My organization here is a bit funny. I wanted to start by explaining how I get Harris as 20 runs better than Punto on offense over 150 games, and how I get Punto as 20 runs better than Harris on defense over 150 games, but it turned out to be perhaps the most unreadable stuff I've ever written. If you find either of these claims contentious, I left my justifications at the bottom.

Harris is basically Bizarro Punto. If you don't do anything creative with them, they have essentially equal value. Harris will make you feel better at the plate, and Punto will make you feel better in the field.

I would propose using Punto and Harris in a defense/offense alignment. Start Punto on defense. On average, the #7 spot in the lineup is going to come up in the 2nd inning, 4th inning, 6th inning, and 8th inning. If you take Punto out whenever his spot comes up after the 6th inning, or if there is a crucial pinch-hitting situation around the 6th inning, say he gets 6 innings of defense on average and Harris gets about 1.5 PA/G. That way Punto gets roughly 67% of the playing time on defense and Harris gets about 35% of the playing time on offense. If that happened, it would be roughly a breakeven strategy from a run differential standpoint. You'd have something like:

-1.7 R/150G -- Defense, 67% Punto
67.1 R/150G -- Offense, 35% Harris

Which from a run differential standpoint is about what you would get if you just ran Punto or Harris out there for an entire game. So, that seems like a lot of work for not much benefit, I suppose.

The main reason I like this strategy is leverage, though, especially with Everett and a banjo-hitting CF also in the lineup. If it's late in the game and the Twins are clinging to a small lead, then by all means keep the good defenders on the field. If the Twins are down by one run, though, Gardy can put Harris in the game either to try to generate baserunners or to drive in baserunners. And say the Twins carry Buscher as the 25th man. If it's the 8th inning and Punto, Lamb, Everett is due up, we can make that Buscher, Lamb, Harris. Will the defense in the 9th inning be ugly? Sure, but you're going to lose anyway if you don't score any runs and the bullpen generally misses more bats than the starting pitchers, so they don't rely as much on the defense.

If it's the 6th inning and Kubel, Punto, and Lamb have loaded the bases with two outs, and you're down by a couple of runs, you have the chance to pinch-hit Buscher, Harris, or Monroe for Everett. If you decide to hit Punto 8th, then you could wind up in a situation where you are down by one run in the 7th or 8th inning and Punto, Everett, and Pridie are due up with Harris, Buscher, and Monroe on the bench. It wouldn't exactly be Murderer's Row, but teams go to so much trouble to leverage their pitchers properly, I don't see why we shouldn't try to leverage the hitters a bit, too.

I'd still try to get some starts for Harris against LHP, especially with our fly ball pitchers on the mound. The mixing and matching could even go beyond just Punto and Harris. With a groundball guy like Liriano, I'd want Punto on the field, but if the Twins were facing a tough LHP, that would be a good opportunity to give Lamb a day off, since the other team is probably going to send a lot of right-handed hitters up to bat, who are more likely to hit the ball to the left side of the infield.

There's a lot of concern in some areas about how many innings the non-Livans can handle, and having better defense at the beginning of the game can help them grind out more innings.

Basically, it's easier to see the opportunities on offense coming than it is to see the opportunities on defense coming, and there will in general be more opportunities on defense early in the game. So start the game with your best defensive squad and send in the subs as necessary. Embrace your inner Punto. You know you want to.

Offense

The 2008 PECOTA projections for the Twins are apparently more controversial than Sinnead O'Connor's SNL appearance, so for offensive projections, let's turn a moment to fangraphs, where they report the Bill James, CHONE, and Marcel projections. To be diplomatic, I'll throw out the highest and lowest by RC/27 and keep the median projection for Punto and Harris. This gives us:

48 RC, 517 PA -- Punto (CHONE)
65 RC, ~509 PA -- Harris (Bill James)

Say the team OBP is around .330. That gives us about 40.3 PA over 27 outs. Teams don't always get 27 outs in a game, but you adjust up for extra innings and down for home wins, so I'm going to leave it at 40.3 PA/G. That gives the #5 hitter about 4.48 PA/G. It can be shown that the Nth hitter gets about one-ninth of an at-bat more than the (N-1)st hitter, so the #7 hitter should be getting about 4.26 PA/G, or 639 PA over 150 games. So if we want to think in terms of a full season:

59.3 RC/150G -- Punto
81.6 RC/150G -- Harris

So Harris has about a 22-run edge on Punto offensively, which is significant.

Defense

UZR doesn't have Harris' time at second base in 2007, and his data is pretty small sample size, but oh what a sample. He's at -38 runs/150G at 2B, -33 runs/150G at 3B, and -22 runs/150G at SS. Our dearest Nicky is at +10 runs/150G at 2B, +13 runs/150G at 3B, and +14 runs/150G at SS.

Using PMR, as a team, the Devil Rays (it was 2007 after all) had the 3rd-worst defense at SS overall last year, and Harris had the third-worst rate amongst shortstops, ahead of only Derek Jeter and Cristian Guzman. A full season of PMR predicted outs at SS seems to be about 500, which would put Harris at about -37.8 plays per season or roughly -19 runs/season.

Punto didn't play enough 2B or SS in 2007 to get mention in the PMR tables. At 3B, he was slightly below average, with roughly -9.3 plays per season or roughly -5 runs per season. In 2006, he was about +9 plays per season or roughly +5 runs per season. In 2005, Punto did get enough playing time at second base to qualify for Pinto's list, and he was nearly off the freakin' charts. If a 2B gets about 450 predicted outs over the course of a full season, then Punto would've been at about +69.7 plays per season or roughly +35 runs per season, which would be positively Everett-esque if I didn't suspect it was a bit of a fluke.

In SFR, 600 balls seems to be roughly a full season at second base. That makes Punto roughly +6.7 runs and Harris roughly -20 runs over the course of a full season at second.

From 2004-7, Harris has a .813 RZR at 2B (only 92 balls in his zone), and Punto at an .832 RZR at 2B. A full season of RZR at 2B seems to be roughly 350 BIZ, so that's a difference of just 6.6 plays per season between Punto and Harris. (Though I would note that RZR doesn't take into account as many plays as UZR, PMR, or SFR, so this is going to tend to underestimate the difference between fielders relative to other systems.) Punto's RZR at SS has been .854 and Harris' RZR at SS last year was .777. A full season of BIZ for a SS seems to be about 400, so that'd make the difference between Punto and Harris about 31 plays in a season.

So, that's a lot of different numbers, but if you throw out the high and low scores from the Soviet judges, then I think that it's pretty reasonable to put Punto at around +5 runs per 150G at 2B and Harris at around -15 runs per 150G at 2B.

+5 R/150G -- Punto
-15 R/150G -- Harris

Which gives Punto a 20-run edge on defense.

Overall

So, honestly, overall I see little difference in total value between Harris and Punto if you're just blindly putting them in the lineup every day. It's kind of a matter of whether you want to win 3-2 games or 4-3 games.