Last week over on Beyond the Box-Score, Sky Kalkman wrote a piece entitled Optimizing Your Lineup by The Book, which summarized the guidelines offered up in The Book by Tom Tango, Mitchel Lichtman, and Andy Dolphin. Kalkman kindly agreed to write a piece analyzing the best lineup for the Diamondbacks, using those rules. It's probably best if you start off by reading the piece, as it goes into more detail about the requirements, but for those of you with short attention-spans, here's the Cliff Notes version:

Leading off, OBP is king - you should have one of the best three hitters on the team here, the one without homerun power.  Speed is nice, but less essential than you think. The second spot is also one of the top batters, with good OBP skills, but the third spot is much less important that you'd think, as he comes up with fewer people on base than the four or five hitters. The cleanup man is the other top-three hitter, the one with power, and the fifth spot goes to your next best hitter, "unless he lives and dies with the long ball." After that, stolen bases are valuable, then the singles hitters in decreasing order of talent. It is marginally better to have the pitcher hit eighth rather than ninth - it gives the top of the order more chances with someone on base - but it's worth only about two runs per season. I'd say that's probably largely negated by the stress on  the manager having to explain it, and on the poor guy who hits ninth.

Kaufman's analysis is after the jump, along with my fiddling with numbers. As a comparison, here's what looks likely to be the actual Diamondbacks' line-up this season:

1. Lopez (S)
2. Young (R) or Drew (L)
3. Drew (L) or Young (R)
4. Jackson (R)
5. Tracy (L)
6. Upton (R)
7. Reynolds (R)
8. Snyder (R)
9. Pitcher

Using the CHONE projections and ordering the projected starters by runs-per-150, we have:

First off, it's obvious CHONE doesn't like the Diamondbacks' offense, although there's plenty of upside for Young, Drew, and Upton.  The top three hitters consist of Jackson and Young, plus one of any other number of guys.  With Jackson being by far the best OBP option, his lack of reliance on homeruns, and GIDP tendencies, he's actually a good choice to lead off.  Chris Young has more power than on-base skills, so he's fourth.  Plus, stealing bases is actually better later in the lineup in front of singles hitters.  Then we've got Upton, Drew, Tracy, and Reynolds to pick from for the second spot.  With Upton's higher OBP and strikeouts (they don't result in GIDPs), he's the guy.

The next to spots to fill are 5th and 3rd.  Drew, Tracy, and Reynolds are the candidates.  Putting a lefty in third would really be nice, and since Drew hits more homeruns, he's the guy.  Then Tracy fifth, again to split up righties.  Six through nine would go Reynolds, Lopez, Pitcher, Snyder based on production.  That would also result in a RRLRLRSXR batting order, which is nice.  One consideration is to put base stealers in front of singles hitters and power hitters behind players who can't advance themselves from first very well.  That would mean Lopez directly in front of Snyder, perhaps, but it's not like Snyder hits all that many singles anyways.  His walks are valuable immediately in front of Jackson at the top of the lineup.

Honestly, that's a strange collection of players.  It's a lot of low-AVG, medium-BB, decently high ISO players.  And many of similar talent level.  In those situations, different lineups aren't going to matter much, as long as you put the top five hitters in the top five spots.  That means a significant rallying cry for Diamondbacks fans would be for Felipe Lopez to be removed from consideration to hit first or second.

1. Jackson R
   
2. Upton R
   
3. Drew L
   
4. Young R
   
5. Tracy L
   
6. Reynolds R
   
7. Lopez S
   
8. Pitcher
   
9. Snyder R
   

Sky Kalkman

I've taken a somewhat different approach, using Baseball Musing's Lineup Analysis tool, but plugging in the same nine players as Kalkman used, and with numbers also taken from the CHONE projections. The best line-up, as calculated by the tool, comes out at scoring 4.935 runs per game, which would be 799 over a 162-game season. That is certainly an improvement on the 720 scored last season, but wouldn't take into account any variation from the lineup either negative (injuries, players needing to be rested) or positive (pinch-hitters). Here's what that reckons is the optimum Diamondbacks line-up:

1. Jackson R
   
2. Drew  L
   
3. Snyder R  
   
4. Young  R
   
5. Upton  R
   
6. Reynolds R
   
7. Tracy L
   
8. Pitcher    
   
9. Lopez S
   

Things that stand out: both approaches reckon Conor Jackson should be batting lead-off for the 2009 Diamondbacks, with Chris Young hitting clean-up, and Mark Reynolds occupying the sixth spot. The surprising thing is probably seeing Chris Snyder as high as the number three spot - however, this is a consistent aspect of the top lineups, with nine of the best ten having our catcher there [the tenth  simply swaps him and Upton]. That may tie in with the 3rd spot not being regarded as very important, per The Book, but may also reflect that the CHONE projections have Snyder with the best walk rate of anyone on the team.

There are obvious downsides to this strictly-numerical approach. Most obviously, it doesn't take into account splits vs. LHP/RHP which would obviously be used to construct lineups in real-life [I couldn't find any projections which offered splits]. Instead, it's a homogenized combination that probably reflects neither the best line-up versus a southpaw or a rightie. Nor does it reflect an ebb and flow over the course of a season: Bob Melvin is clearly a fan of the 'hot hand' and such variations will affect what the best lineup would be. Finally, all the "best" generated lineups have a chunk of four or more right-handed batters, which would be just asking for the opposing manager to bring in a left-handed reliever.

It should also be noted that the amount of time spent - by fans and managers - working on lineups is probably out of proportion to the importance it actually has. Per Tom Tango, "an optimal batting order can get you 5-15 more runs in a season. Not negligible to be sure, but it doesn't seem worth the anguish we felt. every one of the thireen times we saw a solo home-run from Chris Young or Stephen Drew at the top of the order last season.

Jim McLennan