Conversations about Ryan Franklin are kind of like the anti-snowflake - they're always the same, and there's never any chance that some new variation will come up. What I mean is that every discussion manages to cover each of these points:
-Franklin has a bad strikeout rate
-Franklin was good in 2003
-Franklin throws a bunch of innings
-Franklin would be better if Price let him throw the knuckleball!
-Franklin is helped by his environment (Safeco and team defense)
-Franklin could be adequately replaced by an AAA'er for the league minimum
-Franklin is hurt by his low run support
Now, I feel like most of these points have been sufficiently analyzed, so I won't touch on them. Rather, I want to talk about the one that nobody's ever cared to investigate (at least, as far as I know) - that last one, talking about run support. Among pitchers who qualified for the ERA title last year, Franklins average run support of 3.14 was the lowest in baseball. A year earlier, his 4.84 figure just missed being in the lower third. Some people claim that low support forces Franklin to try and be perfect, which would inevitably result in a decrease in effectiveness as he presses too hard.
Grant Fuhr, Old Time Hockey Goalie Extraordinaire, used to "let up" when his team was leading by a bunch of goals, but he'd buckle down in close games. It's not really the same thing, of course; letting up is a heck of a lot easier to do than buckling down, and besides that, it's a completely different sport. The reason I mentioned it is that it provides at least one anecdotal example of a team's offense and defense (of sorts) being intertwined.
With that in mind, I wanted to investigate the numbers and see if there is a distinct performance difference between guys with low run support and guys with high run support versus the league average. To do so, I took the cumulative ERA's of the top 20 pitchers in each category for both 2003 and 2004 - that is, in the end I had 40 pitchers representing each extreme. This is what I got:
League-average ERA for a qualified starter, 2003-2004: 4.12
Average ERA for 40 pitchers with lowest run support: 3.91
Average ERA for 40 pitchers with highest run support: 4.34
"But wait, that can't be!" you say. And you're sort of right - I've forgotten to eliminate a major variable that might have a considerable effect on the numbers. That is, the pitchers with low run support might play in pitcher-friendly environments, which will depress both run support and ERA. By the same token, the pitchers with high run support might play in hitter-friendly environments, which will inflate both run support and ERA. So I ran through the numbers again, this time weighing them appropriately for home park factor. This is what I wound up with:
League-average ERA for a qualified starter, 2003-2004: 4.13
Average ERA for 40 pitchers with lowest run support: 3.97
Average ERA for 40 pitchers with highest run support: 4.27
So the average essentially stayed the same (as you could guess; taking a sample representative of the entire league will yield a neutral park factor), while the pitchers with low run support saw their park-adjusted ERAs climb by six points and the pitchers with the high run support saw their park-adjusted ERAs fall by seven points.
So, what do we have? These numbers suggest that pitchers with low run support actually pitch better than average - that is, that the pressure of not being able to make a mistake actually helps their effectiveness.
And on the opposite end of the spectrum, the numbers indicate that perhaps pitchers with high run support actually do let up a little bit, with the knowledge that one mistake won't kill them.
"No way!" you say. And again, you're right. After seeing these numbers, you might expect me to make some sweeping conclusion that pressure actually helps pitchers perform a little bit better. I can't do that, though - 40-player samples aren't large enough to merit scientific proof, particularly when the same pitcher shows up two times (for example, Derek Lowe had very good run support in both 2003 and 2004, so his performance in both of those seasons was included in the data sample).
What you can take out of this little experiment, however, is that there's no indication that low run support forces pitchers to make mistakes more often, on account of pressing too hard or thinking too much about the consequences of one bad pitch. Whether Ryan Franklin is an anomaly, I couldn't tell you, but I'm inclined to say that the "low run support" argument is just a way to deflect blame for what was a miserable 2004 season. A lot of guys complain about their offense not giving them enough help, but when the data sample shows no significant decrease in performance for guys with low RS, it's hard to take them seriously.