The initial paragraph got clobbered in my cut and paste, (I'm just gonna give up that method). It should've read:
What do you get when you start with assumptions?
sum into pass
mus piss on a ...
u spit on mass
sit maps on us
spam is not us
spasm on suit
ASSUMPTIONS
What do you get when you begin with assumptions?
sum into pass
must piss on a ...
u spit on mass
sit maps on us
spam is not us
spasm on suit
What the above demonstrates is that order is important. On the field, this is easy to see. A homer, triple, double, single, walk (in that order), scores two or three runs. In the opposite order, it scores five. The reality of roster management and team building is not so different.
After the 2007 season, there was a choice about "going for it" in 2008. The club had just won 88 games, and finished a paltry 6 games behind the Angels for the West division title, and the same 6 games out of the wild card spot. The choice at that point was to make a major trade of prospects for Erik Bedard. Bedard was a dominant pitcher ... but one with a short resume, and a history of health problems. But, the club opted to ditch its top near-ready CF spec, a solid veteran bullpen arm, and an additional pack of longer term prospect projects. The result was a complete disaster in 2008.
There were many who railed against the deal, (high on AJ or low on Bedard). There were some who argued, (and in hindsight rightly), that the club was not NEARLY close enough to being a legitimate contender to take that kind of gamble. The 88-74 club had been outscored 794-813. Of course, there were those who argued that the pythag projection for that team didn't tell the whole story, (79-83). So, BEFORE examining the choices the 2009 club has heading into 2010, let us try and identify what errors were made in the 2007 to 2008 assessment, hopefully to avoid making the same ones.
1) Failure to heed pythag -- When wins and run differential don't match, the won-loss record almost always wins the hearts and minds of even the most analytical minds. The danger is in believing that the win/loss record is catching some magical component of the team WHICH IT CAN MAINTAIN. This has been disproven so many times that it is hard to believe that people still fall into this trap. For the 2007 club, the argument was that the horrid back end of the rotation was so horrid that they bloated the runs allowed, so the club was REALLY an 88 win team, to only a few extra runs prevented could push the team to 90 or 92 or 95 wins.
It is sadly amusing to note the complete illogic of the position. The argument is that because the team was losing lots of 4-14 games, that getting slightly better pitchers would result in wins. It won't. It CAN'T. If the assessment is correct, then addressing the very position that caused the abberation will have *NO POSITIVE IMPACT AT ALL*. What will happen is that you will those same games 4-6 or 3-5. But, you will STILL LOSE. The unrelenting reliability of pythag is this --- that while teams can (and do), sometimes have run distribution that doesn't match up with normal run distribution, the ONLY way to maintain a pythag skew is to maintain whatever it is that is causing that abberation. In the case of the foul BOR pitchers, the only way maintain the mirage was to continue to have miserable BOR starters. Does ANYONE believe that is a recipe for success? Anyone?
So, word of warning #1 for the 2010 projections. The 2007 club was outscored by only 19 runs, while winning 88 games. The 2009 club was outscored by 52 runs, while winning 85. Even if one believes they can identify why the 2009 club beat pythag ... before one begins projecting 2010, one would need to find a BUNCH of similar teams that managed to maintain said skew the next season before buying into the notion that they have SUSTAINABLE magic that can help the next season. I'm skeptical that such a list exists.
Comments
You have implicitly assumed that the Bedard trade was directly responsible for the follow-on collapse...or that's the way it comes across. It didn't. In fact the Bedard trade played almost no role at all in determining the fate of the '08 club or nay future club unless you think with all our pitching depth that we're going to depserately miss Tillman and Butler.
Adam Jones didn't have a veyr good 2008 year and has been easily replaced in 2009.
George Sherrill is now and has always been spectacularly overrated by most Mariner fans. He's a nice fungible lefty specialist with stuff good enough to close out games shakily for a last place team. Woo.
The rest of the prospects were replaceable too.
The 2008 team imploded due to many factors...NONE of which were Erik Bedard - aside from Bedard's injury trouble that is.
I don't see JackZ addressing pitching, like the M's did with the Bedard pickup. If one looks at pythag, we won more than we should have, yes, but our weak offense did nothing to help. Our pitching and defense are not the problem, and I fully believe that Jack knows this.
The M's are in a much better position this year, as all they must do is find more offense. The answer to maintaining a winning team and reversing our pythag record is much clearer this time around.
Most AL teams are giving up ~4+ runs per game, leaders and also-rans, and may be expect to do approximately the same in 2010. Ten teams gave up between 732-771 runs, three teams were flat-out shelled, and M's led with 692 runs allowed (fewest). The big difference between the M's and the league leading teams is that we scored ~200 runs less. That's my recipe. Plan the roster to score ~200 more runs, while continue to pitch competitively.
Where do we do that? Ichiro should have scored ~120 runs, but the #2 and #4 hitters couldn't hit themselves out of a bag. Those averages need to hit in the #8-9 spots, or maybe in Tacoma. By position, we suffered sucking chest wounds at C, SS, LF, and 3B. I listed the latter only because we had no acceptable fill-in at 3B with Beltre out with shoulder surgery, and the shot to the boys. Otherwise, he likely would have hit for approximately his career averages. Of the other positions, one might only expect a small up-tick from Guti, and maybe a new DH would improve somewhat over the Griffey/Sweeney DH composite. Otherwise, the 200 runs need to come from the sucking chest wounds. For C and SS, I don't expect much more RBI, though runs scored could improve if they could improve by ~36 hits total (~12-14 more runs). The biggest additions needs to come in LF and 3B, with Ibanez-like numbers or better in left, and Beltre career average numbers at 3B.
Oh goodie, this is gonna be fun this year. I'll keep score. Dr. D 1 - Dave 1. Top of the second and Dave's at bat.
In re-reading my post, I can understand the confusion. But, the post was not intended to even discuss Bedard in any real way. The point was to discuss the ASSESSMENT process before you start making moves. In trying to keep my post-size under control, I only covered the single point -- ignoring pythag is a bad, bad idea.
It was my intent to follow up with some of the other mistakes made in assessing exactly where the end of 2007 club stood, and why a superior job of assessing what existed at the time would have painted a more realistic picture of what was to come, (Bedard trade or not).
What cannot be disputed is that the Bedard trade was made and the 2008 season was a disaster. Doesn't mean that Bedard was to blame, (or the trade), were to blame. But, I think it is very important to attempt to uncover what was known PRIOR to the 2008 season that might have reduced the depth of where the club went.
In re-reading my post, I can understand the confusion. But, the post was not intended to even discuss Bedard in any real way. The point was to discuss the ASSESSMENT process before you start making moves. In trying to keep my post-size under control, I only covered the single point -- ignoring pythag is a bad, bad idea.
It was my intent to follow up with some of the other mistakes made in assessing exactly where the end of 2007 club stood, and why a superior job of assessing what existed at the time would have painted a more realistic picture of what was to come, (Bedard trade or not).
What cannot be disputed is that the Bedard trade was made and the 2008 season was a disaster. Doesn't mean that Bedard was to blame, (or the trade), were to blame. But, I think it is very important to attempt to uncover what was known PRIOR to the 2008 season that might have reduced the depth of where the club went.
Matt: Don't worry about Pythag when discussing the 2009 Mariners. Their bullpen will allow them to repeat and build on the W-L success of 2009.
Sandy: Don't ignore the Pythag when discussing the Mariners' prospects for 2010. Any attempt to downplay its inexhorable force must include a good number of example teams that share the same characteristics offered as reason for doing so.
Have AT IT, boys! I'm with Sandy on this issue, but I sure hope Matt can be convince me otherwise. If Sandy's right, the M's likely still have a multi-year project on their hands to become true contenders in a division that looks very competitive going forward. If Matt is right, we could contend in 2010 or 2011.
After 2007, I was saying the bullpen and the lopsided rotation caused our success relative to pythag and likely meant continued pythag-beating. I was wrong. As I am capable of noticing when I make an analysis mistake, I went into 2009 skeptical of my original point re: the bullpen. I think Sandy is right that bullpen "success" by WPA is mostly situational luck and not something that should be counted on to sustain into future seasons.
No...the reason I'm not worried about the pythag split in 2010 is because it was *NOT* caused by the bullpen...it was caused by a wildly inconsistent pitching staff
And the character of our pitching staff is unlikely to change going forward.
The ace at the top providing consistent success
The fungible core of MOR/BOR fodder constantly being swapped in and out by Z and W
The hard-throwing bullpen with a lot of inconsistent pieces
This type of roster construction (when paired with good tema defense) produces RA results like:
2
4
2
13
5
0
4
0
6
2
3
9
And that has the property of creating pythag ratios that mislead. And no, I don't think that's just luck that will even out. My point is...our pythag ratio this year was misleading for a reason that is persistent.
On top of that, I believe our offense isn't as bad as it currently appears
As i pointed out a couple of weeks back, we're now very YOUNG with the bats...not very old like we were in 2007
A lot more room for upside.
The Ms were unlucky in raw runs scored this year. They were poor in RISP.
The Ms had a 716 OPS vs an opponents OPS of 710.
Ms: .258/.314/.402
Opponents: .247/.316/.394
The Ms' raw WAR adds up to around 83 and considering strength of schedule they pretty much ended up where they should be (a team slightly over .500).
In 2007? The Ms had a 762 team OPS vs opponnents OPS of 780. They were LEGITIMATELY lucky that year. An 88 win team that deserved to have a losing record. Thats something we should have seen coming.
The encouraging thing about the Ms this year IMO is that they are built on the strength of their stars - Felix, Ichiro, and Gutierrez. Lopez, Beltre, and Washburn were the only other players over 2 WAR and the team is full of >1 WAR position players. RRS and Bedard were the only other SPs over 1 WAR. The 6th best position player was Johjima this year. The Ms are 'easier' to upgrade since they are so extremely strong at the top and weak at the bottom.
The Angels aren't going to run a .326 BABIP next year OR run an 833 OPS in high leverage situations. Maybe they'll go +5 wins over their expected total again too, but I wouldn't bet on it. LAA is going to get worse, Texas and Oakland will be even better (Oakland potentially a lot better). The Ms have their work cut out for them, but '10 isn't a long-shot IMO.
The...the Mariners were bad in high leverage plate appearances this year at least in part because they had a few guys who could hit (who would get on base) and then the rest couldn't hit at all (thus resulting in poor performance after there were guys on base).
The reducto ad absurdem on that would be to imagine team with 9 players:
.000
.000
1.000
.000
.000
.000
.000
.000
.000
That team is going to hit ~0.111 overall and ZERO with men on base. Just FWIW.
Maybe thats part of it, but its more likely just one of those things. 2 outs, RISP the Ms ran a .228 BABIP (615 OPS).
Either way its very unlikely to continue in '10 and beyond.
Sorry Matt, but I think you're wrong about *any* trait being able to beat pythag on a consistent basis. I've seen zero historical evidence to support the notion in general - much less one that would support the Ms 2009 team as being a model that could sustain its pythag-beating ways.
That said - I also think your assessment of why the Ms beat pythag is wildly inaccurate, and not supported by the facts.
I charted the Ms run scoring and runs allowed charts against those of the other AL West clubs. On the run scoring side, the Ms have one, (and only one), data point that is significantly skewed from the standard trend -- they scored EXACTLY 4 runs 28 times, compared to the division average 20. When charted, it's easy to spot visually.
Other than that, they had zero games scoring more than 11. The other 3 teams combined for 17 games scoring more than 11, (basically average of 6 each).
On the runs allowed side, they scored exactly 3 runs 31 times, beating the average of 24.25 by a significant amount. (Average of 22 if you remove Seattle from the 3-run mix).
On the upside, Seattle never allowed more than 12. The other 3 teams combined to allow more than 12 in 12 games.
The instinct is to blame the blowout games for the skew. But Seattle actually didn't create OR allow blowout scoring, which removes them from play in regard to pythag.
=========
So, first off, the notion that Seattle's pitching was wildly inconsistent (compared to the competition) is false. The primary cause for the pythag skew was that the club had an inordinate number of games scoring PRECISELY 4 and allowing PRECISELY 3. Other than that, the Seattle run scoring and allowed followed the same basic pattern for other teams. While run scoring was suppressed (on both sides), but the DISTRIBUTION was otherwise normal.
From a mathematical view of the results, AT BEST one might conclude that the pythag skew was effectively evenly split in "credit" (blame?) between the offense and pitching. Moreover, the knowledge that the only significant data points well away from norm lie at the 3 and 4 run marks is FURTHER supported by the fact that the club went 35-20 in 1-run games.
The fact the club had 1/3 of its games be 1-run games is remarkable in itself. Detroit was second in the AL in 1-run wins with 28, (second in 1-run games also, going 28-22). The 2nd best 1-run time by net wins was the Angels, going 27-18, (+9 games). Think about that, Seattle was +15 in 1-run games. That means, not only did they absolutely destroy the rest of the AL in 1-run wins -- they destroyed the #2 team by +6 wins.
So, the evidence doesn't support the wildly inconsistent pitching theory. It supports the notion that the credit for the pythag skew was almost evenly split between offense and defense. And it pegs the lions share of the pythag damage due to the 1-run peaks at 3 and 4 runs.
So ... if you want to convince me that the club has a sustainable pythag beating makeup, then your model must include an explanation for exactly how the team is going to consistently allow EXACTLY 3 runs and score EXACTLY 4 runs. Me? I don't think there's any team design that creates such an expectation.
Additionally, there is strong reason to believe at least 40% of the innings in 2010 will not be devliered by pitchers from 2009, so the notion that they will sustain this alleged (phantom) inconsistency, would require the new arms to repeat PRECISELY this random fluctation. Honestly, what really amazes me is that you're arguing that RANDOM FLUCTUATION is what will make the pythag beating persistent.
===========
...it's not random to be inconsistent. That's a spectacularly inaccurate use of statistical terminology and you shuold know better. Random variation has a consistent pattern of spread when enough samples come in. The argument was that the specific model the Mariners use to build a pitching staff (stars and scrubs) is more likely to produce a series of run allowed figures that are more skewed-right than the league average bias would be given the Mariners' average RA/G. No other AL team allowed as few as the 4-ish runs per game that the Mariners allowed. The variance of RA is directly propotional the mean of runs allowed. So you need to adjust the data to account for the change in variance expected when you decrease RA from the league average to the Mariners' average. Of course the Mariners' pitchers didn't allow more than 12 runs while other teams did...other teams had worse team defenses than did the Mariners. Frickin' DUH, Sandy. But if you check the Mariners' standard deviation of run allowing vs. the all time major league average standard deviation for run allowing for all teams who allowed a similar number of total runs per game, you're GOING to find that Seattle was less consistent than the average 4 RA/G team.
inserted to fix formatting...
This is what happens when people start looking at the surface data willy-nilly with zero real understand of the statstical context. Your 30,000 foot view approach is a good starting point and you are eminently logical in your thinking, but this type of tactic will fail EVERY. SINGLE. TIME. if you don't understand the model. Which you apparently don't.
format fix again
To your larger points...the Mariner team offense was unusually consistent...that is a major component for beating pythag. Studies HAVE INDEED SHOWN that consistent offenses tend to stay consistent when the personnel don't change a lot. Do some background research before you go mouthing off about there being zero evidence. Thank you.
format
On the defensive side, the assumption implicit in Pythag is that the standard deviation of performance will be a constant that is proportional to the mean of performance. NOT that the standard deviation will be constant for all run scoring contexts. Your attepmt at analysis would be right if we expected the Mariners to have the same variance in RA as the Rangers...but we DON'T...and neither does Pythag.
format
I'll post more on this at a later time. You have a point about one-run games...and that compoent isn't likely to persist (but if you correct our 1-run game performance we drop from 35-20 to 29-26 (Baseball Prospectus had a study showing that efficient defensive clubs beat the expected 1-run performance consistently)...that's 6 games...we'd still be near .500...and we'd still be beating pythag).