Luck, or the residue of design?

Jonah Keri this week wrote one of those pseudo-sabertastic the Angels-aren't-all-that columns that you see from time to time, arguing that:

in the end, the Angels' success this season largely boils down to one explanation: They've been really, really lucky.

How could a team riddled with injuries to key performers, and suffering through Vladimir Guererro's worst-ever first half, be considered "lucky"? Because, Keri says, we've scored just a couple more runs than we've allowed, and thus, by the Pythagorean projection of winning percentage, we should be 34-34 right now, and not 41-27; while the hated A's should be at 39-28 instead of 36-31.

But is Pythagoras really the best way to judge an expected record, and does overshooting (or undershooting) its projection by a handful of games really amount to the most significant kind of "luck" out there? I'd answer "no" and "no."

For the problem with using square roots to determine records early in the season, let's go back to Keri:

So, how does a team win so many more games than you'd expect? Getting blown out a lot can skew the numbers. The Angels are 3-7 in games decided by five runs or more — that record includes a 14-2 lambasting at the hands of the A's on April 28. Such lopsided losses have an especially big effect on expected records early in a season, when we haven't processed enough game results to smooth out those kinds of blowouts. But over the long haul, good teams tend to blow out their opponents more than they get blown out
On the flip side, winning close games can help a team look like overachievers. Through Sunday, the Angels were 13-8 in one-run games, the best such mark in the American League.

Note the conflicting modes of record-interpretation here -- the bad record in blowouts is indicative, and the good record in one-run games is illusory. "[O]ver the long haul, good teams tend to blow out their opponents more than they get blown out," ergo, the Angels might not be a "good team." Yet a good record in close games "can help a team look like overachievers."

Regarding blowouts, two very important reasons the Angels haven't been doing any of that, especially since the first half of April, is that 1) their best offensive player has uncharacteristically struggled; 2) two of their other best offensive players missed a month-plus of games, and their respective backups (and backups of their backups) hit horrendously. Let's compare, for example, Howie Kendrick to his replacements ("SC" stands for "scrubs"):

HK: .326/.359/.453
SC: .212/.313/.269

Now the same with Chone Figgins:

CK: .321/.433/.368
SC: .217/.250/.283

Kinda hard to blow teams out when two lineup spots are filled with guys who make Ike Hampton look like Pudge Rodriguez. Also, when it comes to the receiving end of blowouts, the return of John Lackey has meant six starts of 1.83 ERA (in 44 IP) compared to the 8 starts of a 7.36 ERA (in 36) by Moseley/Adenart. When a good team loses its good hitters at the same time it regains a good starter, it's no surprise that those 6-5 games turn into 3-2 affairs. Nor will it be any surprise when, going forward, the Angels offense reverts to its 5-runs-plus-a-game form; which, when coupled with the Lackey/Arredondo upgrades, will start putting us on the right end of blowouts.

What about one-run games? Yes, the Angels have a .619 record in those, but they are also .596 in games decided by two or more, so whatever overachieving is happening there is literally a difference of one measley game. What's more -- and this is important, since one corollary definition of "luck" is "unrepeatable results" -- the Angels have exceeded their Pythogorean projections by an average of 4 games per year from 2005-2007. And not once did they have a better record in one-run games while doing so. It's enough of a trend to make you wonder whether there is something in Scioscia's strategy producing the repeatable "luck," and also whether there are better ways to measure run distribution than a mere average.

Put it another way: The A's are averaging 4.45 runs a game; the Angels just 4.21. That means the A's have had a better offense this year, right? Think again.

You know the concept of quality starts? Where, by measuring the number of games a pitcher has gone at least 6 innings while compiling a game ERA of 4.50 or less, we aim to basically answer the question of "did he keep us in the game?" Well, one could -- and should! -- look at team offenses in a similar way. If you're shut out in three straight games, then score 16 the next day, that's a whole helluva lot less useful than churning out 4 runs every damn day. And when you compare the Angels' offense to Oakland's, you see a team offense that's doing a better job of giving their pitchers a chance to win. Here's the number of times this year each team has scored zero runs, 1 run, 2 runs ... all the up to 15:

0 = 2 6
1 = 3 11
2 = 11 7
3 = 13 10
4 = 13 6
5 = 9 5
6 = 8 7
7 = 3 3
8 = 1 2
9 = 2 3
10 = 3 1
11 = 0 1
12 = 0 1
13 = 0 2
14 = 0 1
15 = 0 1

The main discrepancy just jumps out a you -- the A's have scored 1 or fewer runs 17 times (every fourth game), going 0-17 in the process; while the Angels have done it just five times (going 2-3, thanks to two 1-0 shutouts). On the blowout end of the scale, the A's have amassed 11 or more runs six times to the Angels' zero. Would you rather have an offense that scores three runs or less in more than half its games (going 8-26), or one that scores four or more in 39 out of 68 (going 29-10)?

Is that just a random distribution, or does offensive philosophy have something to do with it? I don't have enough time to crunch all the numbers for the last few years, but I wouldn't rule out the latter. You can neutralize Oakland and other extreme take-n-rake clubs by the mere act of throwing strikes (see Lackey, John). The Angels, with their infuriating-to-many emphasis on aggressive baserunning, the contact play, and periodic one-run strategies, might just be harder to shut out, though remind me of that statement again next time a pus-throwing lefty swings through town.

At any rate, though averages do have a tendency to even out over the long haul, it's noteworthy that Scioscia has beaten Pythagoras by 19 games now over the past 4.4 seasons, despite *never* -- until this year, anyway -- overachieving in one-run games. Seems to me that a more even distribution of runs on offense, and a perhaps larger-than-usual gap between our quality starters and the Dustin Moseley fill-ins of the world, might have something more to do with this than just blind "luck".

This FanPost is authored by an independent fan. Tell us what you think and how you feel.