Thursday, May 26, 2005

Frobby v. Pythag

This time, it's personal!

Nats Blogster Dexys_Midnight posted a rumination today on what it means to root for a team managed by Frank Robinson. (It's turning into a rather memorable experience, isn't it?) Allow me to excerpt an intriguing passage:

My main problem with Frank Robinson is that he far too frequently chooses wrongly on those things upon which reasonable minds should not differ. And that is going to cost a team a few games a season (maybe 5 give or take--I am not citing a stat here, just guessing. Add up 5 games a season and you get pretty close to Frank's woeful career managerial record).

This is an interesting thought. I suppose there is a way, historically at least, to frame Robinson's managerial record beyond "mere" wins and losses---or at least reframe it in a context which might serve to isolate at least some of his navigational skills.

Well, I'll hesitate to go that far, actually; neverthless, a look at the Pythagorean records of the teams that he's managed might prove somewhat illuminating.

Here's a run-down of his full seasons as a manager. I could use Retrosheet for the partial seasons, but you guessed it: that would take way too long. Well, here we go:

Year . . . . . . . W/L . . . . . . . +/- PYTHAG
1975 . . . . . . . 79-80 . . . . . . . (-1)
1976 . . . . . . . 81-78 . . . . . . . (+1.5)
1981 . . . . . . . 56-55 . . . . . . . (-1)*
1982 . . . . . . . 87-75 . . . . . . . (+8)
1983 . . . . . . . 79-83 . . . . . . . (-1)
1988 . . . . . . . 54-101 . . . . . . (~+1)**
1989 . . . . . . . 87-75 . . . . . . . (+4)
1990 . . . . . . . 76-85 . . . . . . . (-1)
2002 . . . . . . . 83-79 . . . . . . . (0)
2003 . . . . . . . 83-79 . . . . . . . (+3)
2004 . . . . . . . 67-95 . . . . . . . (0)

* Strike season
** Okay, I fudged a bit here. Cal Ripken, Sr. was fired six games in (way to give it the old college try, guys!), and Robinson managed the rest of the season. The final record was 54-107, and the Pythagorean record was 55-106. Since we're only talking six games here, I merely assigned a 2-4 record, almost precisely the O's winning percentage for the season, to Robinson's season, giving him a plus-one. Yeah, that's sloppy; so sue me. ;-)

That comes out to + 14.5 wins above the expected record. (This season, Robinson's crew is three games above the expected pace, though I'd say a 47-game sample is a pretty dicey sample on which to make any conclusions. See yesterday's 12-3 loss for why.)

Sure, 14.5 games looks pretty nice, but let's add three rather significant caveats:

1) Pythagorean record is not a measure of managerial effectiveness; it is merely a measure, and discrepancies from actual records on the field can (and have been) chalked up to other factors, most often sheer, dumb luck.

2) To place those 14.5 games in context, we're talking about 11 seasons in the "study." That figures to just over a win per season gained.

3) Over half of Robinson's gains were made in one season, 1982, and + 8 is a pretty significant variation. If that's reflective of luck, then evaluations based on Robinson's Pythagorean record are even more tenuous.

There's also the consideration, I should add, that managers are not static beings. I don't have his play-by-play data handy, so I can't really tell you whether he was a bunting fool in, say, 1977. Leagues are also not static entities; certainly, it makes more sense (at least in theory) to bunt in a low-offense era than it does now.

Interesting look at the team.

Something I thought about doing was ala the model Bill James used in one of his books. (The hall of fame one?)

He worked under the assumption that teams generally revert back to .500, so he used a team's W/L over the previous three seasons to derive a baseline figure of wins -- creating a 'par' figure, if you will.

Obviously that's not a perfect method, but it'd be interesting to see how it turned out with Frank.

Maybe this weekend...
I think it was in the Guide to Managers book. (That's also where he devised a "manager's baseball card" idea, an attempt to put a manager's strategic tendencies in an easy-to-refer form. I guess STATS Inc. does that now.)

Anyway, re: the three year thing: That's basically "the Plexiglass Principle" or the "Whirlpool Effect" from his Abstracts. He used them as different concepts, I guess, though in conjunction they say the same thing: teams are forcefully drawn to the "middle" (whirlpool), so they "spring back" (plexiglass) that way over a period of time.

I'm not sure if the perspective holds much anymore, balance it's tied to what he called "The Law of Competitive Balance": teams that are bad will adjust accordingly because they have motivation to do so, while teams that are good currently won't tend to adjust because they'll have little motivation to do so.

A couple years ago, Prospectus (Jonah Keri, maybe?) questioned the law of competitive balance, arguing that factors inapplicable two decades ago muddy the simple proposition. It's Prospectus, grain of salt, jaded Nats' fan proviso . . . but there's probably some truth there.
"because" somehow became "balance."

I have no idea . . .
Anyone else notice he's totally on pace to be fired mid season? You can't break a nice pattern like that.
Good point!
Generic Zocor Simvastatin drug is an hmg-coa reductase inhibitor or "statin" used to lower bad cholesterol (LDL) and triglyceride levels in your blood or to raise good cholesterol (HDL) levels in adults.

Generic Lipitor ATORVASTATIN drug belongs to the group of medicines called 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors.It works by blocking an enzyme that is needed by the body to make cholesterol, thereby reducing the amount of cholesterol in the blood.

Generic Crestor Rosuvastatin drug is used to reduce the amounts of LDL (bad) cholesterol, total cholesterol, triglycerides (another type of fat), and apolipoprotein B (a protein needed to make cholesterol) in the blood. Rosuvastatin also increases the level of HDL (good) cholesterol in the blood. These actions are important in reducing the risk of atherosclerosis (fatty deposits in the arteries), which can lead to heart attack, stroke, and peripheral vascular disease.

Generic Norvasc Amlodipine drug is used to treat hypertension (high blood pressure) and to treat angina (chest pain).

Generic Plavix Clopidogrel drug is used in the prevention and treatment of heart attack, stroke, blood clots, and acute coronary syndrome.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?