From one aspect, if your goal is to make the best team possible for the purpose of winning the game, doesn’t taking Harper and Puig (injuries aside) make the most sense? They’re some of the most talented players in the game. But if you are talking from a standpoint of performance, can you still justify taking them?
It helps to start with an idea of what we’re looking for. What has traditionally it meant to be an All-Star? Well, there are 34 players on an All-Star team, to start with. It ends up being much more than that, due to injuries and such. Last year, for example, 11 outfielders made the National League team-9 were on the “active” roster, while Matt Kemp and Giancarlo Stanton sat out. So we’ll say the top ten outfielders are the All-Stars, just to give us a rough framework in roster construction.*
*Actually, the past two seasons have had eleven NL outfielders, but we’ll say ten to keep it safe.
Are Puig and Harper among the top ten outfielders in the NL in 2013? Well, fan voting has Harper fourth, but it’s probably better to ignore that based on past voting. Starting with Wins Above Replacement, then, we get Harper with 1.6 (both Baseball-Reference and Fangraphs) and Puig with over 2 (B-R likes his fielding more, 2.6 to 2.0. Both numbers put Puig comfortably in the top ten, while Harper falls to nineteenth in each one. So that’s a point for Puig, at least.
It’s fair to ask about Puig’s playing time, or lack thereof. With only 28 games, he won’t even have half as many games as someone who’s been on the roster the whole year by the All-Star Break. But, at the same time, he’s hitting .440/.466/.743. Weighted Runs Created+ (wRC+) puts that line at a 239, while OPS+ has it at 233. Either way, that means he’s hitting something like 130% better than an average major league hitter. There’s no way it’s sustainable (he has a BABIP over .500), but just because it’s a fluke doesn't mean it didn't happen. He may have half the games, but he’s hitting twice as well.
The only real knock for Puig is if you think the All-Star game should reward the past calendar year. In that case, his 28 games don’t hold up as well...although Fangraphs still puts him 28th in WAR for NL outfielders in that time.
Plus, if you want to use the calendar year as a knock on Puig, you almost have to take Harper instead. Frangraphs has Harper fourth among NL outfielders over the last 365 days, with his 133 wRC+ in that time only rising (he has a 156 wRC+ this year). His 4.2 WAR in that time puts him in a tie with Starling Marte and Jason Heyward, behind only Andrew McCutchen, Carlos Gomez, and Ryan Braun. B-R has him fourth since the start of 2012 with 6.8 WAR behind the same three.
Could Harper even have a case for just this season though? Again, he has 1.6 fWAR, nineteenth in our leaderboard. Some of the players ahead of him are more debatable, too. Evan Gattis has 1.8 WAR, but has 3-and-a-half times as many innings as a catcher, so it seems weird to lump him in with the outfield. Nate Schierholtz has been used more as a platoon bat, with only 29 plate appearances against lefties. It’s fair to assume that those numbers would drop if he were full time.
On top of that, WAR is not exact. Another 3 runs (0.3 wins) would move him up to tenth. That isn’t exactly a large margin. For example, Harper currently loses half a run in his fielding. That isn’t necessarily weird...but fielding metrics are still very rough. Large samples are need to gauge true talent. Might Harper be better than just average? Well, according to his fielding value last year, he was worth a full win on defense. Assuming he’s as good a fielder as his stats last year said (we’ll say around 2.5 runs through 44 games for back-of-the-envelope calculation purposes), Harper’s value is closer to tenth among NL outfielders. Even if you average the two (around +1 run, rather than -0.5), you wind up with Harper in thirteenth, right on the borderline.*
*And again, we’re probably underestimating the number of spots, based on the past two years.
And that’s not taking in to account that he still has over a week of games until the game itself. There’s still room to change, although that applies to the rest of the field, too. But the point is, depending on how you weigh it, Harper as one of the ten best outfielders in the NL just based on 2013 isn’t too outrageous. And if you’re the type who thinks career or talent or time lost to injury or the past year should count for something, Harper is probably one of your better choices.
The same goes for Puig. His play so far has been good enough to merit a spot. He definitely has the talent, as he has been viewed as a strong prospect. Really, the “past year” category is his only lagging area (and if you really want to go by past calendar year, you have no excuse to pass on Harper).
Really, you can't get too upset about either choice. Neither is probably “the best” NL outfielder at this moment, or even worthy of one of the three starter slots, but there are way more “All-Star” outfielders than you probably realize. Once you start accounting for the likely injury sit-outs, you almost have to be trying to keep both Harper and Puig out; putting them both on the rosters isn’t really a bad idea.