OK, no intro, just right into what you're thinking based on the title of this post - what in the world is this guy thinking in suggesting Joe Blanton start game 1 over Roy perfect-game-and-no-hitter-in-the-same-season Halladay?
Well, here's the basic idea up front: what if pitching Joe Blanton in game 1 gave the Phillies a lower chance of winning in the first game of the series but increased their chances of winning in games 2 through 4 as well as games 6 and 7? Would you take that bargain?
So let me back up and explain. As everyone knows, the typical rotation goes from best starter to worst starter. In the world of descriptors, we call the best starter the #1, the second-best the #2, and so on down to the #5 starter. In the regular season, teams start off the season with their #1 and progress through games with each successive starter, looping back to the #1 after the #5 starter pitches.
Without injuries or off-days, what this does is it gives us match-ups of starting pitchers who are roughly equivalent. The Phillies throw their #1 starter against the other team's #1 starter, their #2 against the other #2, and so on down the line. If the teams have similar quality starting pitchers, these match-ups ensure that each team has its best chance of winning each game.
But what if there's a shift that sacrifices one game for a gain in the other four? What if, instead of starting with your #1, you start with your #5 against the other team's #1? Obviously, in that first game, you've taken a big hit. You certainly haven't decreased your chances of winning to 0, but you've lowered them no doubt. If both #1 starters are of equal quality, say the team would normally win 65% of that starter's games, then in a #1 against #1 match up, each team has roughly a 50% chance of winning. But, if now you have a #1 against a #5, you've lowered your chances of winning to, say, 30%. That's significant.
However, in doing so, you've increased your chances of winning the other games. Now, you have your #1 going against the other team's #2, your #2 going against the other team's #3, etc. So, you've increased your chances in four games by sacrificing your chance in 1 game.
It's important to note that there are a lot of assumptions. The main assumption, which is the reason I don't think you ever see this in the regular season, is that rotations quickly change - off days, injuries, weather, lefty-righty match ups, suspensions, double headers, etc. All of these things throw off the 1 through 5 progression so you're not very likely to have the perfect 1 against 1, 2 against 2, etc. lineup. Also, most teams don't have the same quality starting pitching. In the regular season, for instance, most teams didn't have a #1 starter to match Roy Halladay. Even more so, almost no team had a #2 or #3 to match Roy Oswalt or Cole Hamels. The Phillies had decided advantages when those three pitched, no matter where the other teams were in their rotations.
This NLCS is different though. In this series, the Phillies and Giants can line up their rotations however they want. And, in this series, the Phillies are facing a team that has a 1-2-3 punch of Tim Lincecum, Matt Cain, and Jonathan Sanchez that, if not the same exact quality as the Phillies', is surely a good match.
Although Charlie Manuel says that he's not sure who will start Game 1, I'd bet at least one of my children that Roy Halladay is his man. Starting this way means that Halladay will face Lincecum, Oswalt will face Cain, Hamels will face Sanchez, and Blanton will face Madison Bumgarner. Games 5 through 7, if necessary, will be repeats of 1 through 3. Given the quality of pitchers here, these will be some pretty even matchups (although I still think the Phillies will have the edge in all of them, if only slightly).
But we don't have to make these games even odds. Now is the time to shift the rotation and throw Joe Blanton in game 1. Blanton against Lincecum is a mismatch in the Giants' favor, yes, but it's not a lost cause. The Phillies won 11 of Blanton's 18 starts this year. He has pitched well since the start of July, with a 3.84 ERA in his last 17 games over 101 innings. He's no Halladay or Lincecum, obviously, but he's also no slouch and the Phillies can certainly win a game started by Blanton against Lincecum. I wouldn't bet on it, but it's not at all impossible.
Giving up the pretty even match up in game 1 would give the Phillies the advantage in the other games. Game 2 would be Halladay against Cain. While Cain is a very good pitcher, he's no Roy Halladay, and there isn't anyone who would argue he is. Roy Oswalt against Jonathan Sanchez in Game 3 would also favor the Phillies, although it would be closer than the Halladay game. Game 4 would be a huge shift for the Phillies, as Hamels against Bumgarner would be almost as lopsided for the Phillies as the Lincecum/Blanton game was for the Giants.
One big drawback of this approach would be that Blanton would have to start 2 games. The schedule is such that game 5 would require Blanton to make a second start. Again, Blanton versus Lincecum would favor the Giants, but Games 6 and 7 would be back in the Phillies favor. Plus, if there is a game 7, the Phillies can team up Oswalt and Hamels.
There's no doubt this is against the conventional wisdom and is risky. But, because Blanton is a very solid pitcher and because the other games would shift substantially in the Phillies direction, shifting the rotation this way is the right thing to do.