To me the current construction of the bullpen lends itself perfectly to implement a tandem system. Devenski and Peacock throw multiple innings almost every appearance, and Musgrove, Morton, McCullers, and Fiers have histories of not being able to pitch deep into games. Not to mention there are multiple guys who appear capable of handling high leverage late innings. This is how I'd set it up: Keuchel 1-6, Hoyt/Harris/Giles 7/8/9 McCullers 1-5, Devenski 6-8, Gregerson 9 Morton 1-5, Peacock 6-9 Musgrove 1-5, Feliz/Hoyt/Harris/Giles 6/7/8/9 Fiers 1-5, Devenski 6-8, Gregerson 9 With Sipp throwing as a loogy or low leverage arm. If you run into extra innings or have a starter get knocked out before the 5th, you have Feliz and Hoyt with options who you can send down and call up Tolliver, Jankowski, or Diaz for fresh arms.
Since you're basically setting a rotation without regard to score or how players are performing, this just seems like a great way to take your worse pitchers and end up putting them into higher leverage situations while putting your best pitchers in lower leverage situations. It also ends up with Devenski pitching the most innings on the team (180 over a season), and out of the bullpen, which likely means his arm falls off.
Wrong. When you say "worse" (I assume you mean "worst"), I'm not sure who you're referring to. Which pitcher outside of Gregerson and Sipp (both of whom appear to have gotten back on track) haven't been effective this season? In my hypothetical schedule, Devenski would be pitching 6 innings every 5 game cycle, which incidentally is the EXACT amount you'd expect in a typical starter's workload. Not sure how that equates to "his arm falling off". Of course leverage, inning/pitch count, opponent, and recent performance would keep things in some measure of flux, but for the time being this could prove very effective for Houston as it would make pitcher usage more predictable.
I don't know, but there's going to a pecking order as the season progresses. I don't want our great pitchers pitching in a 9-0 game because of some arbitrary pre-made schedule, nor do I want Gregerson pitching the 9th inning of a 1-0 game just because he was scheduled that way. When you are putting pitchers into games based on a pre-made schedule without regard to score, you're by nature using them in less than ideal situations. The whole wow-factor of Devenski, for example, is you can put him in exactly when you need the highest leverage. Instead, you're relegating him to 6 innings in games where he may or may be needed, thus taking him out of games where he may actually be useful. Yes, and there's a reason relievers pitch 60-80 innings a season instead of 180. Predictable and effective are not remotely the same thing.