THE BOOK--Playing The Percentages In Baseball

http://www.baseball-reference.com/pi/gl.cgi?n1=cordefr01&t=p

Cordero has pitched in 57 games this year.  Twice he came into the game in the 8th inning.  Twice in the 10th inning.  All other games, he came on in the 9th.

One 8th inning game: entered bottom of the 8th, 2 outs, ahead by 1, man on 1B.  (LI=2.5) That was May 17.  His previous appearance was May 9.

In the other 8th inning game: entered top of the 8th, 2 outs, ahead by 2, men on the corners.  (LI=2.8) That was Jun 5.  Last pitched Jun 3, and off-day on Jun 7.

Either Cordero is physically impaired to pitch more than 1 inning (injured or whatnot), or Yost is really averse to using him in the 8th.  Since the typical high-end LI for a closer is at 2.3, any situation above that is a definite closer scenario (warmup-time notwithstanding, which is definitely an important consideration).

I’m one of the aforementioned fans who has been calling for Yost’s head (for a lot longer than this season, but I digress).

Yost may be a smart guy, and he may even have bought a copy of The Book, but he has some strange quirks that hurt the Brewers probability of winning.  For one, he is a strong believer in roles.  He decided early in the season that Turnbow would pitch the 8th inning and Cordero would pitch the 9th, effectively limiting the game to 7 innings for the other team (when both pitchers are on their game).  And now he does not deviate from those roles, to the point where local media and sportscasters repeat the mantra that “it’s Turnbow in the 8th and Cordero in the 9th.” Only in hold and save situations, though.  If the Brewers are up by 4 or more, Yost will go to a lesser reliever.

It’s true that many people are piling on the “Fire Yost” bandwagon, and many have bad reasons for it.  That doesn’t mean he isn’t a bad manager.  He’s a bad manager because he uses a lot of small sample size hitter v. pitcher matchup history in his gameplanning, and because he allows roles to become stratified as he ignores actual game situations.  There is also a fair amount of evidence that he plays favorites in the clubhouse, for example, allowing Chris Capuano to struggle for 3 months before removing him from the rotation, but shipping Rickie Weeks off to the minors when he struggled due to a recovering wrist injury.  And now we find Gabe Gross taking at-bats from Bill Hall, as Hall struggles through a tough season, in which he transitioned from the infield to center field.

I respect your opinion, and I suppose I’ll give Ned the benefit of the doubt a little more often, but I think it’s high time legitimate concerns about his management are given their due.

Re: #2

I’m not speaking from a knowledgable position here, but the Capuano vs. Weeks situations (based upon your description) don’t seem inconsistent to me.  If Weeks is struggling due to an injury, it makes sense to send him down until he’s healed.  If Capuano is struggling without an obvious cause, Yost could defensibly rely on Capuano’s history of success for a while.  Reacting to a downturn caused by an injury differently from a downturn caused by simple poor performance is appropriate in my opinion.

Jordan,

I should have been clearer.  Weeks was recovering this year from a wrist injury that caused periodic discomfort (for a day or so), but was structurally healed and healthy.  The issue was caused by some scar tissue that is expected to break away over the course of a year following the surgery.

I can’t say for certain that Rickie’s wrist wasn’t better or worse than described by the team.  I can say that he was sent down to the minors, where he played every day, and so it certainly didn’t appear to related to helping him heal.  Capuano, on the other hand, was given many, many opportunities to work himself out of whatever problems he’s having, and even then was moved to the bullpen instead of sent to work it out in the minors (which may not even be an option, though).

I’m not saying the approach was wrong, but it is one of the examples where Ned appears to favor certain players.  Maybe it’s a lot of nothing, but I follow the Brewers very closely, and while I see a lot of players comment on the close clubhouse atmosphere, most seem to go out of their way to avoid saying anything about Yost, positive or negative, especially in the wake of the big dugout brouhaha between Graffanino and Yost following some nastiness from Yost to JJ Hardy.

What about using Seth McClung, a dreadful pitcher (lifetime ERA of 6.27, much of it as a starter) to start the seventh with a three run lead. That’s not a super high leverage situation (LI = 1.0), but McClung was clearly the worst guy available out of the ten pitchers in the bullpen and was not the optimal choice there.

Marty, don’t get me wrong.  I know very little about Yost.  I don’t know whether he is a good or bad manager.  (And knowing whether a manager is good or bad is about as inexact a science as there is in baseball “evaluations.") I have heard that he bought a couple of copies of The Book, and he sounds intelligent, as managers go, but other than that, I don’t know much about him.  I usually watch a game or two almost every day, so I am sure I have watched the Brewers a fair number of times. Other than this game, I don’t recall anything in particular about his managing style, good or bad.

Using pitchers in “innings roles” is about as bad as it gets in terms of managing your bullpen. It should all be about leverage of course.  Your best relievers, given the situation (L/R mostly) should be used when the leverage is high and yuor worst relievers should be used when the leverage is low, as much as you can (you have other considerations of course).

From what you and others say, and Tango’s look-up on B-R, he is very much of a bullpen “role” guy which is very bad, as I just said.

And using ANY batter/pitcher matchups for decision making, other than as literal tie-breakers, is BAD, as we show in the book.

McClung does not belong in the major leagues.  He is a very bad pitcher and should ONLY be used in very low leverage situations - basically blowouts.  While using him in the 7th with a 3- run lead is probably not terrible, it is probably not so good either.

The bottom line is that all managers can use in-game tactics and bullpen managment a lot better than they do, regardless of what the fans think.  And I pay little attention to what the fans think of manager strategy.  I mean, let’s face it, with all due respect to the baseball fans that frequent this site, the average fan has NO idea what a good or bad sabermetric strategy is.  An example is the so-called knowledgeable fans wanting to leave Shouse in there to pitch to Pence because he has pitched well, by some “garbage standard,” against RHB this year and lately.

At the risk of stepping way out of my depth, Shouse hasn’t been used as a LOOGY in about 4 months.  He was the pitcher of the month in August for the team(not a tough award to win, I think I was tied for fourth in the voting) and had gone a full inning the nite before facing both righties and lefties.  Based on recent performance the first option should have been leave well enough alone and let Shouse get out of it.  The second option would have been Cordero.  At that point in the game there were no other options.

Leaving Shouse in there to face a RHB would have been a terrible alernative.  A pitcher’s future likely mean performance is based on his long-term history regressed toward the mean of all pitchers like him.  It should be obvious from watching Shouse that he is a left-handed specialist.  AND, as I said, his lifetime numbers indicate a large platoon split, as you would expect from a pitcher like him.  So, REGARDLESS of how he has been pitching or how he has been used lately (and “lately” for a reliever is like what, 10 innings, which means VIRTUALLY NOTHING), he is a left handed-specialist and should NEVER face a RHB if a manager can help it.

Joe, with all due respect, and I am glad that you are here, this is a sabermetric site which is dedicated to telling the truth about baseball and exposing some of the myths and “conventional truths” which are misleading or just plain wrong (and widely held by fans, comentators, pundits, and baseball insiders).

One of the more significant ones is that how a player has been doing lately matters much.  It doesn’t.  All decisions need to be made based on an estimate of current “talent” (including platoon splits, G/F ratios, etc.).  Estimates of current talent are a player’s long term performance regressed appropriately (some things get regressed a lot and some things get regressed a little, depending on how much luck and how much skill is involved, which can be figured out statistically, and all things get regressed less and less the larger the sample of the historical data) toward the mean and adjusted for any other RELEVANT facts, such as injuries, etc.  If those concepts are a little difficult to wrap your arms around, I encourage you to read carefully the “statistical” portions of The Book, where much of it is explained, albeit in a somewhat difficult manner.

BTW, can anyone point me to a good Brewers blog or web site where I (or anyone) can create a comment or post with a link to this thread.  I would like Brewer fans to read these comments.

brewcrewball.com, and marty’s site firenedyost.com should get you plenty of opinion on Yost.

I defended him earlier in the year because i saw changes in him this year, far fewer IBB, less reliance of the sacrifice bunt, using the running game to keep the other team honest instead of as a substitute for power strategy. Also he’ll use the “slash” play in a “traditional” bunting situation, I loved it earlier this year (its since disappeared). He’s even, and I’m sure its been hard for him, played less “gritty” ball with all the power potential in the lineup… fewer hit and runs, etc. And, wow, learning about platoon splits… that was huge. That alone… wow.

But he lost his job as far as I am concerned in the Cubs series one week ago. That was an atrocity of baseball management.

He likes to play his “gut.” in fact, he’ll use “Mench was in the flow of the game” for justification. He’s got what I like to refer to as hero syndrome. He’ll roll the dice with the lesser option and give the guy the chance to be the hero. Just like Yost was 25 years ago when hit that one homerun…

He has zero understanding of leverage. When his “7th inning guy” struggles he’ll go to the next rung DOWN with men on base… in a higher leverage situation… he did it with Aquino twice before they sent him down. Lost two games on his own.

He has done it since the acquisition of Linebrink, who struggles then Yost will go to the mop up guy to bail us out. There is a reason the Brewers have blown so many big leads late, and it’s not just bad pitching and horrific defense, it is setting guys up to fail in the 6th and 7th inning.

Eh, rant over. Why the sudden interest in Yost?

Also, I missed where people were calling for Shouse to be left in. Clearly the best option there is Cordero, or any RHP who wasn’t called up to fill out the roster…

I wonder if it is unique to Yost or not but when he brings someone in from the pen they always seem to perform horribly in the first PA after entering the game… I’ve often wondered if Yost is just too slow getting guys up, and “rushing” them into the game.

He’s like a very bad chess player who can only see the current move.

I would add that in addition to having “roles” for the bullpen, Yost’s policy is to allow them to fulfill that role (their designated inning)regardless of how they are actually pitching this day (the relevant facts, IMHO) or until they have blown the lead.  Only then will a change be made.  Additionally, they are brought in to pitch their inning regardless of how their predecessor has pitched.

It was very unusual to see Turnbow removed before the 8th was over or the lead actually lost, and Cordero would never be removed before the 9th was over or the lead lost.

As has been pointed out, Turnbow will walk batters.  He has a hard fastball with a lot of movement and thus is actually at his best when slightly wild.  It’s almost always obvious within two batters if he has control or not, but this does not seem to be a consideration as to whether or not he stays in the game.  On those nights he has good stuff, he will always come out for Cordero if a save situation is in the making.

Another thing, Yost must consider tie games in late innings to be low leverage situations (he will never bring in a good bullpen pitcher into a tie game).  How do you feel about this?

Finally, Yost will leave the starter in to get five innings at almost any cost.  The Brewers actually have, and have had, relievers capable of starting and thus capable of long relief.  I hate to see them give up on so many games so early when the starter obviously doesn’t have it that day.  I understand the concept of “lately” not being terribly relevant, but I would argue that a pitchers’ performance on a particular night may be relevant given what we know of their tendencies.  Thanks for your comments.

First batter faced by Yost pitchers:

http://www.baseball-reference.com/pi/psplit.cgi?team=MIL&year=2007
http://www.bb-ref.com/pi/shareit/neHW

If you are not a subscriber, that’s 536 PA (includes starters, relievers), for an OPS of .694.  Their overall team OPS in all situations is .758.

Among the relievers: Turnbow .412!, Shouse .458, Cordero .404, Villanueva .727, Wise .918, Spurling .900.

Spurling is an overall .836, and Wise is an overall .753.  Those are the only two guys who are worse.  It *is* a small sample size, but Yost’s relievers have been exceptional with the first batter faced.  This doesn’t prove anything, but at the same time, it throws a bucket of cold water on any assertion that first batters faced in the game are a problem.

It seems that when someone’s head is called for, all kinds of things look bad that simply is not the case.

Happy to learn I’m wrong on that. Also, happy to learn that there is nothing baseballreference.com can’t do. Thanks Tango, and thanks Sean!

When there is new, relevant information, at what point does that overtake historical performance? For example, Edwar Ramirez had ho-hum numbers in the minors through 2003, was released and sat out all of 2004. Then he taught himself a changeup and has been lights-out the past three years: 307 K in 188 IP and a 1.39 ERA (that’s including his time in the majors with the Yankees this year).

Now obviously we have three years of stats for Ramirez since he added the changeup...but back in 2005, at what point would we decide that the current year numbers were more indicative of his true talent level than his previous performance?

Anthony, that is one reason (perhaps the only reason) why we weight previous performance by recency, to account for for an actual change in talent based on new pitches learned, a learning curve in general, injuries, etc.  Once you use a weighted system, going back more than 3 years or so does not have much effect on a pitcher’s projection.

Other than that, a good forecaster is free to tweak projections based on “relevant information,” as much as he wants.  One has to be very careful with this however.  It is easy to misconstrue a drastic change in performance as a drastic change in talent.  Without some “hard evidence” that something has really changed (loss or gain in velocity, new pitch, injury, etc.), I would hestiate to tweak a traditional projection model.  On the other hand, if you don’t tweak your projection model for pitchers like Mussina, Schilling, or Pedro, you are going to way overrrate them.  Of course it is more obvious for pitchers who have lost their velocity or otherwise are old or injured.  One has to be especially careful when a pitcher has a long spate of good or great performance.  For example, it is doubtful that Marquis is anything but the bad pitcher he always was, despite a good 2007.  Of course, when a pitcher does have a spate of good or bad performance, it does change our estimate of his true talent.  Marquis is in fact a little better (our ESTIMATE of his talent at least) than we thought he was before the season started.  IOW, our estimate of any player’s talent is always a MOVING weighted average.

Keep in mind, that in our research for The Book, we DID find a small, but not insignificant (around .1 in ERA I think), carryover effect for recency of performance, unlike with batters, where there is virtually none.

Also keep in mind (BVGigBro) that also in our research, we found that it is very difficult, if not impossible, to figure out if a pitcher is “on” or “off” for the day, based on his results for the game thus far.  We did find a few exceptions to this general rule, IIRC.  One is that when a young or inexperienced pitcher gets hammered early in a game, his performance hence is a lot worse than his norm.  Not so for experienced, veteran pitchers. Also, late in a game, if a pitcher is pitching extremely well (at the latter points, not early), he tends to continue (at least for the next few batters) to pitch well.  We found that if a pitcher late in a game gives up a few hits in a row, or something like that, it does NOT mean that he is tiring or should be taken out.  Of course it always amused when a pitcher late in a game gives up a couple of screaming line drives that are caught or strikes out a batter on a 3-2 pitch way out of the strike zone, the manager leaves him in, but as soon as he walks a batter or gives up a couple of seeing eye singles, the manager takes him out.

My only interest in Yost was that I happened to see him bring in Aquino rather than Cordero in that game, and thought that it was worthy of a discussion about leveraging your bullpen and using your ace, among other things (such as that it was 100% CORRECT to take out Shouse).

I’m out of my depth because I don’t understand every calculation made or know all of the different stats, not because I don’t understand the concept of sabermetrics or the math.  It’s not like I’m a sportswriter who’s job is to understand sports but just can’t be bothered to understand sports.  Give me a little credit.

You can argue that Shouse is a LOOGY, always was a LOOGY and always should remain a LOOGY.  But Yost has not used Shouse as a LOOGY this year, or even much last year.  He has faced more righties than lefties this year and put together his lowest career ERA.  Obviously he performs better against lefties, but his 259/326/294 isn’t ungodly worse than Cordero’s 224/254/302 splits vs righties this season.  Since I don’t think the bases were loaded, the 70 point OBP spread isn’t as critical.

And I stay with this season’s number because theorectically they should be the worst for a LOOGY since he is facing such a high percentage of RHB, and Shouse’s career numbers that I could find are all over the map.  His bad ‘streaks’ seem to span an entire year.

Any pitcher who throws 50-60 IP a year of course is likely to have wild fluctuations from year to year, simply due to randomness.

His number of PA this year is not that unusual for a LOOGY.  He has a few more PA versus RHB, but not that many.  LOOGY’s typically have about 50% PA versus RH and LH batters, since they often face RH pinch hitters and sometimes come in to face 2 out of 3 or even 2 out of 4 lefties.

He has a huge platoon split career-wise and this year, as would be expected, as I said, of a lefty pitcher who throws from the side.  In fact, his numbers versus RHB are really bad for a reliever - .893 OPS.  As I said, you don’t want him anywhere near a righty batter if you can help it.  It is just that it is not really possible to have a LOOGY face mostly lefty batters, for the reasons I stated above.

Him having pitched well versus RHB this year in 85 PA is WORTHLESS information as far as decision-making. 

95% of sabermetrics is knowing the value or lack of, of various sample sizes of historical performance, and understanding when and how much to regress sample numbers toward the mean (and what mean to use).  If you are not intimately familiar with these concepts with respect to baseball, you cannot (cogently) argue (well, you can) these types of things (e.g., that Shouse could have been left in the game to face Pence because he has pitched well versus RHB this year and has been used other than a LOOGY).  In fact, if Yost has used Shouse other than a LOOGY this year, that is one more mark against him, despite him pitching well versus RHB this year.  I have friend who went and played BJ the other day for a couple of hours.  He kept standing on 16’s versus a dealer 7.  The dealer kept breaking and my friend won a bunch of money.  That does not change the fact that he is an idiot (for standing on those 16’s).

Just to be clear, I’m did not mean to be arguing here.  Just trying to get a clear picture of what you are saying and improve my knowledge.  I’m not trying to harp on one particular instance, but I looked at the numbers so they are fresh in my mind.  When he was in Texas in 03, 04 and 05 Shouse faced around 200 batters per season with about 30 more lefties than righties per season.  I take this as the typical LOOGY set up.  Since he has come to the Brewers, he faced about 50/50 in 06 and has faced about 15 more righties than lefties this year. 

At what point do you recalibrate based on different data inputs?  If Yost is using Shouse as a pure reliever and theoretically leaving him in to face quality righties instead of the Juan Pierre’s of the world, and his numbers aren’t going over his career mean wildly how/when is sabermetric analysis capable of handling that?

And while I have you answering questions, have you followed what has happened to Matt Wise recently?  Hit a guy in the face a month or so back, next appearance threw 10 straight balls and has gone downhill from there.  How does sabermetrics account for a guy going head case?  Other than pray you never have to use him in a leverage situation, that is?

That does not change the fact that he is an idiot (for standing on those 16’s).

He should be noted as being very smart if he can walk away after a string of such luck, and never again bet in his life.  But, I will guess that 99.99% of people who are on a winning streak (by playing against the odds) will invariably “let it ride”, hoping to keep the streak alive until, eventually, they have no money left.

***

Sabermetrics doesn’t account for change in human thought process, like Rick Ankiel, Chuck Knoblauch, or anyone else, unless those changes are recorded *prior* to the future performances.  Until these players tell us exactly how they feel and what they are thinking, we won’t know what to do with them.

The Billy Beane v Lenny Dykstra paragraph in Moneyball is exactly that.  On some level, having a too-intelligent hitter should put that player on a worse trajectory path, since he’s the guy more likely to overthink everything.  Successful baseball teams may need more Eliza Doolittles.

As far as a guy like Shouse, there is nothing special “sabermetrics” has to do to project his performance.  You simply take a pitcher’s career overall performance versus RHB and LHB and weight for recency (and batters faced of course).  Current year gets more weight than 06, 06 more than 05, etc.  You can (I do) adjust for the quality of each batter faced (so, for example, if he is facing better quality RHB this year for whatever reasons, that is accounted for).  Then you regress those numbers toward the mean of whatever group your player/pitcher belongs to (LH relievers who are X years old, or whatever).  Then you estimate a pitcher’s true platoon split.  For that you use his historical platoon splits and regress that toward the mean, in this case, a sidewinding lefty.  Now you are left with an estimate of that pitcher’s true (and going forrward) mean performance against average RHB and LHB.  THAT is what you use, generally, to make decisions.  Shouse is an easy one. As I said, he is a sidewinding lefty.  Even without looking at his numbers, we know he is going to have a large platoon advantage.  If we look at his numbers, we see that it is true - he has a large actual platoon split.  He is NOT a great overall pitcher.  Even without “running” the numbers, we KNOW that a LOOGY with a large platoon split who is not a great overall pitcher is going to be terrible against RHB and good to great against RHB.  Again, the numbers confirm this (lifetime .893 OPS against versus RHB and .6 something - I think - versus LHB).  Typical LOOGY numbers. 

As far as what to do with a Matt Wise (and players like that), since sabermetrics is about “discovering the truth in baseball” with or without “numbers,” using every possible relevant piece of data, I don’t need to answer that, do I, nor do I know the answer. Sabermetrics is simply using the scientific method to ascertain the truth and to make deicisons in baseball.  That is what responsible and successful people do in all walks of life.  The scientific method is not only “crunching the numbers” it is using (in a rational manner and understanding that the human mind is a poor processor of information, such as the poster above who thought that Yost pitchers were getting killed on batter one, etc.) all other means of available evidence to arrive at the truth and to make decisions.  The scientific method also involves knowing when certain problems or questions are difficult to answer and knowing when we don’t have or have not looked at enough data to come up with answers of any level of certainty.

I wasn’t clear earlier.  I have no problem when pitchers make good pitches and get hit, as happened to Capuano a week or so ago when he made a good pitch a few inches off the ground and it got golfed into the bleachers.

In Capuano’s case, he tends to get tired and begins to throw more and more changeups, leading to trouble.  In Turnbow’s case, if he’s not wild he’s on and has the best stuff in the Brewer’s bullpen.

Late in the game when a pitcher is cruising is precisely the situation Yost likes to start changing pitchers.  We will hear complaints that the starters aren’t going deep enough into the game, but if the lead is three runs or less they are coming out in the 7th inning regardless.

Thanks.  Shouse is just a specific example to better understand the general approach and if/when you say something has changed, he is no longer A but has become B.  Not trying to beat a dead horse.

I wasn’t trying to show a ‘hole’ in sabermetrics with the Wise question, just wondering if there was anything that looked at when a player went head case and tracked a ‘most successful’ approach to rehabbing them back into their old self.  Be it right back into the fire, sitting them for a month, sending them down, putting them in low leverage situations to get their head straight or if conversely there was a list of don’t dos that will worsen the affect.  I guess what I’m thinking about is a psychological study where you looked at differing on field approaches as ‘therapies’ and tracked the outcomes.

I didn’t know if the stats are available to track this stuff, or if there are proxies that can be used to approximate an answer.  Thought it was a stretch, but the variables availabe that I don’t know could fill the ocean.

But thanks for your time.  I appreciate it.

That is what responsible and successful people do in all walks of life.  The scientific method is not only “crunching the numbers” it is using (in a rational manner and understanding that the human mind is a poor processor of information, such as the poster above who thought that Yost pitchers were getting killed on batter one, etc.) all other means of available evidence to arrive at the truth and to make decisions.

Well said, in part. Using “all other means of available evidence” also means hypothesizing based on observed patterns and investigating them. So that poster above may have been on to something, that has not yet been disproved. For example, why is it that the Brewers have a .924 OPS against in the 6th inning? Unless you start theorizing all you have is that data point, and no cause.

Perhaps the problem is that Yost does a horrible job getting his guys ready for an mid-inning replacement, which is also why turnbow and cordero (the good first batter pitchers) almost exclusively start the 8th and 9th inning? He’s actually compensating for one of his shortcomings. Meanwhile, the bad first batter pitchers, Dessens, Capellan, Aquino, Wise, Spurling are the guys he uses first after pulling a starter?

The human mind is a poor processor of information in an anecdotally biased sense, but its creativity is still unparalleled and a vital component to the scientific method.

Brewerfan.net is the probably the place to start a thread at.  If I had caught this post I probably already would have.  I know Tango has posted there a couple times and I can guarantee a good response from both sides of the debate. 

Yost does a very poor job of choosing the best reliever for the situation, but that’s also true of the majority of MLB managers out there.  He also relies too heavily on small samples (Shouse being a prime example) but again, that’s a problem league-wide. As was mentioned, he’ll use Jenkins or Mench against wrong-handed pitchers in important situations if he “feels” like he should.  He also sometimes seems to sometimes play favorites, although I think Capuano is a bad example.  Cappy isn’t even having that bad of a season if you look at the things that matter most (walk and strikeout rates, IMO).  But because he hasn’t recorded a win in about 10 years, fans have convinced themselves that he’s been absolutely terrible.

With regard to the Aquino call, I was sitting in the stands at the time and simply hated the move.  Obviously, Shouse and then Cordero would be the “right” call but short of that, I would have been very reluctant to use Aquino, since he was booed out of Miller Park and demoted to AAA the last time he had pitched in a high leverage situation.  Predictably, the fans started to boo the second Yost made the move to put Aquino in.  Maybe a major league reliever should be able to ignore all that but it was simply brutal to watch the fans tear into him.  I think I would have spared him that.

For all his warts with regard to in-game management, I think Yost has done an excellent job keeping the relatively young Brewers focused and upbeat in what’s turned out to be an incredibly up and down season.  Obviously, you can’t quantify the value of something like that but I wouldn’t ignore it either.  Of course, I’d trade all that for a more sabermetric-friendly manager but who the heck would that be, exactly?

I’ve never been a Yost fan but I’ve found myself having to actually defend him more and more often this year.  I am certain that Yost has cost the Brewers a couple wins but some are convinced that the Brewers might have 10+ more wins, if not for Yost.  Of course, that’s not really an objective estimate; it’s largely an emotional one.  Brewer fans are understandably frustrated and someone must be blamed.  The rule appears to be that the fewer number of the people blamed, the better, so the collapse is Yost’s fault.

Good post, #25.  Especially with regard to, “What managers would NOT do essentially what Yost does?” If I may paraphrase your sentiments.  You are right with regards to Capuano.  Last year was essentially a career year for him, for whatever reason.  Strangely, he walked no one last year, something he has never done before in his career.  As I have said many times, the reason we use an entire career for a player’s projection is that rarely does all of sudden a player change his true talent.  When a player has a banner year in any category (last year, Cap’s walk rate) it merely means that his projection changes a little.  Sure enough Cap’s walk rate this year is better than prior to 06, but nowhere near last year, exactly what a standard projection model would have “expected.” His overall numbers are not terrible at all this year - again, right around what we/I had projected for him - slightly below average GR rate, below average walk rate, and above-average K rate.  Nothing wrong with him at all.  He is and has been this year a solid #4 type starter.  I have no idea what his win rate or ERA is.  I pay no attention to those things as an analyst, other than to see how a player might be underrated or overrated due to a layperson’s (and unfortunately, a baseball insider’s) persepctive where garbage stats actually mean something (ERA is not really a garbage stat, but if you know a pitcher’s ERC or FIP or DIPS ERA, it becomes meaningless - doesn’t add anything).

GR rate should be HR rate…

I really hate fan sites.  Once you start talking sabermetrics and bash some of the nonsense “fanspeak,” while a few people appreciate it, invariably a few idiots start calling you names and tell you to get your spreadsheet out of your a*** or something like that.  I don’t care of course, but at my age, I don’t need that.

If you are referring to Brewerfan.net, they have a sabermetric section to their message board, where that kind of worthless rhetoric is nowhere to be found.  Really, though, the whole message board is pretty open to stats in general.  The discussion Tango participated in was quite civil.

Yost manages his bullpen not just for inning roles, but also around the save stat.  Which is why Cordero has partial inning saves from when the Brewers have had a 4 run lead, including 1/3 of inning outings.  Which wouldn’t be so bad except that it means that Cordero starts warming up whenever there is a baserunner in the 9th and a 4 run lead.

No I definitely was not referring to Brewerfan.net, just fan sites in general, although I have read very few.