How much of Pythagorean differential is really about the manager and how much is about the style of baseball/kind of talent the teams they managed had? (That is, a team built around good pitching and mediocre offense seems like it would have a less great run differential and therefore a less awesome Pythagorean W-L than a team with the same number of wins built around mediocre pitching and awesome offense. Therefore a pitching-heavy, low-scoring team would be more likely to play above their Pyth W-L than a high-scoring team whose offense makes up for its mediocre pitching.) That is, are we blaming stuff on the manager that’s not necessarily about the manager?

HIRE AL LOPEZ.

The exception that proves the rule. A fanpost from someone who joined within the last 48 hours that isn’t total garbage. Welcome, good to have you around.

I’m not sure that I agree with your analysis, since I don’t really know enough about how Pythagorean records are created, and a lot of these kinds of statistics don’t come intuitively enough to me that I can see implications on different styles of play, eras, etc., without it being explained to me. Not that I distrust stats or dislike them, my mind just doesn’t work that way easily.

I usually dislike the knee-jerk, “Fire the Manager” argument. I’m not a huge Wedge fan, however. I do think he did a good job last year (maybe an extremely obvious statement), and his even-keel nature really helped when we struggled in the middle of the summer, and I would contend helped us stay hot once we got going towards the end of the season. Baseball is a long season, there’s got to be something to be said for keeping the lows from being too low and the highs from being too high. But the in-game decisions and constant lineup tinkering are really getting to me this year. I know that these things wouldn’t bother me at all if we were winning, and maybe these decisions aren’t even all that related to the reasons we’re struggling, but it’s something as a fan to latch onto and get very frustrated with. We’re not ever going to be Hargrove’s Indians again, I realize, where he could have photocopied 145 copies of the same lineup on Opening Day and be worried about running out of them, but Wedge doesn’t seem interested this year in giving players a chance to develop, the luxury for (certain) players to not look over their shoulder every time they make a mistake, or putting the players who will give us the best chance to win long term on the field.

Interesting post (I’ll concur with CU Adam), I recall seeing something like this on THT last year. Wasn’t quite the same metrics you used, but was Pythag based.

Anyways, one comment on this paragraph:

It’s immediately pretty obvious that 2006’s -11 performance is probably the main cause of Wedge’s place so low on the list. Remove that year, and his average gets bumped up to -1, barely a half run worse than the average for managers with as much time in as he had. It’s also worth noting that none of the negative years hurt the Tribe’s playoff chances in any way. In ‘03 and ‘04, the Indians weren’t anywhere close to the playoffs, in ‘05, 3 games would have put them at 96 wins, still 3 less than Chicago’s 99, and in ‘06, even adding 11 wins to the 78 the Tribe won wouldn’t have been good enough to crack 3rd place, with 3 other teams in the division picking up 90+ wins on the season.

In 05, another three wins puts us in as the Wild Card over NY/Bos.

In 06, if some those 11 wins were taken from Detroit and the Angels, we end up with the Wild Card. Besides, the law of averages should mean we end up with a few seasons of a positive Pythag, so instead of -11, let’s say we were +3, that puts in the playoffs.

While I agree Pythag isn’t the best way to determine how “successful” a manager is, one would think a team would have close to a zero sum of Pythag seasons. Under Wedge we haven’t.

This obviously is a topic worth batting around, but it remains difficult to get a fix on — one of the best things about this post is that its author is very clear on how unclear all of this is.

I have made plenty of criticisms of Wedge lately, but three quick points to make in his defense — or actually, in the interest of criticizing him with more precision:

First, the cumulative effect of tactical mistakes has been overblown. Many of the mistakes that get cited around here are easily counter-argued at least into dubiousness. The run-expectation defying SB attempts? They almost all come with two strikes and two outs, which is a standard move, and if you had a run-expectation chart that showed bases-outs-balls-strikes rather than just bases-outs, I betcha it would show up as the correct move, because that run ain’t scoring from 1B with two strikes very often! The bunts? Even giving the anti-bunt argument every benefit of the doubt, the negative effect of a sac-bunt on run expectation is MINISCULE, and as we saw the other night, many times a sac-bunt will reduce RUN expectation even while increasing WIN expectation. So I would submit to you that when you weed out the dubious claims and add up the effects based on pure math, this just doesn’t amount to much.

Second, any discussion of Pythagorean issues must consider the effect of the manager on raw production. Every Wedge season until 2008, we have seen a handful of significant players substantially outperform projections in terms of the sum of the individual players’ stats, and the team in sum has outperformed. Looking at 2005, we probably got some 50-60 runs out of Sizemore and Peralta alone, above what was projected, and they weren’t the only ones. So you could make the case, the team outperformed in the raw numbers by 5-10 wins, but then underperformed Pythagoras by 5-6 wins. The manager clearly has a role in both performances, in each case hard to discern.

Let’s talk about that 2005 bullpen for another minute, and here I’ll contradict myself a little. Here you’ve got a bullpen with seven slots, and eight guys performing on a level of some of the top setup men in the game. (Rhodes, Cabrera and Miller were all excellent and essentially rotated through two roster spots.) The eighth-best guy was Sauerbeck with a 4.04 ERA, but he allowed a huge chunk of his runs in blowouts — in April alone, he allowed six runs in three games that we lost by a score of 28-3, and keep in mind, he only allowed 18 runs all season.

So the bullpen was basically cutting run expectation from around 0.5 per inning to about 0.3, all the time. Usually, you’re one or two runs down, you’re putting in your third or fourth best reliever. But if you’re 3-4 runs down, you’re probably using your fifth or sixth best — thing is, though, they’re all pitching the same, so you’ve essentially got high-leverage performance in low-leverage situations. So we’re behind, and we probably can’t win, but with that bullpen and offense, in a typical game, we’re going to score 1-2 runs while allowing 0-1 over the last three innings.

So what’s the result of that? Well, usually, we’re going to lose anyway, because we were behind 3-4 runs and will only close the gap to 1-2. Sometimes we’ll catch up and win, but just as often we’ll take a 5-6 run deficit and — since all the relievers are awesome — catch up to within 1-2. So basically, this is an unusual way to run up a large number of one-run and two-run losses — which we did — which is great for your Pythag record, meaningless to your real record. So while the performance of the top 3-4 relievers should help both the real record more than the Pyth record, the performance of the other 4-5 guys only minimally helps our real record while boosting our Pythag record. And the manager looks like an idiot.

Two things to drill home about that. One, while the performance of our big three — Wickman, Howry and Rhodes — was outstanding and well above the norm, the performance of the rest of the pen was much further outside the norm — on most teams, the bullpen outside of the top three guys won’t do much better than a 5.00 ERA. So that goes to raw performance but not Pythag. Two, it should be clear that the effect of the top guys’s excellence on the real W-L will be much greater than the effect of the bottom guys’ excellence, because that’s the very definition of “leverage.”

I would also point out that, in general, averages tend to skew numbers in such discussions (the classic Bill Gates walks into a room – how does the average income in the room change scenario). I would be interested in seeing the above charts using the median for each manager instead.

Wedge seems, in this case at least, to have an average similar to his median, which would be -3. This represents the typical performance that he has established as a manager and tends to discount the outliers. I would be interested to see how eliminating the outliers would rank the other managers on that list.

Let’s assume that an above avg pen improves a manager’s pythag credits and vice versa. Because pens are so volatile year to year, we should be able to conclude that no manager experienced such a great or lousy 5 year pen run that this span would cease to be an effective measurement. The fact that Gardenhire came out in the top list only serves to boster the notion that the superior deployment of relief pitchers is a huge difference maker. On the other hand, why don’t we see Tom Kelly in the top list? He seemed at least as effective as Gardenhire in managing his pen.