Where's the Evidence that Wedge Should be Fired??????

[Jennifer]

Since the title of the thread says what this thread is about and someone has already accused me of being a Wedge supporter I thought I'd begin with a general explanation of my own views about Wedge. Simply put I have no position regarding whether Wedge should be fired. Shapiro apparently feels Wedge is doing a good job and he is in the best position to know. That doesn't mean that he is right or that the real reason he is supporting Wedge has something to do with his own agenda. The problem in evaluating any manager is lack of transparency. We simply do not know what is going on behind the scenes.

Then there are those on various boards who claim that Wedge should be fired. I read each of their posts carefully to see if they offer a fact, an insight or an opinion that goes on the side of the balance sheet that says Wedge should be fired. If I'm not persuaded than I challenge the poster giving him or her another chance to persuade me.

It seems to me that when there exists questions like whether a manager should be fired that you start with the status quo as the presumptively correct answer and that supportable reasons must be offered to rebut the presumption.

What I see on various boards, which is reflective of real life, are people who make up there minds with precious few facts and then view everything in terms of how does it support their positions. For some no matter what Wedge does they will find a way to say he was wrong (or that is motives were insincere) and should be fired.

Usually the allegations against Wedge are pretty general in nature and with little or no specific support. For instance The Homeless Rhino recently wrote (and I apologize to him for singling him out):

Fundamentals and defense are in part a function of effort, and that's something over which a manager can exert some influence.

Who's defense is he talking about and what control can he exert over it? Same questions regarding fundamentals.

Up until recently it was very popular to argue that Wedge had "lost control of the team" and that "some players had given-up" Who were these players and what support is there that Wedge had lost control?

In responding please keep in mind that poor little old defenseless me cries easily.

Lassiez les bon temps roulez.

[Editor's note: Jen has just gone and hidden under her bed]

[swerb]

Jen, first off, I've appreciated your contributions here. As these message forums become more frequented, I can only hope we can attract more regular contributors like yourself.

As far as Wedge goes, the following things disturb me greatly.

~The teams record in one run games.

~The teams inability to play small ball.

~His inexperience, and the subsequent effect that has on the influence he has on his players.

~His laissez faire attitude. The "easy as it goes, this is a long season mantra" is one that he carried way too far into the season IMO.

~The fact that all Wedge said all spring was that the team was focusing heavily on fundamentals. The team has been the worst fundamentally in all of baseball.

~The influence Shapiro lets him have on personnel decisions. This is Shapiros fault, not Wedges, but its another reason to get rid of him ... as Wedge will likely to continue to have said influence as long as he is manager.

~The teams epic last week melt last year, and their terrible starts to the last three seasons. When the pressure is really on this team to perform, they play their worst baseball.

~The various statistical baseball websites (help me out Jumbo) that all show that the Indians are terribly underperforming (from a W-L standpoint) based on their team statistics. This is the third season in a row that the numbers are showing that this team *should* have a much better record than they do. In one isolated season, I would put no creedence to these stats, but when I start to see it three seasons in a row ... red flags go up for me.

~The porn stache from last year. I'll never forgive him for that.

[furls]

Jen,

First let me start by saying that I think your recent contributions to the board have been terrific.

I also follow the same general belief that you do, that we do not have much insight into the happenings behind the scene. We, the fan and follower, cannot directly tell how much control he has or does not have over this team. It is very difficult to determine how you think control can be determined indirectly, so you have to base it on trends.

The things I look at when deciding how a manager is doing are these:

1.) Defense: No one really likes to play defense (except Omar Vizquel) because it does not pay in free agency. As long as you are serviceable at your position, you get paid if you can swing the bat, but defense is an important part of the game and frankly the Indians are grossly underperforming in this aspect of the game and are ranked last or near last at almost every position, that is a shocking TREND.

2.) Performance of players vs. what is expected of them: I understand that some players just have off years for one reason or another. This is NOT indicative of a manager, but when an entire team has an off year? Well that is a much different thing altogether. That tells you that it is organizational. There is OBVIOUSLY a TREND here on the team.

3.) Record in close games: Rich covered this.

4.) Fundamentals: This is such a crapptastically vague topic that folks always seem to just toss it out there. I will be specific with what bothers me. This team completely lacks the ability to advance runners and score runners from third with less than two outs. Those runs should come across, but they don't. They don't in a lot of cases because of poor plate discipline and lack of fundamentals. How many guys on this team can bunt? Ever seen it? It is UGLY, yet this is a crucial aspect of the game and really puts the tribe behind the eightball in close games.

You know what happens when you get greedy and try to pull an outside pitch with runners on base? It usually ends in a double play grounder to short or second. I do not play major league baseball, nor have I ever, but I still know that little tidbit of information. So why is it that I continually see these double plays end rallies?

Baserunning: we don't need to even mention this. If you have watched the Indians this year you know what freaking adventure that is. It is like seven starters went and got baserunning advice from Manny Ramirez in the offseason. It is an epidemic through the whole team.

5.) Decision making: It is a statistical inevitability that Cliff Lee will get in trouble in either the 6th or 7th inning and have to be pulled out of the game, yet Wedge will usually let things get pretty ugly before giving him the hook. Actually, he pretty much does this with every pitcher. I understand that the bull pen has not been stellar this year, but you have to play all the cards you have. I cannot count how many times I have seen him leave a pitcher in the game long after the pitcher has "lost it."

You see these same issues in shuffling the line up. How many months of .250 batting did it take for him to move Peralta from the three hole to the bottom of the line up? How long does Casey Blake have to hit the hell out of the ball to move up from 8th or 9th? How long do we have to watch Aaron Boone play horrible 3rd base and hit into double plays? How many times do we have to watch guys like Jeff Conine run on V-Mart before we say, this guy is a liability behind the dish that we cannot afford?

Then there are the lack of decisions: platooning NEVER works make a decision and stick to it. It always works out better that way, but he cannot seem to make the commitment and relegate some guys to permanent bench slots.

6.) Attitude and personality: This is a young team and frankly, young players need a boot in their asses and need to be held accountable, yet we keep hearing the, "it is a marathon not a sprint" bull crap. Playoff teams are usually decided by a few games, so what that tells me is that you cannot afford to take nights off. While the season is 162 games long, it was abundantly clear last year that these games that don't seem to matter at the beginning of the season really do. If I have to hear about how long the season is from this guy one more time....

This is why I am dying to pull the plug on the comatose Eric Wedge era. While I cannot directly say how much control he has. I can say that if he does have control, I am actually even less impressed with him than this post would imply (and that says alot).

[consigliere]

I agree that Wedge probably needs to go.....but, we can wish for it all we want because I just don't see it happening. Not at least until after the 2007 season (when Shapiro may walk). The only way Wedge is fired (or let go) is if Shapiro is fired or walks after 2007 when his contract expires.

I agree Wedge has some problems with in-game managing....but I do have to give credit to him in that this team still plays hard for him. At this point in the season when the team has changed course from contending to evaluating, they are still playing hard and haven't mailed it in. That is a testament to the players, yes....but mostly to Wedge.

[furls]

they are still playing hard and haven't mailed it in

Kids who get called up to the majors in mid season always play hard. Not being a smart ass, but quite frankly (screamin a smith impersonation) that is what is driving this team right now. 1/2 of the guys seeing playing time today are AAA kids getting a shot and the remainder are pretty much guys fighting for their jobs (with the exception of sizemore, pronk, and peralta). Competition always makes people play harder.

[pup]

DAMN! My favorite topic since my pursuit to get Chuck Manuel shit-canned(actually it started with the begging to not hire him), and Swerb and Furls have already made every point, and done so quite well.

The thing that boggles my mind the most is how you can be so bad in one run games. Things in baseball ALWAYS even out. Maybe not over a week, or even a month, but 3 years, everything statistically evens out. One run games, 3 run games, batting averages and fileding percentages all usually work their way back to your actual ability. For any team (and sorry for the lack of time to do the proper research) to go 30-70(?) in one run games is downright scary. I don't know if you could out there and say I am going to make every move possible to try to lose one run games and be that effective.

Teams have a pulse. When you are spending as much time with 25 other guys as a baseball player is, the team as a whole becomes a character. The best managers always have a pulse on what their team needs. A manager has to know when to fire away and rip some guys and when to stroke some egos. It doesn't appear that Wedge has any clue when to do which one, and that leads to the uninspired play that we see every night, until the kids showed up.

Jennifer, I would hate for you to think that I am someone jumping on the anti-Wedge bandwagon late. Here is a link to my opinion of him, which was written in early May:

http://swerbsblurbs.com/article_detail.php?id=366

And welcome to the boards. Fresh takes from new keyboards are always good to get things going on here!

[Jennifer]

Thanks everyone for responding. I intend to respond to every comment offered eventually flush out the opinions offered in some instances which is a somewhat daunting task. I'm going to try to limit my responses to a single point made by one or a group of posters in an effort to try to keep the lines of discussion clean. Unfortunately, the issues raised do overlap. I'm doing this in an effort to get some extended discussions. In my short time here it is obvious how intelligent and knowledgable so many of you are (far more so then me but I never learn anything if everyone always agree with me).

First response is an easy one. Mr. Swerbinsky wrote:

The various statistical baseball websites (help me out Jumbo) that all show that the Indians are terribly underperforming (from a W-L standpoint) based on their team statistics. This is the third season in a row that the numbers are showing that this team *should* have a much better record than they do. In one isolated season, I would put no creedence to these stats, but when I start to see it three seasons in a row ... red flags go up for me.

Rich - I assume you are talking about "expected wins" based on "Pyth") Do you have any basis for your implied assertion that the Tribe's negative number is Wedge's fault? My own considerable expertise on the subject and the results of my many years of research failed to find any significant correlation.

To determine whether managers have an impact or whether randomness preveils the coefficient -- R-squared, is frequently used -- is useful. Essentially one compares the manager's ability to exceed his team's Pyth record across seasons. If the manager does have the ability to affect his team's actual W-L record when compared to his Pyth W-L R-squared should be closer to one (perfect correlation) than to zero (total randomness).

In sampling over 1,000 seasons, R-squared is 0.002. In other words a manager's previous season has for all intents and purposes no correlation with his current season with regards to exceeding Pyth W-L. Even larger sample sizes (sets of even-or odd numbered years are used) R-squared is only 0.030, or still nearly zero.

Class dismissed.

[Editor's note: Jen is a math dunderhead. The above is really a paraphrase from Chapter 4-3 of Baseball Prospectus's Baseball Between the Numbers ]

[Hornless Rhino]

Jen,

No need to apologize for singling me out.

I don't have much to add to what Rich, Furls and Pup have said on the subject, other than to say that I think the team's errors and ghastly overall fielding percentage speak for themselves when it comes to the question of "whose defense" is the issue.

As for what Wedge can do about it, well, if you buy into my view that effort is part of the problem when a team experiences a continuing breakdown in defense, baserunning, situational hitting, etc., a manager has options ranging from not writing somebody's name in the lineup card to dumping the post-game buffet all over the locker room carpet.

[consigliere]

I'm still on the Wedge fence.

By the way, in 2004, he was 26-20 in 1-run games. Maybe 2007 evens things out?

[furls]

Rich - I assume you are talking about "expected wins" based on "Pyth") Do you have any basis for your implied assertion that the Tribe's negative number is Wedge's fault? My own considerable expertise on the subject and the results of my many years of research failed to find any significant correlation.

To determine whether managers have an impact or whether randomness preveils the coefficient -- R-squared, is frequently used -- is useful. Essentially one compares the manager's ability to exceed his team's Pyth record across seasons. If the manager does have the ability to affect his team's actual W-L record when compared to his Pyth W-L R-squared should be closer to one (perfect correlation) than to zero (total randomness).

In sampling over 1,000 seasons, R-squared is 0.002. In other words a manager's previous season has for all intents and purposes no correlation with his current season with regards to exceeding Pyth W-L. Even larger sample sizes (sets of even-or odd numbered years are used) R-squared is only 0.030, or still nearly zero.

Class dismissed.

I specialize in probability and statistics, but I refuse to apply this to baseball. For those that are confused or mathematically impaired, let me briefly explain the "pyth." The Pyth or "Pythagorean Method" is a crude method of attempting to correlate wins and losses with Points scored and points allowed. Basically, you take the points allowed to some exponent power and then divide it by the sum of the points allowed added to the points scored to generate a winning percentage. It is used in lots of sports and it is very VERY crude.

so the formula looks like this:

Let X be points scored,
Let Y be points allowed,
and let Z be the empirically (based on observations) determined exponent:

Expected winning percentage = X^z/(X^z + Y^z)

Here are some general (and obvious) observations: If y is very small (say approaching 0) then the formula looks a lot like X^z/(X^z +0) this reduces to one, or a winning percentage of 1.000. That is pretty reassuring and a good start to the model, teams that allow 0 points dont lose games!

Now lets say that our team never scores a run, so X=0. Hmmm, now our ratio turns out to be 0. That too is very reassuring, for teams that score 0 points win 0 games.

Unfortunately masked in our early work here is one interesting fact, the variance is also 0. Variance is the measure of "randomness" of our variables X and Y and in our first to tries, in order for a team to score 0 points or allow 0 points it must do so in EVERY game.

The problem with using the Pyth for anything is that the day to day variability is masked in the sum, and it ignores trends. This variability and these trends are absolutely essential in making any type of extrapolation. Simple things, like mop up time and close loses and blow out wins skew these expectations greatly.

Now before someone busts out a "Correlated Gausian Model" on me, let me start by first saying again that this will again fall into the "faulty generalization" you see in the first example due to data loss. While this does account for "variability" by applying the standard deviations, it does not allow for trends that skew the empirical distribution from its expectation. Again, a team that wins by blow out and loses close games is not accurately represented, nor is a team that wins close games and loses blowouts.

Take this years Indians for instance. A little research indicated that baseball stat geeks typically use 2 as the exponent (z). This yields a Pyth of .541, almost 22% higher than their actual performance this year. Why is there this enormous discrepancy? The fact that the Indians lose close games and win by blow out. According to my rough (and prbly inaccurate count) the Indians are about 10-20 in one run games. That might explain it. Something nice came out as a result though, it just so happened to be exactly 30 games. That is nice because most statisticians place the minimimum threshold for applying the Central Limit Theorem to 30 trials.

But what does having a winning percentage of .445 and a close game percentage of .333 tell us about the Indian's ability to win the close game? 3rd grade math tells me that the Indians should have won 13.35 of those games.

If you dont like math step away from your computer and skip this paragraph: Whether the Indians win or lose a game is a Bernoulli trial with probability P. The net sum of Bernoulli trials generates a binomial distribution parameters {n,p}, {30, .445} and therefore using the moment generating function, variance = p(1-p), but since there are n trials (30 in this case) that can be written as np(1-p) or 30*.445(.555)=7.40925. Standard Dev= (Variance)^.5= 2.72. Applying the Normal approximation to the Binomial distribution yields Z= 1.23 and therefore, the probability of the Indians, based on their winning percentage, winning 10 (or less, the "tail probability) of 30 games = (.5(1-Z(1.23)) = .05465. It stands to reason that this would be a small probability event due to the large number of trials and the fact that is over one standard deviation from the expectation.

Ok, the coast is clear for non math geeks! Basically, the last paragraph shows that the odds that the Indians (independent of all variables but winning percentage) lose 20 of 30 games is about 5%. This leads me to the conclusion that there is more at play here than the obvious, either this is a statistical anomally or there is a bigger problem than just a couple of games. In either case, neither the Pyth nor the Gausian Correlated Method are going to account for this.

Personally, I think the reason is that Eric Wedge is a crappy manager and he cannot effectively use his team to win games that they should, but hell, give me a couple of more seasons, a few beers, and I think I may be able to show it more convincingly.

As far as using crappy models like the Pyth, I would not recommend it. These things are generally oversimplified models that do not take key aspects of the problem into account. Essentially, most of these simple models assume that all the samples are unbiased when in fact it is pretty obvious that there are extremely difficult to measure "biasing" factors at work.

As for the fact that the model seems to sort of work a lot of the time, yielding an answer that is fairly consistant with the overall final results, I would say to consider the fact that those exponents are empirically determined based on "curve" fitting to do exactly that, with little attention to the factors that determine the other variables, Points scored and points allowed. Therefore, this model falls into the "statistical generalization" trap. I would say look at the runs scored and runs allowed for any team and then guess the record of that team, I bet most of the time you will be fairly close also.

[pup]

Yeah, what he said.

Holy shit. I don't know what is more amazing, those 2 posts, or me sitting here reading them both in their entirety at 2:00 AM!!!!!!!

[Jennifer]

Furls - In quoting me you failed to quote the part where I admitted I was clueless about what I was saying. However, your response was very reassuring. Conceptually I have suspected that a Pyth calculation of winning percentage tells us very little. It seems to me, especially after reading your response, that the premise is the greater the run differential the higher the expected winning percentage and that this percentage can be predicted by using a Pythagorean formula (I believe BP uses 1.83 instead of 2.0) or Pythagenport. A team that wins a number of games by high scoring blowouts but loses many more games that are close will have an actual winning percentage lower than than what is predicted. A team that wins a lot of close games but suffers some high scoring blowouts on the other hand will have an actual winning percentage higher than its expected winning percentage.

Since, unlike you, I could not demonstrate based on statistical theory have very little Pyth contributes to the discussion my earlier post accepted the premise of Pyth having some validity beyond looking at raw run differentials (since many think that it does) and then went on to show that even a prime advocate of Pyth, BP, also asserts that that managerial skill does not explain the difference between actual and expected winning percentages.

In other words, whether Pyth gives us something better than raw scoring differentials it offers no generalized explanation about why teams's expected winning percentages differ from actual winning percentages.

You, as well as others, also raised the number of one run games the Tribe has lost has an indicator that Wedge is a poor manager. In my mind, and it appears that you agree, that expected winning percentages and performances in low scoring games are separate issues. In any event, I believe they are, and it is my intention, as I noted in my last post, to eventually get around to addressing it (too much thinking all at once gives me a headache). It is comforting to know in advance, however, that you will not pullout some statistical model that I won't understand to challenge me.

[furls]

I guess to I should post the following to clarify my points:

1.) Essentially what I was saying in the beginning is that the pyth is an overgeneralized piece of crap that really offers no insight into much of anything, nor do any of the other "simple" models. These are complex problems with many variables related by varying degrees.

2.) In the second part, I was trying to show that their is in fact something wrong with the Indians winning % in close games. It should be much higher than it is. This is obvious to anyone who pays attention to the Indians, but I was trying to show just how improbable their performance is as compared to winning % alone. Basically that their have to be other factors at work. What are those factors? I cannot say for certain. Could it be that Wedge sucks and gets out managed in close games? sure. Could it be that his players choke in close games? Sure. It could be any number of possibilities, but the fact of the matter is that something is rotten in the state of Ohio.

[Jennifer]

In the second part, I was trying to show that their is in fact something wrong with the Indians winning % in close games. It should be much higher than it is. This is obvious to anyone who pays attention to the Indians, but I was trying to show just how improbable their performance is as compared to winning % alone. Basically that their have to be other factors at work. What are those factors? I cannot say for certain. Could it be that Wedge sucks and gets out managed in close games? sure. Could it be that his players choke in close games? Sure. It could be any number of possibilities, but the fact of the matter is that something is rotten in the state of Ohio.

Yes, I did understand that was a point you were trying to make. Sorry if I did not express it better.

[consigliere]

I um...erm....uh.....nevermind.

Jesus H.

All I can add to this is that I think Travis Hafner is a very good hitter.

[Jumbo]

So, I can't follow all the advanced math - as the cliche goes, a little knowledge is a dangerous thing - so I can only repeat what's been said in other places, or rely on "simple" models. I also am biased in favor of run differentials...at least, they give us something positive to hope for next season, i.e., the thought that the Indians' core talent is sufficiently productive that if they can change whatever it is that causes them to lose close games, they will be better.

There are a few related points I'd like to comment on:

1) I think Furls' conclusion #2 in his most recent post is really the important point. The Indians record in winning one-run games is messed up...whether that cause is Wedge, choking, the bullpen being relatively more important than other parts of the team, etc. Even if the math doesn't add up exactly, it seems like it would have an impact on the Pyth numbers. More to the point, somebody (I think it was Rob Neyer, when he was first popularizing statistical analysis) wrote (paraphrased), "Good teams don't win close games. Good teams win blowouts." I.e., one-run records (in a given year) tend to be fairly random, but teams with the best records tend to be the ones that win games by four, five, or more runs. According to BP, they are 10-2 in games decided by 10 or more runs, 20-20 in games decided by 4-9 runs, and 23-42 in games decided by 3 or fewer runs.

2) As related to BP's book, which Jennifer cites as stating:

To determine whether managers have an impact or whether randomness preveils the coefficient -- R-squared, is frequently used -- is useful. Essentially one compares the manager's ability to exceed his team's Pyth record across seasons. If the manager does have the ability to affect his team's actual W-L record when compared to his Pyth W-L R-squared should be closer to one (perfect correlation) than to zero (total randomness).

One question that immediately arises (and that may have been covered in the book), is that we aren't looking for "managers" ability to impact their team, but rather a particular manager: Wedge. This is somewhat related to that ongoing issue over whether pitchers can influence batted balls in play, and one conclusion (don't know whether it's the correct one, but it's just an example) is that while pitchers as a whole don't influence balls in play, some particular pitchers (e.g., knuckleballers, perhaps extreme groundball pitchers) might.

Some people have done studies (can't find anything good on google right now) that suggest that some managers, over the course of several seasons, do underperform or overperform their pythag. The question then is whether these are stastically significant. (But again, we get back to issue #1 and sample size: The Indians record in 1-run games in recent years, and perhaps by extension their expected W-L, is off by a factor that seems greater than pure chance. Is it?)

3) Even if run differentials and records in 1-run games are complete, 100% bunk, they at least give us something to work with and compare among teams. Things like fundamentals and attitude are always easy to observe when you are watching your own team for 16, 82, or 162 games, but its hard to compare your team to others that you don't watch on a nightly basis...or have such an emotional attachment to. The result is that these kinds of things, often resulting from selective memory and really bad results, tend to be the bugbear of fans everywhere.

I think this is entirely natural...and, as a related digression, the "vocabulary of complaint" is one common to all sports fans. Consider the following two parapgraphs:

"And I could spend the next 3,000 words ranting and raving about the unacceptable performance of the Dolan/Shapiro regime since the team began to contend last season -- the catastrophic Johnson/Vazquez signings; lowballing Millwood/Howry, then half-heartedly chasing other free agents; overvaluing Byrd (a genuine disappointment) and Michaels (a colossal disappointment); undervaluing their own prospects (Brandon Phillips and Josh Bard); getting litte in return for solid veterans; dealing Riske and Rhodes without knowing about the strength of the bullpen; allowing Fernando Cabrera to pitch in the WBC (he's a walking corpse now); letting Howry and Crisp go; handing Boone that unconscionable extension (I yelped out loud when I saw the headline); and we haven't even mentioned last winter's soap opera with Dolan and the payroll yet -- but I don't want to ruin my chances of getting a key to the office next season. So let's just say that everyone did a swell job and I fully support every moronic decision that was made. Now where's my key?"

The work of Regano? No. Bill Simmons...with just a few edits:

The most painful edit, of course, is the phrase following "since" in the first sentence. It would be nice to have such problems.

"And I could spend the next 3,000 words ranting and raving about the unacceptable performance of the Henry/Theo regime since they won the World Series -- the catastrophic Renteria/Clement signings; lowballing Pedro/Damon, then half-heartedly renewing talks at the last second; overvaluing Beckett (a genuine disappointment) and Crisp (a colossal disappointment); undervaluing their own prospects (Hanley Ramirez and Anibal Sanchez) in the Beckett trade; freezing at this year's trading deadline; dealing Arroyo without knowing about the health of Wells and Foulke; allowing 40-year-old Mike Timlin to pitch in the WBC (he's a walking corpse now); letting Roberts and Cabrera go; handing Beckett that unconscionable $30 million extension (I yelped out loud when I saw the headline); and we haven't even mentioned last winter's soap opera with Theo yet -- but I don't want to ruin my chances of getting a key to the office next season. So let's just say that everyone did a swell job and I fully support every moronic decision that was made. Now where's my key?"

4) Jennifer's comment citing BP uses pythag as a means to get at the effect managers have on games. Not the issue of whether pythag itself is valid (which furls goes after). At the very least, the BP people place some validity in pythag, or at least some form of run differential analysis: there's a reason why they still rank the Indians #11 in baseball.

Also, and this is getting way way outside the topic and into pure speculation (if it hasn't already strayed that far), but there seems to be a certain "connection" between a lot of these advanced metrics. For example, I think the key assumption of WARP (wins above replacement player) is that you can estimate wins on the basis of runs (VORP), which you can estimate on the basis of individual performances at the plate. If the wins/run differential connection is faulty, then wouldn't the stat be meaningless? (This is relevant to the fact that conventional wisdom is that a team should maximize its spending when that spending would take the team from ~85 wins to ~95 wins, into playoff contention. If the Indians can't predict, based on their roster and offseason targets, that free agent X could provide them a certain number of runs, and thereby provide them a certain number of wins, and thereby increase their chances of making the postseason by a certain amount, then it's difficult to make a rational(*) estimate of what to offer free agent X.

My assumption that decisionmaking in sports is rational is, of course, ridiculous.

[Jennifer]

It is a statistical inevitability that Cliff Lee will get in trouble in either the 6th or 7th inning and have to be pulled out of the game, yet Wedge will usually let things get pretty ugly before giving him the hook. Actually, he pretty much does this with every pitcher. I understand that the bullpen has not been stellar this year, but you have to play all the cards you have. I cannot count how many times I have seen him leave a pitcher in the game long after the pitcher has "lost it."

In responding let me start with some general observations. There seems to me to be at least five major reasons to remove a starting pitcher from a game (1) injury; (2) going beyond a
certain pitch count (when a pitcher goes beyond a certain number of pitches he is more prone to lose his mechanics thereby increasing risk of injury; There is also some evidence that a pitcher exceeding a certain individual pitch count is not as effective in his next few starts); (3) lack of effectiveness; (4) actual loss of mechanics from being tired (this is a decision made by the manager based on input from the catcher and his pitching coach); and (5) late in the game for matchup purposes.

Obviously, starters should always be pulled when injured, starts to get beyond the designated pitch count or begin to lose their mechanics due to being tried. Whether pulling a pitcher for
the other two reasons, however, does require some judgment. One factor you mentioned is the quality of the bullpen but even with a usually reliable bullpen there are often reasons not to go
to the bullpen. The bullpen might be tried. The bullpen might be reasonably rested but the next game's starter might be pitching poorly and an effort must be made to save the bullpen as much as possible for fear that it will have to be used extensively the next day. The score is also important. If the team is up or down a bunch of runs keeping the starter in awhile longer is a prudent decision for trying to achieve long-term success.

A manager has to try to be satisfied if his number five starter can routinely give him five good innings but less than a routine six for starters one-three is just not acceptable. No bullpen can regularly handle multiple starting pitchers going less than six full innings. Part of the reason for the Tribe's bullpen success was due to four starters averaging more than 6 innings per start and I believe Elarton was just below that threshold.

Some pitchers stuff just is not good enough to allow them to get successfully through the order more than twice. Others have good enough stuff but either lose concentration as the game goes on or have not learned how to pitch when facing the order a third or fourth time. Leaving them in games, particularly when up or down a bunch of runs, gives them experience in dealing with concentration problems and facing the order three and four times. In my opinion routinely
taking out a pitcher who has been effective in the fifth makes no sense whatsoever unless he fails a large part of the time to be ineffective after the fifth at it is an important game or that the pitcher has demonstrated to the satisfaction of the manager and pitching coach that he is not capable of going beyond five innings.

This brings us to Mr. Lee. By my count he has started the sixth inning nineteen times this season:.

4/5 - 5.0 innings, 3 ERs, 1 in the 4th and 2 in the 6th:
4/11 - 6.0 innings, 2ERs, 2 in the 6th;
4/16 - 7.1 innings, 1 ER, 1 in the 4th (6th and 7th no runs);
4/26 - 6.0 innings, 1 ER, 1 in the 4th (no runs in 6th);
5/6 - 6.2 innings, 4 ERs, 3 in the 4th, 1 in the 7th (no runs in the 6th);
5/12 - 7.0 innings, 5 ERs, 1 in the 2nd, 1 in the 5th, 3 in 7th (no runs in the 6th)
5/18 - 6.0 innings, 4 ERs, 1 in 1st (unearned), 1 in 2nd, 3 in 7th (no runs in the 6th);
6/3 - 7.0 innings, 2 ERs, 2 in 6th (no runs in 7th);
6/9 - 7.2 innings, 4 ERs, 2 in 1st, 1 in 6th, 2 in 8th (no runs in 7th);
6/15 - 6.2 innings, 3 ERs, 1 in 4th, 2 in 7th (no runs in 6th);
6/20 - 7.2 innings, 1 in second (unearned), 1 in 6th (no runs in 7th);
6/26 - 6.0 innings, 2 ERs, 1 in 3rd, 1 in 4th (no runs in 6th);
7/1 - 6.0 innings, 4 ERs, 3 in 4th, 1 in 5th (no runs in 6th);
7/6 - 6.0 innings, 4 ERs, 3 in 5th (1 unearned), 2 in 6th;
7/13 - 5.2 innings, 2 ERs, 3 in 6th (1 unearned);
7/18 - 5.0 innings, 6 ERs, 5 in 4th, 1 in 6th;
7/30 - 7.2 innings, 3 ERs, 1 in 2nd, 1 in 3rd, 1 in 4th (no runs in 6th or 7th);
8/4 - 6.1 innings, 4 ERs, 2 in 6th, 2 in 7th;
8/10 - 7.0, 2 ERs, 2 in 4th (no runs in 6th and 7th).

I didn't examine the scores of the 6th, 7th innings and 8th innings which Lee was permitted to start nor how many inherited runners were permitted to score (nor how many inherited runners the bullpen prevented from scoring) because I did not find it necessary based on the data that I did look at. To me the data does not establish "a statistical inevitability that Cliff Lee will get in trouble in either the 6th or 7th inning and have to be pulled out of the game, yet Wedge will usually let things get pretty ugly before giving him the hook." Without further analysis it appears to me that Wedge based on my introductory discussion and the above data has been pretty reasonable in his use of Lee beyond the fifth.

[furls]

By my count, based on your data, Cliff Lee has given up 28 runs from the 6th inning on. Also by my count, Cliff Lee has pitched 26.2 innings after completing the fifth inning.

This gives him an ERA of 9.61 after the fifth inning. That is problematic, I think even Carmona's ERA is lower than that. Every pitcher in the horrible bullpen has a lower ERA than that, therefore, it may be a wise to choice to make a move.

Now consider that Lee has given up, according to your count, 51 ER total this year in the starts you showed. This implies that Lee has given up 23 runs through the first 5 innings of those 19 starts giving him an ERA of 2.17 until the fifth inning in those same games.

Now all together this year, Lee has given up 77 runs (28 after the 5th), this gives him an ERA (for all 25 of his starts) of 3.57 before the sixth inning.

I did not mean to imply that Wedge should pull Lee after the fifth in every game, but I would definetly start warming someone up in the pen and I would pull him at the first sign of trouble late. It would be super duper terrific if he could go into the seventh or eighth, but it just doesnt seem to go well for him there.

A key part of managing your people is knowing their limitations and strengths and sometimes it is more important to minimize the former rather than maximizing the latter. Don't get me wrong, I like Cliff Lee very much. I think he is a terrific pitcher, up until the 6th, and I think he is one of the important "core players on this team." I just hate to see him mismanaged.

Now, Jennifer, on a more personal note, I don't want you to think I am attacking you or trying to be a jerk or something. Frankly, I think you have done a fantastic job of riling up some pretty compelling dialogue here that goes well beyond the typical, "Eric Wedge sucks and needs to be replaced because the Indians should be better and I am tired of rooting for a loser."

[Jennifer]

Now, Jennifer, on a more personal note, I don't want you to think I am attacking you or trying to be a jerk or something.

Furls and everyone else - There is no need to treat me with a pair of kid gloves (although anyone wanting to send me a nice pair of kid gloves should feel free to do so ). I would not be raising things that some of you feel very strongly about if I did not expect strong disagreement in response. Moreover, none of my opinions are so strongly engrained that good arguments won't convince me to rethink or change my opinion. Not to be repeatitive but I never learn anything if people always agree with me. To me a willingness to admit being wrong is a sign of a secure ego not a character flaw.

Now turning to the matter at hand.

I did not mean to imply that Wedge should pull Lee after the fifth in every game, but I would definetly start warming someone up in the pen and I would pull him at the first sign of trouble late. It would be super duper terrific if he could go into the seventh or eighth, but it just doesnt seem to go well for him there.

Compare with the post I was responding to:

Decision making: It is a statistical inevitability that Cliff Lee will get in trouble in either the 6th or 7th inning and have to be pulled out of the game, yet Wedge will usually let things get pretty ugly before giving him the hook. Actually, he pretty much does this with every pitcher. I understand that the bull pen has not been stellar this year, but you have to play all the cards you have. I cannot count how many times I have seen him leave a pitcher in the game long after the pitcher has "lost it."

In your last post you also asserted that:

By my count, based on your data, Cliff Lee has given up 28 runs from the 6th inning on. Also by my count, Cliff Lee has pitched 26.2 innings after completing the fifth inning.

This gives him an ERA of 9.61 after the fifth inning. That is problematic, I think even Carmona's ERA is lower than that. Every pitcher in the horrible bullpen has a lower ERA than that, therefore, it may be a wise to choice to make a move.

Now consider that Lee has given up, according to your count, 51 ER total this year in the starts you showed. This implies that Lee has given up 23 runs through the first 5 innings of those 19 starts giving him an ERA of 2.17 until the fifth inning in those same games.

Now all together this year, Lee has given up 77 runs (28 after the 5th), this gives him an ERA (for all 25 of his starts) of 3.57 before the sixth inning.

So much to disagree with.

1. Under your reasoning Wedge should actually get someone warming up after the third because Lee in the starts I have identified has given up 16 ERs in the 4th inning over 8 of the 19 games I showed. Moreover, my sample ignored games he did not make it past the 5th.

2. From your earlier post you made it clear that you would yank Lee as soon as he gets into trouble after the 5th. Yet in a number of games he gave up runs in the 6th and was then effective in the 7th.

3. In your earlier post you claim "yet Wedge will usually let things get pretty ugly before giving him the hook." Perhaps are definitions of "ugly" differ. In only one of his starts did he give-up more than 2 earned runs in the 6th, 7th or 8th (in one game he gave up more than 2 runs (he gave-up 3) and that was unearned.

4. Your warming up assertion I believe you will agree is an large over generalization since it ignores actual game situations. In some of these games the Tribe had healthy leads going into the sixth, seventh and eighth innings. In some of these games I would presume that Wedge needed the innings from Lee for bullpen reasons. In some of these situations Wedge may have had someone warming up in the pen either because Lee was obviously tiring or nearing his pitch count limitations. You also seem to ignore the introduction in my response regarding the pitchers learning how to go deep into games.

4. In originally singling out Lee and then asserting that someone should be heating up in the pen after the fifth you seem to be ignoring the bigger picture. If all number one-three starting pitchers were looked at I believe you would find that the large majority of them do give-up runs in the sixth and seventh innings. This is the point that they are beginning to face the order a third time. If you look at hitting stats you will find that many hitters BA (and other hitting stats) show a general tendency to improve as they face a pitcher in a game an additional time.

5. If Lee were the only starter on the staff that the Tribe had to "worry" about after the fifth your point might be worth considering but he is not (see point "4"). From a bullpen management standpoint you can't have relievers warming up after the fifth on a routine basis. Relievers can only get up and sit down two times or three times before you lose them for the game.

6. Perhaps you disagree with the philosophy but managers look for pitchers to give them "quality starts" (at least six full innings allowing 3 or less earned runs). Getting six or seven innings out of a pitcher is more important then how the runs are distrubuted over the course of those innings. Giving-up a run or two in the sixth in many games is less important than getting those six or seven innings. The key is not whether a pitcher gives-up a run or two but whether the manager and pitching coach believe that the pitcher can still be effective if left in. In Lee's case he has shown that even though he gives-up a run or two in the sixth that he in the instances where he has been left in that he can be effective in the seventh and even into the eighth.

[furls]

Moreover, my sample ignored games he did not make it past the 5th.

I did while calculating that adjusted ERA, 3.57, before the 6th.

Decision making: It is a statistical inevitability that Cliff Lee will get in trouble in either the 6th or 7th inning and have to be pulled out of the game,

I thought the sarcasm was pretty clear by the overstatement regarding statistical inevitability.

1. Under your reasoning Wedge should actually get someone warming up after the third because Lee in the starts I have identified has given up 16 ERs in the 4th inning over 8 of the 19 games I showed. Moreover, my sample ignored games he did not make it past the 5th.

It is rather interesting that you point that out, because I was thinking something similiar. Obviously teams are making adjustments to Lee and punishing him the second time through and it looks like he gets roughed up even moreso the 3rd time through. Not a shocking revelation, but not exactly contradictory evidence.

2. From your earlier post you made it clear that you would yank Lee as soon as he gets into trouble after the 5th. Yet in a number of games he gave up runs in the 6th and was then effective in the 7th.

Yep, and lots of times he doesnt get out of the inning either, so what is your point?

3. In your earlier post you claim "yet Wedge will usually let things get pretty ugly before giving him the hook." Perhaps are definitions of "ugly" differ. In only one of his starts did he give-up more than 2 earned runs in the 6th, 7th or 8th (in one game he gave up more than 2 runs (he gave-up 3) and that was unearned.

This is not just a statement about cliff lee, this is a generality covering most of his decisions, hell it even applies to the line up and etc. As you stated, this does not accurately reflect inherited runners and what the situation was when he left, nor does it accurately represent what his control and mechanics looked like. Times when I am sitting there watching the game and saying oh crap, this guy is losing it after he walks a batter and goes 2-0 on the second. OK, now please jump out and say, "oh we should pull every pitcher who falls behind in the count late in the game," because that is EXACTLY what I just said.

4. Your warming up assertion I believe you will agree is an large over generalization since it ignores actual game situations. In some of these games the Tribe had healthy leads going into the sixth, seventh and eighth innings. In some of these games I would presume that Wedge needed the innings from Lee for bullpen reasons. In some of these situations Wedge may have had someone warming up in the pen either because Lee was obviously tiring or nearing his pitch count limitations. You also seem to ignore the introduction in my response regarding the pitchers learning how to go deep into games

In some games I am sure they did have good leads, in others I know that they did not and these are exactly some of the decisions that lead to the Indians losing close games, at a rate that is inconsistant with their statistical expectation. Again, this is not limited specifically to Cliff Lee, and further you are ignoring EXCEPTIONALLY compelling evidence that Lee has an ERA of almost 10 after the fifth. I would expect an ERA to increase with fatigue, but should it really be allowed to increase 3 fold late in the game?

4. In originally singling out Lee and then asserting that someone should be heating up in the pen after the fifth you seem to be ignoring the bigger picture. If all number one-three starting pitchers were looked at I believe you would find that the large majority of them do give-up runs in the sixth and seventh innings. This is the point that they are beginning to face the order a third time. If you look at hitting stats you will find that many hitters BA (and other hitting stats) show a general tendency to improve as they face a pitcher in a game an additional time.

exactly my point. You know that Lee is going to get touched up on the third time through, some guys only have 1-2 times through the order stuff. That is life, you and I recognize it, so do lots of other folks, so why would you want to make a mistake that is a general tendency, as you put it. Isn't that part of managing? Identifying trends and adapting to them? Or is managing a baseball team more about sitting on your ass in the dugout and giving speeches about "marathons?"

Giving-up a run or two in the sixth in many games is less important than getting those six or seven innings.

Giving up a run or two in the 6th or 7th inning for the sake of a "quality start" is a foolish consistency, which to paraphrase Emerson, is the hobgoblin of little minds. Don't get me wrong, I understand that your pen cannot go four strong everynight, but to say that Lee has been effective (consistently) after the fifth is real stretch.

The key is not whether a pitcher gives-up a run or two but whether the manager and pitching coach believe that the pitcher can still be effective if left in.

and this brings us back full circle. I wrote this as part of a greater issue regarding Wedge's decisions, yet it has turned into a detailed discussion about Lee's late game effectiveness. Ultimately, it is the fact that Eric Wedge believes Lee will be effective later in the game that is problematic. With an ERA approaching 10, it is clear that he is not.[/quote]

[Jennifer]

Furls - There is much I intend to respond to with regard to your latest post but I need first to respond to some of the other Wedge posts.

I do, however, want to thank you for taking off the gloves and responding aggresively. I hardly cried at all.

Note to TBTB - You really need a much greater variety and number of emoticons.

[consigliere]

I second the emoticon request.

[furls]

thought I would bump this, since it is Indians season and therefore, fire Eric Wedge season.