One of the byproducts of sabermetrics has been the change in how we view managers. For starters, general managers are now the off-the-field face of the organization. There was a time when Earl Weaver had a big role in determining his 25-man roster or Davey Johnson could tell Frank Cashen he wanted a 19-year-old Dwight Gooden on his team. Now the general manager pretty much says, "Here are the players I'm giving you." As a result, we discuss general managers and roster building and the like as much as we discuss in-game decisions. Where we used to rail against managers, we demand that GMs be fired.

Think about this also: We talk about managers in terms of, well, managing. How they manage players and their egos. How they manage the bullpen. How they manage pitch counts. How they manage a young player. How they manage the media.

Less often, we talk about them in terms of strategy and tactics. This picks up in the postseason, of course, as we scrutinize every pitching change and sacrifice bunt and realize nothing Ron Washington does seems to make sense, but the regular season is dissected and analyzed more in a big-picture mindset.

Then sabermetrics piles on and says a lot of the decisions managers make aren't really all that important: Lineup order doesn't matter all that much, one-run strategies are overrated. Even all the shifting we see these days? That's coming from the front offices and the stat nerds, not the manager on the bench.

But then we get games like Wednesday night's at Citi Field between the Atlanta Braves and New York Mets, a reminder that the big picture consists of 162 little pictures, and some of those little pictures depend on a key decision from the manager. Score a big 3-2 win here for the Braves in their battle for the wild card, with a big tip of the cap to skipper Fredi Gonzalez.

Here's what happened. After Andrelton Simmons made perhaps the defensive play of the season to save a run in the bottom of the eighth, the Braves took that 3-2 lead into the bottom of the ninth.

All-Star closer Craig Kimbrel came on and showed some of the wildness that has made him a little less dominant this season (his ERA entering the game was all the way up to 1.76, and he'd blown four saves). Eric Campbell singled sharply to right field on a 3-2 fastball. Matt den Dekker got ahead 3-0 and eventually walked on a 3-2 fastball.

Due up for the Mets: Wilmer Flores, Ruben Tejada and the pitcher, Flores hitting .224 and Tejada .228. With David Wright and Daniel Murphy both apparently unavailable with injuries, the Mets' bench was thin, so manager Terry Collins didn't really have any pinch-hitting options since he had to save a hitter for the pitcher.

Collins elected to bunt with Flores. That itself is debatable. I would have swung away, my theory being that getting the tying run to third base against Kimbrel is less valuable than against most pitchers because Kimbrel's strikeout rate is so high. Plus, he had just walked a batter and has been wild all season, so who knows what happens if you don't give him an out. I'd rather hope to go 1-for-3 than 1-for-2.

Flores got the bunt down and both runners moved up, bringing up the light-hitting Tejada, who has just 12 extra-base hits in over 300 at-bats. Gonzalez faced a tough decision: Bring the infield in to cut off the tying run but increase the probability of a grounder or line drive going through the infield and winning the game for the Mets, or keep the infield back to at least preserve a better chance of keeping the game tied and sending it into extra innings.

This is a situation in which the numbers can't provide a "right" answer. You could attempt to analyze the probability of Tejada hitting a ground ball (41 percent of the time when he puts the ball in play) against Kimbrel, who allows grounders on 43 percent of his balls in play. But then you have to factor in that Kimbrel didn't look sharp. And you'd have to factor in the odds of Tejada hitting a hard grounder or a slow grounder, let alone a line drive.

Oh, and you have about 10 seconds to make your decision. Good luck consulting the charts there.

Gonzalez had to make a snap decision. Maybe it wasn't that difficult; after all, with Kimbrel you have a good chance of a strikeout anyway, even against a solid contact hitter like Tejada. But it's one with enormous risk, no? Most managers are going to play it safe there; managers, by nature, are risk-averse. If Tejada hits a seeing-eye single through the drawn-in infield, the Braves lose and Gonzalez is vilified by the fans and the media.

I'm guessing that Gonzalez's primary consideration was that Tejada doesn't hit the ball hard. With that in mind, he brought the infield in.

It worked. Tejada hit a slow-roller to third base and the Braves got the out at home plate. Kimbrel then got pinch hitter Kirk Nieuwenhuis to fly out to shallow left and the Braves were a win closer in the wild-card standings, one game behind the Cardinals.

Sabermetricians often talk about the "process" -- stick to the right process and things will eventually go in your favor. Sometimes a right decision will backfire and a wrong decision will work. But it's the process that matters.

Well, sometimes it's the result that matters. Fredi Gonzalez went for the win and got it.