I think you may have simply presumed I thought the idea was far-fetched; in fact, I only qualified the idea as being (a) currently ambiguous in the case of Tyler Yates, and (b) difficult to describe in terms of acute areas of improvement ("somehow").  My exact words were:

"I am not comfortable completely discounting a very large sample of data just because a player may have somehow benefited from a surgery."

"His FIP last year was 3.84 to give you an idea of what he could do with league-average luck."

I guess I question whether to believe Yates is the 3.84 FIP pitcher with a little bad-luck with BAbip (as in 2007), or to believe he is the 4.70 FIP pitcher with good luck on balls-in-play (.273 BAbip in 2006).  My fear, with respect to his future, is that he is likely somewhere in between - especially at age-30 - and hence will oscillate within the 4.20-4.80 ERA range.

"I don't know what you mean by "material success", but if you mean having a low ERA, then there's little any pitcher can do about that.  ERA involves a helluva lot of luck, and sometimes the batted balls are just gonna fall in."

I understand this point perfectly well; however, ERA does measure something, especially for starting pitchers, and especially in the aggregate and over a large sample-size.  For instance, if two pitchers throw 3000 innings, and one carries a 4.50 ERA over that span, while another carries a 3.50 ERA, earned-run-average is quantifiably-sound enough to justify a qualitative description: the pitcher with the 3.50 ERA was clearly the superior pitcher over the course of their respective careers.

(Of course, in less discrete cases, issues like park-effects, league-averages, and overall run-environments much be more prominently accounted -- such as in the qualitative distinction between a 3.50 ERA pitcher from the 1960s and a 3.60 ERA pitcher from the 1990s - but in those cases, ERA+, dips/FIP, et cetera once again become necessary and highly advantageous.)

In the specific case of Tyler Yates, however, I simply meant run-prevention in high-leverage games situations, which are frequent in the middle innings of games.  With relievers, ERA is further problematized by factors such as runs ascribed to the previous/relieved pitcher (mid-inning uses of relievers) and vastly smaller sample-sizes; hence, I will use OBP/SLG-against because it measures more directly the reliever's interactions with the batters faced (positive and negative), because it accounts for walks and extra-base hits, and because it is easily broken down into game-specific situations through interweb technology.

Employing such a heuristic, we find that Yates was worse in most higher-leverage situations in 2007:

.250/.353/.409 in "Late & Close" situations*
(106 PA)

.253/.354/.398 in +- 1 R situations
(100 PA)

.394.487/.667 in tie-game situations
(41 PA)

In 2007, Yates pitched through 294 total PA and produced a line of .251/.338/.400 for the opposition; his best game-specific situational category (besides >4 R) was +- 2 R, and his .241.340/.358 in those 163 PA certainly helped his overall numbers.

Perhaps, either way, the issue will be Cox's usage pattern - whether it be specific game-situations or frequency/days-of-rest.  And, while I very much appreciate your arguing against the premise of the decision, and not the reality of the decision, I think many are simply glad to see Yates depart because - by removing him from Cox's purview - we will no longer have to witness Yates struggle with run-prevention, even if he isn't the predominant source of culpability.

(The usage patterns of major-league managers are certainly highly impugnable: for instance, I'd prefer Cox use Soriano in instances of the highest potential leverage - regardless of inning or save-situation.  However, most managers are psychological uncomfortable with this conception of bullpen-usage; moreover, most elite-quality (or even above-average) relief pitchers would object to this usage model, because the "Save" statistic has such cachet and the accompanying fiscal benefit in contract negotiations: Cf. Francisco Cordero.)

*Baseball-Reference defines "Late & Close" as "PA in the 7th or later with the batting team tied, ahead by one, or the tying run at least on deck."

"LAKE BUENA VISTA, Fla. -- When Chris Resop arrived at Spring Training, he still felt a little tentative about the fact that he was coming back from right elbow surgery. But six weeks later, the right-handed reliever finds himself with a high-velocity fastball and reason to feel confident that he'll be on the Braves' Opening Day roster.

'I definitely didn't expect my velocity to be where it's been,' said Resop, who was claimed off waivers from the Angels in October. 'It's higher now than it's ever been. I've always been a 93-95 [mph] guy, and for some reason this spring, I've been higher, a lot higher. I'll take it gladly.'"

Thus, even if your injury theory is correct, it may apply to both Yates and Resop; considering, Resop's superior track-record and the five-year age disparity benefiting Resop, it may not have been a mistake to stick with Resop if one of the two had to go.

---

An Open Letter