Brawndo the Thirst Mutilat

I had him in third place before this, behind Sabean and Sabean Jr. (Colletti).

Not really. McClung had 250 terrible IP with Tampa. That's actually more than Mota had thrown since leaving LA (before this year). Now, I'm not 100% convinced that McClung has in fact turned it around, but leaving that aside...there is a difference between the two, but it isn't in the amount of performance to analyze. McClung was in his early 20's when he established that he was a terrible pitcher, and is now at his peak. Mota was in his early to mid 30's (or older if his official age is inaccurate). Looking back, Mota really only performed like an elite short reliever for 1 2/3 seasons, from 2003 through being traded in 2004. He has been average to lousy the rest of his 10 year career. The only other time he ever posted an ERA+ above 100 was his rookie year, which was a product of some good BABIP luck. It is deeply frustrating to cheer for a team that is dumb enough to think he might be worth keeping around, for any reason.

What an embarrassingly feeble return for 2 1/3 years of Rauch at under 3 mil per yr. Bonifacio didn't even make Rotoworld's midseason top 150 prospects lists that was just released. He won't hit well enough in the bigs to be a regular at any position other than SS, and he doesn't play defense well enough for that. He doesn't walk and he has struck out an awful lot in the minors for a guy with no power -- in other words, he's exactly the sort of guy that projects as a much lousier hitter in the bigs than in the minors. He's a base stealer. Yippitty Doo. Bowden just boldly leapt to the top of the dumbest GM standings.

I was going to post links to everything I said in the transaction rumors subforum thread before it got locked, but it is just easier to post this link to Jeff Sackmann's trade recap at Brew Crew Ball. Then I can just say, "What he said."

What I cannot figure out is how it hasn't occurred to Yost (or any of the other apparent imbeciles that manage MLB teams and insist on having a pitcher fill this role) that if you don't want to "bury" an average-ish pitcher by restricting his usage to totally inconsequential innings in complete blowouts, maybe the problem is with that usage pattern. Does this mean that you must keep a terrible pitcher on the 25 man roster at all times, because otherwise you're not optimally filling the role that has somehow morphed into nothing more than a way to hide a terrible pitcher? And don't we already have 2 terrible pitchers on the 25 man roster we're only willing to pitch in blowouts? And wouldn't it be nice to have another good short reliever (without further depleting the diminishing supply of prospects in the high minors), so that Villy, Riske, and Shouse (or whoever it turns out at the time hasn't blown a hold recently) don't have to pitch in every single game if you don't have a blowout for a while? I will somehow restrain myself from additional commentary on the manner of statistical analysis driving this decision.

Why not? I mean that sincerely. Has there been exhaustive research into seasons that are so aberrant for established major leaguers that come back and continue to be average or better major leaguers? I don't know of anything I'd call "exhaustive". Sure, guys have fluke bad seasons. Guys also have fluke good seasons. It is somewhere around equally likely (given his age and position) that Ray Durham is now legitimately as bad from a true talent perspective as last season's numbers suggested (thus, this season's goodness is a statistical aberration), as it is that Durham is now legitimately as good as this season's numbers suggest (thus, last season's badness was a statistical aberration). But Marcel isn't doing what you say it's doing -- it's not projecting him to hit as badly as he hit last season -- it is simply splitting the difference between those two equally not-very-likely extreme scenarios. And doesn't that make the most sense? That Ray Durham today is somewhere between how bad he was last year and how good he's been so far this year? I think most people use "is" to describe recent and current performance, given large enough sample to work with. There isn't any doubt that Durham has performed better this year. And most of us expect Weeks to be better. But I don't want to gamble this year's playoff hopes on the regression to the mean of a player that has performed as poorly as Weeks has this year. Then most people are using the present tense when they should be using the past tense. And I don't want to gamble this year's playoff hopes (at a fixed cost of one sorta useful minor leaguer) on a bad bet that Ray Durham will somehow continue to defy Father Time and the laws of statistical probability (in the form of regression to the mean). Not to mention that once you factor in defense, Weeks and Durham have been worth about the same number of runs this year so far anyway.

Maybe I'm off, but I would guess your "scientific" method of determining how a player will perform in a small sample size assumes some sort of pre-season standard deviation for that player and then regresses/progresses that player's YTD stats to reach the assumed standard deviation. The problem is that Weeks is so far off of any assumption anyone would have given him to start the season that he will have to perform well above average simply to reach to lower reaches of his bell curve. In other words, he'll have to play very well simply to have a bad season relative to what was expected of him. Conversely, since Durham has played well after a poor season last year, it is natural that such a projection would assume that Durham would have to fall off significantly over the final 60 or so games to regress to him assumed standard deviation. Actually, he's not too far over his career .787 OPS (which is low due to his poor seasons when he was in his early 20's), but last year would really have screwed up any SABR projections for him. If the projection systems did what you suggest they did, they would be guilty of the gambler's fallacy. They do not. Durham's overperformance so far raises his Marcel for the rest of the year, Weeks's underperformance lowers his Marcel for the rest of the year. But the original gap in their projection was such that not even a gap as wide as has existed so far this season is enough to compensate for it. As for the sample size argument, I am not telling you that these systems (for which I claim no credit) are some sort of perfect crystal ball that tells you exactly how well they'll hit. Nobody would - that would be crazy. Small sample size caveats apply in this direction as well -- anything COULD happen over the course of 1/3 of a season, and obviously no one knows what WILL happen. All I'm trying to say is that there are better informed ways of deciding how good a player currently "IS" than simply looking at this year's numbers and expecting them to continue. That's all these projections are for.

This is stated far too definitively. Marcel is great for large pools of players, but it doesn't know how to account for some things. Durhams year last year was clearly out of line with the rest of his career. I would think his season so far and his career should be weighted far more than the aberration of last year. You can't simply throw out an entire season worth of data. Maybe it was an injury-ridden, struggle-filled campaign. But Durham is old. And he has a lot of miles on the tires. That kind of thing is par for the course for players in his age group (not to mention 2B age worse than any position other than catcher), and is very relevant in projecting his future performance. Marcel does care about how old a guy is. You're right that I probably needed some qualifiers in there -- 'probably's and 'best bet to's and the like -- but nobody bothered to use qualifiers like that in the dozens of posts in this thread that quoted this season's hitting stats and used the present tense in declaring how much better than Weeks Ray Durham "is". We can't know true talent in a 100% definitive sense -- our best guesses are always at their core still guesses. But there are better and worse informed ways of making such guesses.

I'm assuming these projections don't factor in the likely scenario (based on his numbers) that Durham won't be playing against lefties. Is there any way to factor that into the projections? Sure, at least approximately. Checking his career platoon differential out at B-R, it seems reasonable to add 10-15 points of OBP and subtract 15-20 points of SLG, if you assume they'd use him exclusively against RHP. He's basically the same hitter from each side of the plate except for this small difference, which is essentially a wash anyway. Half-year platoon splits are worthless for projecting performance when you have all the additional info you have for Durham. He was better against LHP as a righty in 05, 06, and 07. I didn't feel it was necessary to go back any further than that.

Ask and ye shall receive. I used Sal Baxamusa's Marcel calculator (found here) for all the relevant big leaguers who would be affected. Here are their projected performances for the remainder of '08: Durham: 259 / 335 / 412 Weeks: 247 / 353 / 409 Dillon: 271 / 348 / 413 Durham still projects as the worst hitter of the three. Yep, even worse than Dillon. Marcel doesn't know about park factors, so Durham may get a slight boost moving from AT&T to Miller Park, but it is doubtful that it would be enough to make up the difference, especially since OBP is where he's hurting. Everyone who thinks he'll fix the team's OBP problems, please note that. Now, a Marcel projection isn't the best way to project someone like Dillon's performance, as he doesn't have very many ABs so his projection is regressed very heavily to the mean. But if you used ZiPS or CHONE or any other smarter projection system (one that filled in statistical gaps for guys like Dillon with MLEs from their minor league performance), he'd look even better (ZiPS pre-season was 280 / 338 / 449, CHONE was 280 / 356 / 492. 80 PA this year wouldn't change those projections too significantly - I guarantee that they'd still be higher than the above Marcel.) Assuming this deal does go down, both the guy losing playing time to Durham and the guy losing his roster spot to Durham will be superior hitters to Durham, even considering his resurgent performance so far this year and the respective struggles of those other guys. As for defense, Weeks is rather clearly better. Dillon is probably comparable as a bench defender. Somebody already brought up Durham's woeful RZR and OOZ numbers - it is likely that his range is totally shot. There is no way he comes out as a better overall regular player than Weeks, or a better bat off the bench than Dillon, except if you only care about this year's hitting stats. If that's how you like to project performance... This is a nothing move. If anything, they are adding salary to make the team slightly worse. I don't care one iota about Ford being traded, but I like Hammond and I think he has a chance to somewhat useful in the majors (if only as a long relief guy) beginning right now. He is easily the best pitching prospect in the Brewers high minors (not that there's much competition -- but that's another reason not to trade him). I hate this move in all respects. The only genuine upgrade to Weeks who might be available at a reasonable enough price to justify the move (Roberts = too exepensive) is Mark Ellis, and there's no guarantee he's actually even available (latest word was the A's were talking extension, and they're still in the race).

Nothing. We're pretty likely to make the playoffs as is. Deadline deals are virtually always sound and fury, signifying nothing (other than a pointless loss of young talent). Most of the moves being kicked around on this subforum right now are for players that would either be prohibitively expensive (given the modest improvement they represent over a third of a season), or for players that only look like improvements if you think that the best way to project performance for the remainder of the season is simply to look at this year's partial season stats. We already got the crown jewel of the available players, a month early, for the best prospect in the system. Be content with that, lest you get what you asked for in the form of Linebrink trade v. II (this time featuring Mat Gamel!).

The whole thing just goes to show the capacity of the human mind for cognitive dissonance and overlooking the obvious. I'm as guilty as everybody else, heck, I was complaining about this phenomenon as recently as Sunday in the in-game chat. But looking at the data, and just thinking about it calmly and rationally for a minute, it's pretty obvious that it's a chimera. Ask any baseball fan about how the lousy pitchers on his own team perform on a start by start basis, and s/he'll complain about the "inconsistency", meaning he's good or ok for 3 or 4 starts, but then he'll just get lit up horrifically. All it takes to understand why the Brewers seem to underperform against lousy pitchers on other teams is to turn that perception around, and realize that their lousy pitchers aren't any different than our lousy pitchers.

Sure, just look at RLuz's spreadsheet and count. Not sure how wide you think the "expected range" should be, but they're above the opponent's projected ERA 8 times and below it 11 times. If you use, say, 1 run above or below as "within an expected range", that only happened 3 times out of 19 starts. I get 6 "above expected", 3 "within expected", and 10 "below expected". But, as has been pointed out several times previously in this thread, this is precisely what you should expect, and a totally meaningless finding. It would only be significant if start by start ERA was normally distributed around the starter's mean ERA, which it isn't. Check the game logs for a couple of those bad pitchers and you'll see that their opponents would be "below expected" for MOST of their starts, and their ERA is only as high as it is because of a few times when the opponent "destroyed the guy" which "pulled the entire number way up". Maybe that's what you were getting at -- not sure. Mark Hendrickson gamelog ER by start- 6, 1, 1, 2, 3, 2, 3, 1, 4, 1, 5, 10, 4, 5, 3, 4, 5, 6, 4, 2 Livan Hernandez gamelog ER by start - 2, 4, 0, 3, 4, 7, 1, 1, 3, 3, 5, 6, 5, 7, 7, 1, 1, 4, 3, 6http://www.baseball-reference.com/pi/gl.cgi?n1=hendrma01&t=p&year=2008 Both of them gave up 3 or fewer runs in at least half of their starts (10 of 19 for Hendrickson, 10 of 20 for Livan that would be "below expected"). That leaves just half of the starts to divide between "within expected" and "above expected". 4 or 5 runs might be either, depending on how many IP, but I've put in boldface the only starts that are obviously bad ones (3 of 19 for Hendrickson, 5 of 20 for Livan). This distribution of performances looks EXACTLY like the Brewers performances against bad pitchers from RLuz's spreadsheet.

Per the bottom line ticker on ESPN, Carlos Marmol is an All-Star, replacing Kerry Wood. Heaven forbid if half the players on the NL roster aren't Cubs!

Inman is filthy. They were talking to Melvin during his inning about trading LaPorta for Sabathia, and it fell to Doug to point out that he had traded Inman as well. Surprisingly low "and that" count -- unless I missed some, I heard just one.

Josh Kalk looked at this over at Brew Crew Ball. It turns out that what you describe is very close to what actually happened, except that it was 2 starts ago that his velocity was way down and then it returned in his last start with Oakland. According to Pitchf/x, his horizontal movement was down and his vertical movement was up along with the lower velocity 2 starts ago, so Mr. Kalk hypothesizes that his arm angle was just out of whack.

It's only hard to save the best guys for the one run deficit situation because you've presupposed all of the fully LaRussa-ized modern bullpen roles I so utterly despise. Now, every modern manager seems to use them, which is why I put this in the vent thread rather than the Yost thread, I guess. But you're absolutely right that the first run is the most damaging when you're down 1 run. That's why it's a perfect spot for your relief ace (closer), since he's probably the guy who is most likely to have a clean inning. A 3 run lead, any 3 run lead, whether the 8th or the 9th, is a much much better spot for a "medium resources" guy, because the first 2 runs he gives up don't cause a terribly dramatic change in your odds of winning. That's quite a cushion for a mediocre reliever. Maybe not for Guillermo Mota. We're in agreement there.

I always found this to be an interesting idea -- that we might be able to measure manager quality by looking at win-loss record relative to Pythagorean expectation. I'm not convinced it's a good way to try to assess managerial competence. In order to be responsible for his team outperforming Pythagoras, the manager would have to be able to affect how runs scored and runs allowed are distributed from one game to the next. He might be able to a little bit -- for example, by using the best relievers in close games and the worst relievers in blowouts, you might win more than your fair share of the close ones and make your blowouts more extreme (then again, don't all managers do that?) -- but it seems to me that the lion's share of the credit has to go to the gods of random variation. A more interesting avenue, and one that to my knowledge hasn't been researched thoroughly (if at all), is whether you can pick up any sign of managerial competence by comparing a team's expected runs scored and runs allowed given its individual offensive events to its actual runs scored and runs allowed. In other words, look at BaseRuns and BaseRuns allowed, or EQR and EQRA, vs. actual runs and runs allowed. I think that most of the managerial decisions we evaluate -- when to sacrifice, when to steal, when to hit and run, when to pinch hit or not, lineup order, etc. -- would have a much greater chance of causing noticable effects here than in pythagorean record. BaseRuns and EQR tell you very accurately how many runs a team should have scored if all of the offensive events had happened in a vacuum. But they don't happen in a vacuum, and a manager is responsible for a lot of stuff that, if done well, might allow a team to squeeze extra runs out of it's component offensive stats. I still think that it's mostly the gods of random distribution -- you're going to score a lot more runs if you have 9 hits in one inning than if you have 1 hit an inning for 9 consecutive innings, and the manager can't do much to control that -- but I'm less sceptical that a good manager's skills could show up, given a large enough sample, using this kind of analysis. Half a season isn't such a sample, but for what it's worth, we're only the slightest bit short of target: 437 EQR vs. 430 actual, 410 EQRA vs. 411 actual. A grand total of 8 runs shortfall, or maybe 1 win. ................................................................................. Oh yeah, a grade for Yost...I'll give him a D. Practically every hitter on the team is underperforming this year (even the ones that made the AS team are doing worse than last year). It's not totally fair to lay that all at the manager's feet, but I do think that the "kids" have found the going much tougher once the pressure was on, and I blame Yost for failing to keep them loose and confident. In fact, given his personality and method of leadership ("It's my first time in a pennant race too"), I can't imagine anyone staying loose and confident under that guy. And what did he do when he felt like the team needed to blow off some steam? Oh yeah, he brought in Larry the Cable Guy to take batting practice and amuse the guys with his hilarious brand of revoltingly insipid dreck. Every 20-something I know that likes hip hop enough to pick the walk-up songs these guys picked just LOVES Larry the Cable Guy.

In an ideal world where there were no concerns about Parra's innings, you send McClung. If they're going to pull Parra from the rotation at any point, however, now is the time. Whoever is sent to the pen, I just hope (probably in vain) that Yost finds a more creative and useful role for him than the one Mark DiFelice used to have and apparently Villy has now.

Two excellent Terry Gilliam kids films: The Adventures of Baron Munchausen Time Bandits An offbeat personal favorite lots of other folks hate, but I loved as a kid and still love now: Joe vs. the Volcano

My point isn't that Villy wouldn't be available tomorrow, but rather that Yost won't use him tomorrow (or any game) unless the game is out of hand. DiFelice was the long man before he was sent down. How many close games did he see? Villy is the long man now.

Villy is the "long man". I used to think that meant he would pitch in situations where the starter leaves early-ish, or the pen is tired, and you need a couple of innings, but it turns out what that actually means is "pitches exclusively in blowouts". What a useful, constructive way to use a roster spot! And what a good spot for one of your better young pitchers!

It's late and I'm a little drunk, so I think that might be a bridge too far for me right now, but Fangraphs has average fastball velocity data for the season, and that's easy enough. For the whole season, Harden is averaging 92.6 mph. Last 14 days: 92.4 mph Last 7 days: 91.6 mph So there's some dropoff, but according to Fangraphs it's nothing like 5 mph; in fact it doesn't look like it's enough to be a significant concern. That doesn't mean he won't turn out to already be too injured to make more than a couple semi-effective starts before they have to DL him. Or, he could hold up the whole season. I guess I'm saying you're better off trying to figure that out with Tarot cards or a Ouija Board than the Fangraphs data. Maybe Pitchf/x sees it differently.

Strange move from Tom H., suggesting a roster move based on nothing other than what his gut was telling him during the post-game presser. Maybe he's feeling all full of beans after "breaking" the Sabathia trade (he didn't break it, really, but he was probably the first reputable MSM guy to say that it was finalized and definitely going down), and has decided he can become Milwaukee's version of Jim Plaschke, the L.A. Times guy that ran DePodesta out of town and has become essentially a de facto co-GM to the new regime. Now the Brewers almost have to get rid of Mota. Tom H. said it was going to happen. All the fans want it to happen (fans have notoriously short fuses for lousy performance). Even if they want to keep him, or feel compelled to keep him to avoid the scenario outlined by FtJ above, they're not in a great position to do so. I'm opposed in principle to this manner of malarkey, but I'll be glad enough to see Mota gone to look the other way this once.

Regarding the use of Mota tonight -- I think it is pretty clear from his recent usage pattern that Mota HAS lost his "8th inning guy" job. We didn't have the lead tonight. Whoever the "8th inning guy" currently is (my money is on Gagne), that guy only works when there is a save / hold situation to be preserved. My vent concerns this philosophy and its complete and utter ludicrousness. Using Christopher Shea's neat Win Expectancy Finder (link), let's compare a situation where the Brewers would have been trying to preserve a 3 run lead to tonight, where they faced a 1 run deficit. First, starting the top of the 8th with a 3 run lead: Runs allowed in 8th..........Chance of winning 0.................................................98% 1.................................................96% 2.................................................89% Now, with a 1 run deficit: Runs allowed in 8th..........Chance of winning 0.................................................30% 1.................................................15% 2.................................................7% Which of these two situations seems to you to call for the better pitcher? Why, in the name of all that is holy, does Ned Yost believe that the superior short reliever should be used in the case of a 3 run lead, while the designated bullpen millstone (even Yost has to think that's what Mota is right now) is the right guy for the 1 run deficit? My other vent -- What kind of post in this thread requires research? I can't even vent right.