premium blogad:
blogads:
advertisement:
sitemeter:
|
|
|
February 27, 2007
The Pitchers Are Alright
Hey, did you guys hear? The Mets rotation is too old! And while we’re at it, it’s too young, too! And inexperienced! And mediocre! I don’t know if you all have realized this, surely this hasn’t been covered ad nauseum by the media, both local and national, radio and television! Just thought I’d spread the news! It seems every day there’s at least one article — and typically, there’s about three or four — written about how the Mets rotation isn’t up to snuff. Oh, sure, the lineup’ll be swell and the bullpen’ll be terrific, but those Mets are sure to have a hard time winning relying too heavily on veteran/rookie/lousy/fill-in-the-blank pitchers. Perhaps these writers don’t remember the Mets winning 97 games with a rotation that was, at one point, an injured Pedro Martinez, Tom Glavine, Steve Trachsel, Jose Lima, and Geremi Gonzalez. Maybe they don’t remember the Mets winning 97 games while starting thirteen different pitchers last year. They surely don’t remember the Mets winning 97 games without ever having a set rotation at any point last season. If they did remember that, they probably wouldn’t worry all that much. They’d probably recall the fact that the Mets gave Mike Pelfrey four starts of 5.48 ball, Dave Williams five starts of 5.59 ERA ball, Victor Zambrano five starts of 6.75 ERA ball, Oliver Perez seven starts of 6.38 ERA ball and Alay Soler eight starts of 6.00 ERA ball. That’s twenty-eight games with a combined ERA of 6.04 ERA — and you know what the Mets’ record was in those games? 15-13. Yes, despite the fact that those pitchers pitched twenty-eight games — seventeen percent of the season — with an ERA over six, the Mets still managed to play two games over .500 baseball. And that’s not including the aforementioned Lima/Gonzalez starts, or the fact that the Mets turned Steve Trachsel, a man with a 4.97 ERA, into a fifteen game winner. The fun thing about Spring Training is that the season hasn’t begun yet, and everything is started anew. The miserable thing about that is there really isn’t too much to discuss, so attempting to write about a team can be difficult. That’s why the biggest story out of Mets camp so far is that Cliff Floyd, now a Cub, thinks Willie Randolph was “confused” in the playoffs. And across town, all that mattered for week was that A-Rod and Jeter are not BFFLs anymore. With nothing much to focus on writers tend to grab a hold of three or four concepts early and milk them dry through February and March, whipping most casual fans into a frenzy. But some closer observation can sometimes disprove these thoughts. The fact of the matter is this — the Mets had some pitching issues last season, and they overcame them splendidly with a terrific lineup and a nasty bullpen. Now, in 2007, they have some pitching issues… to go along with a terrific lineup and a nasty bullpen. So, why should there be any more concern about it this year? Right now the rotation is Tom Glavine, Orlando Hernandez and John Maine, leaving two spots in the backend to be filled accordingly. Fighting for those spots are Chan Ho Park, Phil Humber, Mike Pelfrey, Oliver Perez, Jason Vargas and Aaron Sele. Six guys, all with track records of success on various levels, to fill two spots — and we’re supposed to expect them to do worse than a 6.04 combined ERA from the schlock the Mets threw out there last year? Heck, the guys fighting for spots this year combined for a 5.22 ERA total in ‘06 — if we had that last year, that’d mean over 100 wins! Realistically, nobody can say the Mets have a good rotation at this point, because it’s just not true. As a matter of fact, they don’t even have a rotation at this point, unless it’s 1890 and we’re expecting Tom Glavine to go Charley Radbourn on us and start sixty-eight games. But what they do have is potentially far better than what they had last year. Are there some older fellows? Sure, Glavine’ll be forty-one come Opening Day, and Hernandez is anywhere from thirty-seven to fifty years old depending upon who you ask. But — as Roger Clemens, Curt Schilling, Greg Maddux, Jamie Moyer, Tim Wakefield, Woody Williams, David Wells, John Smoltz, Randy Johnson, Kenny Rogers and Mike Mussina have taught us — age only matters if you suck. Luckily for us, Glavine (15-7, 3.82 ERA) and Duque (9-7, 4.09) do not. Are they inexperienced? Yeah, but everybody in the game is inexperienced at some point. It’s what you do after your development year — or, in some cases, years — that really matters. Last year was a chance to prove he could do it at the big league level for John Maine, who allowed only sixty-nine hits in ninety innings of work, to go along with one of the best fastballs in the majors. It was a shot for Oliver Perez to start re-learning his delivery and develop a gameplan from one of the best pitching coaches in the game in Rick Peterson and, though the final numbers don’t look too spectacular, did you know from September on Oliver Perez went 39.9 innings with a 4.28 ERA allowing only thirteen walks while striking out thirty-nine? It was also a year in which Pelfrey and Humber combined to go 12-6 with a 2.77 ERA with 188 strikeouts throughout the Mets minors. And those numbers came from one guy coming off Tommy John surgery and another who threw a pitch all season he decided to scrap after the year. Is there the possibility that the Mets will throw some lousy retreads out there? Of course, but unlike years past these lousy retreads have a chance to actually be good — or at the very least serviceable. Park is still only thirty-three years old, healthy for the first time in years, and but a scant four years removed from a career so stellar it landed him a sixty-five million dollar contract back when that number meant something. Frankly, it’s just as conceivable he’ll be terrific as he’ll be bad, and if he does perform terribly it only costs the Mets some loose change to find that out. Sele’s a notorious fast starter, 96-52 with a 4.35 before the All-Star break for his career, and was terrific for the majority of his time with the Dodgers last season. The important thing to keep in mind, however, is that neither of these guys will be leaned on heavily. The season will not be made or broken by how well Aaron Sele pitches. The Mets’ pitching will get this team by. True, they’ll never be compared to the Tigers’ or Red Sox’ rotations, but they’re not going to be smacked around on a nightly basis, either. This season will depend on what a season typically depends on — the offense. Does good pitching win ballgames? Absolutely, but despite the old adage to the contrary, good hitting will typically beat good pitching, and the Mets’ offense is the best in the National League, possibly all of baseball. You think the Mets’ lineup is gonna be afraid of Freddy Garcia and Adam Eaton? Mike Gonzalez and Rafael Soriano? Henry Owens and Matt Lindstrom? The hitting will be there. The fielding will be there. The bullpen will be there. The rotation? Relax, it’ll be there. 66 Responses to “The Pitchers Are Alright”Leave a ReplyYou must be logged in to post a comment. |
|
Pitcher by commitee worked a bit last year until the summer when A. We get a healthy Pedro back. or B. Humfrey tear up minor league ball or C. Milo tears it up and Omar trades him for Dontrelle?.
ROFLMAO! I want the bumper sticker…
great piece andrew
Great piece, Andrew. It encapsulates what I’ve been saying to people for a couple of months now, but with the stats to back it up. This piece should be required reading for every dooom-saying Mets pundit in the NY area. Bravo!
The one statement I would take issue with, however, is that good hitting beats good pitching, aphorisms to the contrary. In fact, I firmly believe that good hitting beats decent or mediocre pitching — but great pitching beats great hitting. Of course, there are only la handful of pitchers capable of pitching consistently great games, which is why this isn’t a problem during the regular season, but is definitely an issue in the playoffs (or ask the Yankees with that killer lineup they had for several years after 2000 that didn’t win a single WS because great pitching shut em down).
But apart from that — and noting that that has little significance in regular season anyway — everything else is spot on.
I agree with the argument, but the one thing that does trouble me is that the Mets managed to wring 82 Quality Starts from their 13 different starters.
Pedro, Glavine, El Duque and Maine provided 56 QS (~60% of starts), while the rest of the staff provided 26 (~40%). I think it’s going to be very tough to match those numbers without some serious improvement on the back end of the rotation.
Yea no doubt. I would like to add look at the Dodgers in 1988. Sometimes one good pitcher can beat good hitting. I know im still hurt over it too :-(
I sort of look at that as a positive. In almost half the games, we did not get a quality start from our pitchers — and we still won 97 games. To me, that shows how relatively unimportant SP can be.
You might try another calculation too — how many QSs did we get from the five we broke camp with last season? (Pedro, Glavine, Trachs, Zambrano and Bannister) Just shows that who you start with doesn’t necessarily tell the whole story for the season.
Exactly!!!
So out of 162 games played last year we had 82 quality starts. Roughly half the games we played was a quality start. Im not sure thats any good at all. We should do better than that this year I would hope.
Actually, 82 QS is pretty good.
Here’s the QS for each of the playoff teams last year. It’s easy to see why we would’ve been considered favorites, had Pedro and El Duque been healthy.
SDG - 90
DET - 88
LAD - 81
NYY - 79
OAK - 76
STL - 74
MIN - 73
I definitely agree with this statement.
Wow thanks phl, I wouldve never thought that. I gotta keep reminding myself its a different game today, relief pitching has basically become way more important.
Andrew, you should email that article to a few of the writers who have been doom-sayers and see their reaction…. ( please include Steve Phillips.) I’d be curious to see any responses and their arguements against your statistics.
PHL, you would be shocked the number of times the Mets hit right at the QS metric, so I think that is a little misleading. The overall scope is a little more important I think.
The Mets starters last year ranked 23rd of the 30 major league teams in innings pitched. They were the lowest of any team in baseball that made the playoffs, and were just in front of such luminaries as Tampa Bay (24th) and Baltimore (25th). If you just look at the National League, the Mets were 14th out of 16 teams, only in front of the Nationals and Cubs. They leaned heavily on the bullpen last year.
Tne Mets were right in the middle of the National League in ERA, 8th out of 16. But they were only .06 ahead of the 11th place NL team in ERA (Milwaukee). Considering the pitcher’s park they play in, it’s safe to say they were really below average in terms of overall ERA as well.
Mets Bullpen ERA: 1st in NL, 3.25
Mets Bullpen IP: 3rd in NL, (of course, only behind Washington and Chicago)
The Mets starters lead the National League in wins despite their below average impact. The Mets relievers also led the National League in wins.
What does this all mean?
The bullpen didn’t blow leads, kept the team in games when they were behind, and the offense scored runs. The starting pitchers survived. That was the formula last year. And that will be the formula again this year.
I don’t disagree with any part of that statement. Actually, my argument is that 4 of 13 starting pitchers did the bulk of heavy lifting in terms of keeping us in games through the 7th inning. Of the four:
1. Pedro will be unavailable until … ?
2. Glavine will be awesome 70% of his starts.
3. Maine is expected to regress, but we’ll see.
4. El Duque will be awesome, too, but we don’t knowhow many starts he’ll give us.
So, our starting pitching gave our team excellent chances in more than half of the games played last year, which might be counterintuitive to many fans. Expecting the same from our current prospective rotation might be presuming too much at this point.
But the point is, you don’t know what to expect from the rest of our starting pitchers. Is it unreasonable to think that Park and/or Sele/Vargas/Humber/Pelfrey are going to pitch better than Lima/Gonzalez? I don’t think so. I would expect the best of that group to pitch better than Trachsel did last year. So imo it’s not unreasonable to expect additional QSs from our revised staff this season.
I don’t understand why it is presuming too much at all. Glavine will be roughly the same.
El Duque will make more starts for us this year. And for all the talk about his injuries, he made 29 starts last year. There is no reason not to expect 25 starts or so from him.
Pedro was great for 11 starts in April and May. Pedro has a good chance to be great for us in August and September this year. And Pedro was terrible after May because he was pitching injured. Terrible. His post All-Star break numbers were 31 IP and a 7.81 ERA.
Maine will probably be more inconsistent than last year. But even if he regresses, he will have 15 starts as good as his 15 starts last year, and then his other starts just have to be better than Lima and Geremi Gonzalez. Not too much to ask.
As much as I am not a fan of Chip, his production will probably be better than Trachsel’s. He will at least eat more innings than Merlot did.
Oliver Perez is the real wildcard. Andrews stats about Ollie and what he did from September on are very telling. This guy was lost, but now found. Was blind, but now… you get the point.
all i know is I want Glavine to throw his curve lots and lots this year. in fact i think he needs to!
Great article Andrew. Full of great information and points and it cracked me up! :)
Put another way:
Let’s assume that Glavine, El Duque and Maine combine to provide 47 QS in 85 starts–55%, which should be a reasonable but conservative estimate. In the remaining 77 games, the starters would have to pitch in 35 QS, or 45%. That’s gonna be tricky for the back end of the rotation, unless Perez is more Glavine-awesome than Trachsel awful. (Trachsel, the 15-game winner, provided 13 QS in 30 starts.)
The good news is that Perez in 2004 was Glavine-awesome, providing 21 QS in 30 starts.
The bad news is that Perez in 2006 was Trachsel-awful, with 7 QS in 22 starts. (Only 1 of 7 with the Mets!)
I don’t believe that QS is a strong predictor of win percentage. However, I do think it is a pretty intuitive way to measure how much weight your starting pitchers are pulling.
The QS metric is pretty useless when assessing Steve Trachsel.
Guess how many times Trachsel pitched to a batter in the eighth inning last year? Zero.
He completed 7 innings only 3 times.
Boy, good thing that QS metric is only 6 innings.
Eh, despite devoting quite a bit of time and words to this, I don’t feel very strongly about this. I do think this year will be tougher, but that’s about it. The sky is not falling.
We’re more or less looking at this the same way. I’m a little skeptical, because I’m discounting Pedro contribution this year altogether and I think we’ll miss his starts.
Really? Based upon what, a 4.09 ERA last year? Sorry, but for every outstanding game Duque pitched last year he threw in a stinker. I recall more than a few comments here and in the MSM questioning whether his level of concentration was consistent, meaning whether he tried as hard in, say, a July game in Milwaukee or Pittsburgh as in a postseason game. Just look at his performance in Arizona before the trade. I expect Duque, over the course of the season, to even out somewhere between adequate and above average, but “awesome” is just not in the cards (at least not until October, I hope)..ie
PHL, I completely understand where you are coming from. The Mets rotation could definitely fall on its face.
But last year’s was not a good rotation. It was a below average rotation and they won 97 games. That is my overriding point.
I think this team can at least recreate a slightly below average rotation.
Actually, from a wins perspective, 12 of 20 QS is pretty awesome. Especially when he pitches deep during good starts, and gets pulled early during poor ones.
I think the real issue is the offense. Starting pitching will be about average. The bullpen will be better than average, but the real issue will be whether the offense can perform at the same level as last year.
So much talk has been focused on pitching with the assumption that the offense will be as good as last year - but in my mind there are as many or more questions about the offense - and it’s an area Omar could have done more about in the off-season.
Consider the Mets had career (or near career) offensive performances from many players including Jose Valentin, Endy Chavez, Paul LoDuca, Xavier Nady and even Carlos Beltran and Jose Reyes. While it’s true that Moises Alou’s addition strengthens the line-up, imagine what will happen if Valentin, LoDuca or Beltran regress more towards their mean performance. The result could be many fewer leads for the bullpen to hold.
What was so difficult to watch this off-season was Omar’s lack of moves to improve the offense. When it became clear that pitchers were going to cost more money than he was willing to spend (and he drew the right line in the sand), he showed no flexibility in using his extra dollars to further bulk up the offense.
He re-signed Valentin to a $4mm contract which essentially percluded him from upgrading that position (either Ray Durham or Marcus Giles would have been a better more reliable procuder to sign). Also he didn’t address RF with a good veteran right handed platoon that would have really helped the team (Jose Cruz Jr. a.300/.400/.500 right handed hitter with a former gold glove would have been an excellent choice and cost only $700k for a one year contract - instead we got Ben Johnson.) And finally when you look at the composition of the bench - it looks particularly weak (save Chavez) with Newhan, Easely, Castro/Alomar, Franco and Johnson - none of those guys really seem strong to be good platoons, pinch hitters, or potential starters should the front line guys go down.
So as we enter spring training I worry a lot more about the offense than the pitching. To me it felt like we were a bit “lucky” last year to get everything we got. In the end I can only hope that Omar’s selection of offensive talent has as great of a year this year as his selections did last year.
i am surprised that no mention is made of aaron heilman and his 2007 role in this otherwise well researched, thoughtful article…
davidg, I couldn’t disagree more.
Delgado’s all patched up, Wright’s removed from the HR derby, Reyes is bigger and stronger, Alou’s adding right-handed punch, the kids are a year older and the veterans don’t seem like they’re regressing. If Valentin is a question mark at 2B, what would you call AHern, who was the starter going into 2006? The offense is going to be at least as good, but I think better.
And on the bench, is Newhan worse than Chris Woodward? Is Ben Johnson worse than Eli Marrero/Ricky Ledee/Michael Tucker? Is Easely worse than Matsui? No and no and no.
Agreed 100%, Lucien.
I have to strongly disagree with anyone who argues, “Our rotation sucked last year and we won 97, so it can suck as much as it wants this year and we’ll still win games.”
Bad starting pitching has destroyed many otherwise-solid teams’ playoff hopes. Many. Many many many.
Now, those who argue that our starting pitching probably won’t be all that bad, I personally agree with. But I think there is a not-insignificant chance that it will be that bad, and thus I can’t fault all the reporters for going back to it.
Oh, and if the offense, defense, bullpen and rotation all perform like last year, we should expect to win 91 games (a la last year’s run differential), not 97. Teams don’t generally repeat a +6 over “expected wins” 2 years in a row.
It wouldn’t surprise me if the Braves or Phils find a way to win 91, so I suspect our rotation may have to be better than “tolerable”.
I think we’re in the thick of it in August and then we pull away when Pedro comes back. :)
If that was directed towards my argument, that is not what I said. I said if the rotation performs slightly below average again this year, and the offense and bullpen remained static, the Mets would be just as good. My argument was that the rotation does not have to be above average to repeat 97 wins, it just has to be close to average.
A team with a good bullpen and offense will always out-perform their run differential in wins. Every single time. If you are deficient in one of those two, then no. But if you are excellent in both, you will. You’re going to win more close games with a top-notch bullpen and offense.
Now, you can argue that the bullpen or offense may not be as good and therefore the Mets will win less games with similar starting pitching, and that would be accurate. But if the bullpen remains a quarter of a run better than the rest of the league and the offense is a top 3 or 4 offense, the Mets can get by with below average starting pitching.
Wasn’t directed toward you, Danny. More toward an argument I’ve heard many times from various sources… and I think a similar logic is implied in Andrew’s article.
Er, wait, is there some research on this that I’m unaware of?
Or is this your opinion?
If it’s your opinion, as far as I know, you’re dead wrong. Bullpen is indeed a potential factor in team over- or under-producing their expected wins… to the tune of “not much”. Offense doesn’t even factor into it.
About a year ago I read a very detailed analysis which rated teams’ over/under-performing expected wins as attributable to:
1) LUCK by a mile
2) Bullpen maybe, probably a little
3) Manager maybe, possibly, hard to say
Since then, I’ve read a few shorter pieces that have said pretty much the same thing. I think one of them was by Neyer, I can’t remember for sure.
If that’s what you got out of the piece, you may need to read it again. I feel I made enough points to come across on the positive side for the Mets rotation next year. I think they’ll be fairly strong. Glavine’s a dependable number one, I think Maine’s got the stuff and the ability to win 15+ games next year, and I think Chan Ho Park — a guy who guy went 80-54 with a 3.80 ERA in his Dodger career, averaging 213 innings a year as a full-time starter — will be a big surprise for this club. He’s still young, he’s still healthy, and Peterson will help him get his mechanics back in whack.
If El Duque can stay relatively healthy, and Oliver Perez can perform as he did after the calendar turned to September, then the Mets have a shot at having a very good rotation. Even if one of those things doesn’t happen, the rotation has a shot at being solid.
Nowhere, however, do I say that the rotation will suck but it’s all gonna be okay because the offense is good. I’ve seen enough Texas Rangers game to prove the contrary, ;-)
I don’t buy this…never have.
I haven’t seen a study, and I wonder if argon or anyone else has, showing how a teams performs vis-a-vis its “expected wins” based on the quality of its bullpen.
Commonsense and a rudimentary knowledge of statistics lead me to believe that if a team has a terrific bullpen, it will win a lot more 1-2 run games than it loses, which will skew its “expected wins” to the positive side.
What I would like to see is the “expected wins” vs. real performance for the team with the top 3 bullpens (ERA wise) in the league. My gut tells me that they almost always out perform their “expected total.”
That being said, with the Mets bullpen likely to be as good or better than last year (a healthy Wagner and hopefully a healthy Sanchez all season) the team is likely to outperform its expected win total again this year.
Furthermore, I would expect the offense to be better because of the HUGE addition of Alou to hit behind Wright, a healthy Delgado, and (hopefully) continued maturation and improvement from Wright and Reyes. Plus, I think either Green by himself or a combination of Green/Johnson/Milledge will put up much better offense from RF this year.
Argon, I’m not so sure. My mental model of the PythWin calculation likens it to one season-long running score, with wins and losses tallied by randomly sampling the score at the end of an inning 162 times. If that model is correct, then you should expect PythWin to come up short when there’s a substantial difference in starting vs. relief ERA–i.e., there will be discernibly patterned run clusters every 9-ish innings.
(For the purposes of this thought experiment, offensive production is presumed to be evenly distributed throughout innings and thus a non-factor.)
I would like to read this if you know where it is, because it doesn’t make sense to me.
Say you have the best bullpen in the lead.
Your starter leaves in the 6th inning and the game is tied. You will win many, many more of those games than you will lose, and many times by one run (especially the walk off type of wins).
Logic holds that the fewer runs your bullpen gives up, the more one run (and closer) games you will win. The more one run games you win, the more likely you will outperform your expected wins total.
If your bullpen sucks, you under perform for the same reason.
Let’s look at the last two Mets teams.
In 2006, arguably the best bullpen in the league, over performed expected wins by 6
In 2005, crappy closer, weak bullpen, and IIRC, didn’t they slightly underperform?
(Although 2005 was skewed heavily by that insane series in Arizona where they outscored the Diamondback by 30 runs in 4 games).
I say a rotation of Glavine,Duque,Maine,Park,Sele and Perez as the long man would be great. Sele gets off to a good start as usuall and is replaced mid season by Pelfrey. Humber comes in of anyone gets hurt after a couple months. If someone goes down early stick Perez in as starter. Perez was great for 4 innings just about every game last year so long man would suit him well whether he goes 3ip or 4ip in a spot.
His five hit complete game shut out of the Braves in September was the single best game and Mets pitcher threw last year.
Factor that in with his Game 7 performance, and I think Ollie gives us a sub-4.00 (closer to 3.50) ERA.
I still say Maine+Ollie win 29 games.
Andrew, I know you didn’t outright say “our rotation can suck and it’ll be fine!” That’s why I said “implied” — and I wasn’t referring to your whole article, just the part that goes “what happened last year . . . what we can expect to happen this year.” It’s that latter kind of thinking that I’m trying to caution against.
I actually share your optimism regarding Maine and Park. And I do expect Ollie to figure things out enough to be above-avg eventually (though likely not by April).
I see no reason why the mets cant outperform they’re pythag again.
Are we also gonna lose a ton of 1-run games this year cuz we were so “lucky” last year and everyone kept saying. I dont buy it one bit.
I also think its a silly little calculation that ppl put way too much weight into and then hide behind luck anytime someone suggests that maybe the way the team is built can affect win totals more than a simple calculation based on RS and RA.
How many 5-game sessions did the Mets have the same 5-man rotation?
How many 5-game sessions did the Mets have the same 5-man rotation?(in a row)
I think y’all anti-Pythag people are approaching it incorrectly, pretending that “Runs Scored vs Runs Allowed” covers less ground than it actually does. E.g.:
Yes, you will. And what will be happening in those 7th-9th innings? You will be scoring more runs and allowing fewer.
That is one specific outcome out of many. How often does a tie game in the 6th turn into a 1-run win? How often does a 3-run game in the 6th turn into a 1-run win? How often does the team with the worse bullpen hand more of those key innings to starters who might actually excel in them?
Forget a tie game in the 6th. Basically, you’re saying that teams play 1-run games, and teams with excellent ‘pens win more of those than their overall run diff. would predict. As far as I know:
1) that doesn’t consistently happen
2) the calculations suggest that the number of added wins by a ‘pen with an ERA a run lower than avg, with avg usage, will tip the 1-run games by maybe 3 all season.
There’s no reason we can’t. We just shouldn’t expect to, because, historically speaking, that’s very rare (to any degree as significant as +6).
No! Of course not! That’s my whole point! Last year was last year! If we’re gonna project 2007, I’d prefer to look at all of baseball history as opposed to just the 2006 Mets!
I don’t think that the matter of which innings you outscore your opponents in will dsitort expected wins at all.
For that mental model you’re using (which strikes me as neither “correct” nor “incorrect”, just a matter of perspective and abstraction), you can just as easily play a game backward. “Ooh, the bullpen has given us the lead, can the starting rotation hold it?” In the end, you allow how many runs you allow.
I worry that I’ve missed your point entirely… if so, sorry…
But doesn’t this say the Mets bullpen was singlehandedly responsible for almost all of the Pythag difference then?
Mets bullpen ERA: 3.25
Average NL bullpen ERA: 4.17
Mets bullpen IP: 542.2 IP
Average NL bullpen IP: 504.1 IP
So they were almost a run better than the league and had a larger workload of 40 innings. So 3 of the 6 additional Pythag wins were purely a function of the bullpen then, no?
check this out, I had never seen bullpen #’s before in “close and late” situations:
Actually, by randomly sampling a running score 162 times, you’re basically abstracting out the rules of baseball. There is no forward or backwards.
Have seen this? I’m staring at way too many papers as it is, but a quick skim suggests that, if the derivation of PythWin is at all usable, the simplifying and problematic assumption is that the same shape and scale parameters are used for modeling the distribution of both runs allowed and runs scored. I’m arguing that the Mets pitching staff invalidates this assumption.
Danny, yes, that’s certainly possible. I mean, to know for sure, we’d have to do some ridiculously detailed analysis of every relief inning of every 1-run game the Mets won last year. Without doing that, I think the statmasters would say, “The +6 was probably +5 luck and +1 bullpen, but in a given year, who knows? Coulda been -1 luck, +5 bullpen and +2 brilliant managing for all we know. Still, don’t bank on things turning out similarly in 2007.”
If anyone wants to “project” that we’ll win 1 to 3 more games in 2007 than our Pythag says, I’m not gonna argue too hard with that.
Historically speaking, though, I think the safer bet is that we hit our expected win total. (But, to be fair, “historically speaking” is not Wagner/Heilman/Sanchez/Mota/Feliciano/Show/Burgos.)
Either way, a 2nd straight +6 would be a definite outlier. (Although, that could be because so few relievers maintain performance from year to year? Hmm…)
Hmm. I still don’t see why that assumption is required for Pythag wins to be usable. I’m coming at this from an experimental angle, i.e. “We have a few thousand team-seasons in the books to look at, let’s see what’s happened.” Well, Pythag wins does a good job of explaining what’s happened.
I’m certainly open to the idea that a more sophisticated method of projecting a particular team-season is possible, but I haven’t bothered because 162 games seems a small enough sample size for the influence of luck to be significant. I mean, what if one of the guys who “was supposed to make us win more than our Pythag” gets injured? I’m just not motivated to work too hard on it. If anyone else is, though, please lemme know what you discover!
Dep, you article is alternately interesting and terrifying. It still doesn’t mean the Phillies catch us, but probably means they were slightly unlucky late in games and Manuel sucks.
argon, I wouldn’t be surprised if the volatility of bullpens was the real equalizer when it comes to the absence of occurrences where teams consistently outperform Pythag, and the bullpen was in fact a very key compenent in a team’s actual W-L relative to its Pythag W-L. Interesting.
Dep, that’s amazing. Where did you find that?
Here I thought the Phils’ pen just sucked. In reality, they were utterly fantastic when it didn’t count.
Any Phil-fan lurkers wanna argue that trading Abreu turned around the team vibe and made Geoff Geary more clutch? :)
We should totally try to convince someone to research that. :)
Okay, lemme back off a bit and say that a performance that good probably will effect your record in 1-run games. I mean, that 2.20 has got to include some stellar 1-run-game performances. Wow.
If we can do that again, sure, screw Pythagoras. I think we may miss the ChadBrad…
Am I reading that right in that someone tied us for that 2.20 mark??? Who??? The other top team in wins over expected is the Brewers at +5, followed by the Reds at +4… two teams that had major trouble finding closers.
it was from a phillies blog. sorry for not linking before, i always forget.
http://www.philliesflow.com/120106.html
I think that Tie relates to the 68 runs allowed not the ERA, i cant imagine anyone tied that!
In 1999 BP ran a study that looked at how bullpens affected a team’s likelihood of outperforming their Pythag record. The study ran on ESPN but seems to no longer be up. The follow-up can be found here.
The study showed that teams with great pens outperformed their Pythag record by 1.3 games, and teams with awful pens underperformed by 1.6.
Personally, I don’t like adjusting a team’s Pythagorean record for it’s bullpen. Pythag record is purely a predictive tool, created for one purpose: to do a better job of predicting a team’s record the next season than actual record, which it does well. The problem with making a second adjustment for your bullpen is due to the nature of bullpens–very few of them are great in successive years. They’re just too chaotic to predict beyond “looks like it should be good.” I can’t, for instance, look at a bullpen and say, “oh that one’s definitely gonna mean at least two games over Pythag.” Especially when a lack of timely hitting could swing things three games in the other direction.
I’m not arguing that PythWin isn’t usable; empirically, it fits well for much of baseball history. However, by referring to that attempt at a derivation I’m trying to provide a more rigorous theoretical justification for why PythWin doesn’t/won’t work for the Mets.
That’s cool, and I’d be happy to discuss said justification if I was clearer on the derivation. I’m just not sure whether that’s a useful/accurate way to look at the Pythag calculations or not.
I haven’t read your link yet, maybe that’s the problem. :(
Thank you Alex!
I think the actual number (1.3 wins) makes my point a lot better than my guesstimate of 3 wins.
At the end of the article:
So, we come full circle. But in fact, we had both a great bullpen and a high percentage of QS from 4 of our starters.
I new there was a positive correlation, but I am suprised it is as small as it is. I would have predicted more 2-3 games a year (and 3 games is a significant amount, as a look at the 1985 NL East standings will tell you!)
The fact is if “good” bullpens win 1.3 more, than “great” bullpens probably move to 2+ wins a year. Which further reinforces argon’s statement that the 2006 Mets pen probably accounted for +3 of the wins!
I also wonder how much the expected record vs. actual was effected by the team shutting it down at the end of the season. I mean how many less runs did the Mets score in the last three weeks (when Beltran sat a lot) then they would have otherwise.
There is so much “noise” in just one season (harken again back to the 20+ run differential in a 4 game series in 2005) that can effect “expected wins” and losses.
I know this system looks at historical average, but controlling for all of these other factors is impossible when you look at a team’s season.
So, between clinching so early and the ultra-dominance of the Mets bullpen, I am not sure that there 6 extra wins can be attributed to “luck.”
all we need is for Maine to pitch past the 6th inning at least .500 of the time early in the season, Perez to throw more strikes then balls, and for Humber to take that 5 slot with Vargas and Pelfrey ready to pitch at triple A. Humber is the man for me he can flat out pitch. Mets future Humber, Perez, Pelfrey, Niese, Vargas, Mulvey, Deolis, good lord are pitching after this year will become a strength instead of a ? You got some guys who can throw gas by people, and with Niese and Deolis still being teens im quite sure they can add 3-4 mph’s on their fastballs before they turn 22. :)