead, if you really got something behind your bluster stop that highly comical, "I'm ignoring backlasher" and lay your cards on the table. See if you can engage in discourse without the "anybody who knows anything about..." and solipism. See if you can lay on the table some real evidence.

Because everyone knows the game, when data is put on the table, you run, run, run. Rauseo already pointed that out when you once again rolled out your silly banter.

another

What I found interested in the googling is an old sawbone of Treders that advocated

2 Starters

5 Swingmen and

3 relievers

as the optimum staff. One of the earlier quoted threads will show that the "swingman" was the role where pitchers performed the worse.

And as for the threads and data on that, you should be able to run the queries yourself to see that reliever to starter era has gotten better in modern utilization and its the "swingmen" that really stink up the joint.

Here's hoping he proves us both wrong.

Oh well. At least we could hope! :-)

Oh well. At least we could hope! :-)

Hope is all you got ;):):::)))))))

You have no evidence, no ability, and you don't even have the legion of "Please give me a well said Uncle Steve" any more.

You can just hope nobody calls you on your silliness.

Your worry is well-founded, because that is of course his extremely predictable behavior.

LOL. That of course should be an interesting read for some Primates. Remember, Treder claims he "ignores me" and doesn't read my posts. Obviously, this is affirmative proof that he is a liar on this front. His "ignoring" is just a rhetorical device.

And it also shows you another favorite trick. Any time he sees a delay in response, he'll affirmatively claim that other people don't respond. He's done this multiple times, only to see people come back and post right away.

But this thread is classic Treder. He throws out "my subjective take... bar bar bar." then goes "If you want to persuade me that they aren't compelling or convincing, I invite you to present compelling and convincing refutations based on evidence." Like any thing could rise to the level of compelling and convincing to someone that thinks Lee Lacy is proof that Bonds didn't take steroids.

But this is the best Nah, it's right. Wilhelm was even studlier than I thought! ;-) . I guess the guy doesn't get embarrassed no matter how often he royaly screws up. Stevie, did you do those calculations yourself?

Except that leverage is maximized in the current reliever deployment model.

This is a ridiculous statement, especially in a thread originally about the Yankees' reliever deployment on Tuesday. You may or may not be right that current deployment is better than the 70s relief ace situation, but many teams treat a three run lead in the ninth as being higher leverage than a tie game in the eighth, which goes clearly against Tango's work that you reference.

You can also check Tango's web site for the LI studies.

This is excellent advice for anyone trying to get beyond th bluster in this thread. For anyone seeking enlightenment, click here.

Tango certainly doesn't argue that current relief patterns are better than the old. Comparing Sutter to the top relivers in 2002 he shows:



We have a list, if not the top 10 relievers in the league, arguably 10 of the 20 top relievers in the league. How did their managers use them?

Here are their LI, over the same time period. For a frame of reference, Bruce Sutter's career LI is 1.90, while an average LI would be 1.00. Bob Stanley's career LI is 1.30.


FirstLastLI
RobbNen1.85
TrevorHoffman1.85
MarianoRivera1.72
BillyWagner1.68
UguethUrbina1.56
KeithFoulke1.30
SteveKarsay1.29
PaulShuey1.29
OctavioDotel1.28
MikeStanton1.08

His overall conclusion is that Some managers are effectively using their best relievers, and some are pitching them in more secondary roles. Hardly a ringing endorsement for current practice.

I think current bullpen management does have some advantages. Psychologically, it is almost certainly helpful for relivers to know when they are going to be used. But there are big problems with distortions created by the hold and save stats. Too many managers will use better pitchers to prtect a three run lead than to try to overturn a one run deficit or even keep them in a tied game. And there is too much fear of having your best relievers go two innings in critical situations. Lets not pretend the current situation is perfect.

I think current bullpen management does have some advantages. Psychologically, it is almost certainly helpful for relivers to know when they are going to be used. But there are big problems with distortions created by the hold and save stats. Too many managers will use better pitchers to prtect a three run lead than to try to overturn a one run deficit or even keep them in a tied game. And there is too much fear of having your best relievers go two innings in critical situations. Lets not pretend the current situation is perfect.

Just about precisely sums up my view, too, Adam. I guess that makes it "well said!"

This is a ridiculous statement, especially in a thread originally about the Yankees' reliever deployment on Tuesday.

No its not a ridiculous statement, and it is in fact a correct statement. You might want to check out some of the links that were actually posted. LI is higher in the current model. Because for every career of Sutter and Gossage, you have a Percival that is even better over a career, and for every Foulke year, you have a Tekulve year. I'm not sure what happened to the Tango study that actually showed the LI bands by years. I think the archive beast got it, but if Daly is around, he might be able to find it. But hey, if you want to take one article that deals with Foulke and not look at the eras and pretend that LI is not higher through some inference chain, knock yourself out. That type of analysis is what gave us DIPS, and its repeatedly used by Treder.

Lets not pretend the current situation is perfect.

Let's not put words into other people's mouths. You have the link of the thread where this was discussed in its greatest length.

Just about precisely sums up my view, too,

People can read the links to see about that too. Its not that hard and its not that long of a thread. You can also see the disappearing act of Treder and people demanding he come forward with evidence and address the other arguments.

You can see the "Throw 10% more when they are teenagers." The "2 starters and 5 swingmen"; and all the other daffy rhetoric that lead to Rauseo giving him a spanking earlier in the thread.

Haven't found the study, but here are some blast from the past quotes that will tell you a few things about LI.

"I'd say the 1.9 for Sutter is interesting, because that is about what Bill James guessed a typical closer's value was, in his relief study in the NHBA." - Objective Joe Domino

"Craig: Mark Eichorn, in 1986, was 1.32. If you look at the list of 20 players I listed in my followups, you will note that Bob Stanley is 1.30. I think this is probably what you'd find with your multi-inning non-closing firemen." - Tangotiger

"The impact of an 80-inning reliever is no more than that of a 160-inning starter. And that's how we should view them" - Tangotiger

People can read the links to see about that too.

They can. And they can also read what I've written on this subject, in this article:

Pitchers get hurt a lot; they always have, and 15 years into the era of significantly reduced workloads, they still do. If I were a major league GM, I would work on instituting a conditioning and pitcher-use program throughout my organization that would strive to develop starting pitchers capable of throwing at least 10% more pitches per season than the modern norm. I'm confident that in the long run such a program would provide a significant competitive advantage, without producing greater injury rates than are occurring now.

Please understand that I'm not saying that there is no place for pitch counting in monitoring and handling pitchers, nor am I saying that pitch count limits aren't appropriate for young pitchers (and of course for amateur pitchers). I'm saying, as are James and Malcolm, that there's a reasonable deployment of the tool, and there's an unreasonable, counterproductive fixation upon it, and over the past decade and a half we've left the former behind and driven ourselves right into the latter.

And this article:

Questioning the wisdom of how pitchers are being deployed in the modern era should absolutely not be construed as an assertion that pitchers, most particularly young pitchers, aren't very susceptible to injury, and shouldn't be handled with care.

And this article:

Every previous model of pitching staff/bullpen deployment offered strengths, but weaknesses as well. Sooner or later, the weaknesses came to be seen as intolerable by one or more teams. Someone eventually had success with a revised model, competitive dynamics stimulated others to follow, and soon a new paradigm had taken hold. Then, sooner or later, the weaknesses of the new paradigm would become apparent.

I see no reason to expect anything different regarding the current situation.

And this article:

So, even though their usage pattern provides them with the best possible conditions -- high proportion of platoon advantage batters faced, and extremely short stints, necessitating no pacing and exacting no in-game fatigue -- it would be unrealistic of us to expect LOOGYs to be really shutting their opponents down. Nonetheless the bottom line remains: LOOGYs don't really shut their opponents down. Even the very best LOOGYs, the elite deployed in the most extreme hard-core pattern -- the Myers, Plesac, Orosco class -- don't produce rate stats that compare with those of the elite pitchers deployed in the more challenging roles of Closing or Starting. And, especially as the number of pitchers deployed as LOOGYs mushrooms, the typical LOOGY is less and less accomplished, resembling an elite pitcher less and less. The LOOGY is a class of pitcher deployed in a very high proportion of high-leverage situations, but without a particularly impressive record of effectiveness.

And this article:

History strongly indicates that, sooner or later, the only constant is change. There are numerous reasons to expect that this eternal pattern will reveal itself here again.

To assert this is not to say that what will likely happen is a return to the precise Ace Reliever pattern, or that that’s what should happen. Far more probable, and far more reasonable, is the innovative development of a new pattern of bullpen deployment, something we haven’t quite seen yet. As such, we can’t describe it today in detail. But to say that a new development of some kind won’t likely occur is to make a statement that would have been proven wrong at every previous point in history.

No its not a ridiculous statement, and it is in fact a correct statement. You might want to check out some of the links that were actually posted. LI is higher in the current model.

I'm looking at Tango's post #106 in the Fog thread, where he says in response to a similar claim by you that:

Actually, what I've shown is that the LI of the relievers of today and of yesteryear are "similar" (more below). There is still room for improvement, in both cases.

The biggest difference is the number of innings per game. Sutter and Gossage were brought in alot in the 8th innings. The chart referenced by Mikael in post #87 is very illustrative of the issue.

What you really want to maximize is IP x LI.

QUOTING BACKLASHER "LI has been better in the closer model than all other models."

I don't know about better, but they are pretty close. Bruce Sutter's career LI is 1.90, Gossage and Lee Smith at 1.76, and that's pretty much in line with Trevor Hoffman, Percial, Rivera et al.

What's happened is that managers give more of the 3-run leads to the closers of today, and they gave more 7th and 8th inning games to the firemen of yesteryear, and the firemen of yesteryear did not get all the high-leverage 9th inning games.

The conclusion that I draw from all of this is that neither the 70s model nor the current model is perfect. There are clear ways to improve the leverage of your top reliever. Expecting him to be able to go two innings with a one run lead or tie game in the eighth is the most obvious.

Current usage is clearly not maximising levarge when Scott Proctor pitches over a rested Mariano in the 9th inning of a tie game.

An excellent look at the leverage issue is this article by Studes.

His conclusion:

A closer is two-and-a-half times more likely to be brought into the ninth with a three-run lead (75% of the time) than with the score tied (30% of the time). Excuse my bold formatting, but this makes no sense at all!

Three-run leads are gimme situations; fans are heading to the exits. On the other hand, tie games in the ninth are the epitome of crucial situations. Yet most managers would rather use their closer with a three-run lead. What gives?

The issue isn't that what should occur is a reversion to any previous model. The issue is to comprehend the weaknesses of the current model (and comparing its strengths and weaknesses to those of previous models can be useful for that purpose), and thus imagine the forces that will drive its inevitable evolution to the next model, and thus also imagine how that model might look.

At this rate, Rivera will never pitch again!

Here is another good one. In it, there is even the prediction that Steve will again change his position and claim its what he always said.

The conclusion that I draw from all of this is that neither the 70s model nor the current model is perfect. There are clear ways to improve the leverage of your top reliever. Expecting him to be able to go two innings with a one run lead or tie game in the eighth is the most obvious.

Current usage is clearly not maximising levarge when Scott Proctor pitches over a rested Mariano in the 9th inning of a tie game.

Then read the threads and you can determine what everyone's position is rather than assigning a position.

The issue isn't that what should occur is a reversion to any previous model. The issue is to comprehend the weaknesses of the current model (and comparing its strengths and weaknesses to those of previous models can be useful for that purpose), and thus imagine the forces that will drive its inevitable evolution to the next model, and thus also imagine how that model might look.

I think your "slavish devotion" to the Bill James simulator is on record.

But let's see if we got this straight, you no longer think that ACE RELIEVER is the model to aspire too; you no longer think that little Billy Wagner should pitch 120 innings; you no longer think that kids should throw every day; you no longer think teams should carry 10 men staffs with 2 pitchers pitching on 3 days rest and 5 SWINGMEN.

Everyone can certainly read where you have said these things from the links alone. So let's establish a few things. Rather than just spout "we shall overcome" and then expect magic to happen, and pretend that is what your BAR BAR has been all the time --- why not go on record and tell us what the Treder model is.

Or maybe we can be treated to more of your diatribes because Bobby Cox uses Ray King over John Smoltz when John Smoltz was injured.

Why not quit talking around the point like a Dilbert caricature of an executive. Why don't you tell us what it looks like.

I've been on record. In the linked thread, I even declare the optimum decision paradigm, which Adam isn't reading or he would know my take on the use of a healthy Proctor over a healthy Rivera.

But trevise isn't going to email you any numbers for this one. Vinay albeit inadvertently already exposed your inability to do simple math.

Quit pretending like you have done "years of reasearch on the injuries of pitchers" and shout out pseudo-executive speak. Give us a model, or reaffirm your 2 starter (w 3 days rest); 5 swingment; and 3 relievers plan.

Quit bemoaning pitch counts on one hand and then try to say you advocate "monitoring" young pitchers on the other.

You aren't going to play an "Expert" card, and you aren't going to convince anyone of years of research, which amounts to one USA Today article.

As before and as shown by Rauseo, you have offered absolutely nothing new to this discussion. You haven't addressed existing points.

Do you really want to stand on your prior record on this point? If not, better throw out a bunch of well saids and send a few emails to see if someone can do some research for you. You aren't just going to be able to filibuster until someone else joins the fray and "well said" their efforts then co-opt both sides of the argument. I'm interested in this, do you still want to stand that a pitcher would have to throw 34 Innings at a 120 ERA+ and give you Rivera's output to match Wilhelm's output? And goodness, I'd love to see you to go to LI on that one. If you remember, Torre was once lauded for the LI of Rivera in comparision with whippin-boy Kenny Williams and his use of Foulke.

Isn't the answer to this entire thing to just have good pitchers?

If you had LOTS of good bullpen pitchers instead of one guy you had to go to over and over again...

he'd be less likely be injured.

That being said, I'm a heavy advocate of the following:
1) Pitchers not pitching 2 days in a row, unless they're a LOOGY
2) IF somehow pitchers pitched 2 days in a row no matter the circumstance they don't come out to pitch the 3rd day.
3) No injured pitcher pitches. EVER.

I generally buy the "ace reliever" theory. I'd much rather my best reliever pitch out of jams, rather than just get 3 outs with nobody on base, even in a 1/2 run situation. Even BL you'd agree that a situation with runners on base is more "dire" if you will than a situation with no inherited runners.

I've read this thing around here advocating a starting pitcher come out for short relief 3 days after his last start. Supposedly it's his day to "practice throw" anyway, so why not use a really good starter to pitch one/two hitters to get out of a jam? Hey man, Masaichi Kaneda, 400 IP's a year! Relief AND Starting!

I was really tweaking with the idea of having "teams" of bullpen pitchers. Especially for short tournaments (like the WBC) where you could have a deep pool of pitchers to choose from.

The staff would consists of
1) 5 starters
2) One really krappy pitcher for strictly mopup duties.
3) 4 Good relievers split up into two man "teams"
4) One LOOGY

Hypothetically... since it's the WBC, I can pick anybody I want off the USA roster, I pick say Nathan, Smoltz, Wagner, Shields, and Ryan as the LOOGY. Hey, because Americans love America.

Anyhoo...
Team 1 would be Shields and Wagner. Those guys pitch one day
Team 2 would be Nathan and Smoltz. Those guys pitch the next day.

And they'd rotate by day. The first pitcher on that list would be the "Ace Reliever", pitching one or two innings, getting starters out of jams, etc, UNTIL the 9th inning. Which is when the second pitcher on that "team" comes in and does the regular "Closer", but maybe I'd use him to get 4 outs instead of 3, since he's not throwing the next day anyway. The LOOGY can be used anytime "between" the starter and Ace reliever or Ace reliever and closer to "bridge" pitchersi f you will, and get a big out in a big situation. I chose Shields and nathan to be the "ace relivers" on their "teams" because they can go more than one inning. Wagner and Smoltz are more traditional "3 out" guys, but I'm relatively sure they can get 4 outs.

Oh this is a total hypothetical situation. This "Could" have been tried in the WBC, but the US never had a lead.

Why doesn't Treder simply lay out his beliefs on optimal bullpen usage, BL respond like a damn adult and maybe an interesting discussion can ensue?

Why doesn't Treder simply lay out his beliefs on optimal bullpen usage

If my sarcasm detector is under the weather, then my bad, but are you serious with this question? I believe I've laid out my beliefs on optimal bullpen usage 94 million times, to the point that sane people are sick of hearing about them. If you have a specific question about what I see as how the prevailing norm could be improved upon, please feel free to ask.