Do Match-ups Work?

With three weeks left to go in the baseball season, there's not a lot of roster maneuvering left that's going to make much difference in the overall standings of any fantasy league.  The trade deadline has passed and your players will either play like you think they are capable or they won't.  It's pretty much out of your hands.  But if one hit or one strikeout can determine who wins your league, it's not a bad idea to pay attention to the hitters' match-ups.  I'm not talking about player versus team match-ups.  I'm talking about breaking the game down to it's most basic components: the batter versus the pitcher.  Thankfully, there are several sites, being the most notable, that provide these breakdowns.

For example, Brian Jordan finally looks healthy and is playing well.  But how much playing time will he get?  Gary Matthews Jr. has earned Buck Showalter's praise for his defense so as soon as he's healthy enough to play every day, he very likely will.  Kevin Mench has been hitting the cover off the ball this month so he'll likely get the everyday nod as well.  Jordan will have to compete with Laynce Nix and Dave Dellucci for at bats the rest of the way. 

He is slated to face only lefties, but neither Dellucci or Nix is hitting the ball often (.174 and .125 in September respectively) so he may get more than a couple days a week.  Even if he ends up with a just straight platoon, Jordan's production could be pretty decent.  He's hit .333 versus Oakland's threesome (Mark Mulder, Mark Redman and Barry Zito) but not so great against Jarrod Washburn (.167), the only other lefty starter in the division he's faced. 

Conversely, Dellucci has had little success against Oakland.  Tim Hudson is the only starter he even has a hit against and even then he's had only marginal success - just three singles in 11 at bats.  Against Mulder, Zito and Rich Harden he's hitless with no walks.  Against the Angels and Mariners he's only slightly better.  He's hit Colon well (.429) and acceptably against Gil Meche (.286 with on base of .375) but against Aaron Sele (.231), John Lackey (.125), Escobar (.250) and Ryan Franklin (.143) he hasn't done a whole lot.  I'm not saying he couldn't do well given enough chances.  But when a manager is filling out his line-up card and trying to decide between two players, if he doesn't go with his gut instinct he'll look at the hitters' batting averages against the pitchers they're facing regardless of how small the sample size is and decide from that.

As for Nix, he's hit Hudson (.400), Meche (.500) and Sele (.600) but has had little success against Harden (.167), Escobar (.111), Colon (.250) or Lackey (like Dellucci against Hudson, a completely empty .273).

Assuming Jordan will get the nod against the lefties, Showalter will be looking for the best match-ups against the right-handers and his answer may be the same for many of them.  Jordan has superior numbers against Escobar and will probably get the nod against Lackey and Harden because he's doesn't have their record of failure, plus he's the hotter hitter (.350 in September).
So despite his .214 average for the season, Jordan makes a good percentage play for the final few weeks in an AL-only league where an at bat or two can make the difference.  Yes, the sample size with each pitcher is small, but taken in aggregate with all the pitchers each hitter will likely face, you can get a decent idea of what to expect.

As with pitchers, it doesn't always work out.  For example, Nate Robertson faced the Twins on Sunday.  Minnesota is one of the worst hitting teams against lefties in the AL, both in terms of runs scored and OPS.  Yet they pounded Robertson.  Bloops, bleeders, bombs... they did it all in 4 innings: 10 hits, 2 walks and 6 earned runs.  Robertson hasn't been as good at home as he has on the road, but the split wasn't so dramatic as to expect that kind of thrashing.

But anything you can do to reduce the risk of falling in the standings or losing out in a head-to-head match-up has a good chance of making the difference in a close race.