The 2018 season is still a few weeks away, yet you already know a few things about what will happen. For example, there's almost no scenario where top teams like the Astros, Dodgers or Yankees aren't among the best in their leagues, thanks to incredible amounts of depth and talent. At the other end, it's very difficult to see enough things going right all at the same time for rebuilders like the Marlins, Tigers or Royals to end up in the playoffs.
But in the middle, there's a ton of possible outcomes. We saw this last year, where everything that could have possibly gone wrong for the Giants did as they lost 98 games, while more things went right for the Twins than they could have possibly expected, pushing them from 59 wins to 85 and an American League Wild Card berth.
It's that middle ground that we're most interested in. Which teams are the most volatile, which is to say -- which teams have the most different possible outcomes? Who are the most high-variance clubs? You couldn't possibly imagine the Yankees losing 100 games, for example, but for a team like the Mets, who have a talent level exceeded only by their propensity for injury, you could envision a division title or a fourth-place finish.
How do we define that? Let's be honest: There's a lot of math here, probably more than you're interested in, so rather than going too deeply into the weeds, we'll give you the short version.
First, we looked at 2018 numbers from the respected Steamer projection system, specifically the difference between a team's 90th percentile projections (essentially the best-case, everything-went-right scenario) and 10th percentile projections (the darkest timeline, as it were).
For an established star like Chris Sale, that's relatively narrow, because if he's healthy, you know he'll be productive. While there's not a breakout coming (how could he possibly be better?), it's almost impossible to see a collapse, either. That being the case, the gap between his best-case scenario (2.44 runs allowed per game) and worst (4.48 runs per game) is barely two runs. But for a total question mark like Timothy Lincecum, his range is much wider, with a large 4.32 run difference between his best-case (3.53) and worst (7.53) likely outcomes. You could see him being surprisingly effective, or out of baseball by May.
It's the same thing for hitters, too. For the veteran Nick Markakis, there's relatively few likely outcomes, and the gap between his best-case (.350 wOBA) and worst (.286) is only 64 points. But for a talented yet unproven prospect like Houston's Kyle Tucker, we're talking nearly 200 points of uncertainty here. 
We did that for hitters and pitchers on a team level, finding the clubs with the smallest gaps between the best- and worst-case outcomes. Then second, we excluded the seven teams currently projected to lose 90 games, because we're trying to find teams potentially in the Wild Card mix with the most possible outcomes.
Doing it this way, we find that the Dodgers, Cubs and Astros are the three "safest" teams in baseball, in terms of expected outcomes, and that makes plenty of sense. The most volatile team is the White Sox, and that makes sense, too, since they're relying on a ton of young, unproven players. But they're knocked out since they're projected to lose 90 games, which leaves us with these five feast-or-famine clubs.
Orioles
Here's what you can be confident in with the Orioles: Adam Jones will probably put up the same roughly league-average season he does every year, and … well, that might be it. In the past five years, Chris Davis has had two superstar seasons (2013, '15), one good year ('16) and two disasters ('14, '17). Manny Machado is a young superstar, but his 2017 was a huge step back from '15 and '16. Timothy Beckham had one of the most dominant months of any player in baseball last year, yet little in his track record indicates that's just "who he is" now. It's impossible to know what to expect.

Apply that to the rotation, too. Kevin Gausman and Dylan Bundycould both be good, because they're talented, but each had ERAs over 4.00 last year. Chris Tillman used to be good, but was arguably baseball's worst starter last year, and Zach Britton is injured. There's talent here, it's just incredibly hard to know what versions of these guys we'll see.
Rangers
Texas, like Baltimore, is coming off a disappointing season, having lost 84 games, and despite a string of moves aimed at reinforcing the pitching staff -- adding Mike Minor, Chris Martin, Doug Fister, Matt Moore, Jesse Chavez and Lincecum -- its current projection is all of 78-84. The pitchers are still projected to allow the fourth-most runs per game, at 5.10.

This one isn't hard to explain. All of the names mentioned above, as well as returning Rangers like Cole Hamels, Rougned Odor, Nomar Mazara and Shin-Soo Choo, have had past success. The problem is that very few of them had good seasons in 2017, which is how you lose 84 games in the first place. We know Adrian Beltre is a stud, yet he's also 39 years old next month. If everyone plays to their potential, and Willie Calhoun mashes like he's capable of, this team could be in the American League Wild Card Game. If even a little goes wrong, 90 losses are just as possible.
Phillies
Philadelphia lost 96 games last year, but it has an impressive young core, and it has since added Jacob Arrieta, Carlos Santana, Tommy Hunter and Pat Neshek. Everyone expects this team to be better, and even the modest 75-87 the Phillies are currently projected at would be a nice nine-game jump, if not into the postseason conversation.

Now, could the Phils be better, good enough to get into the race? For example, the extremely consistent veteran Santana has a relatively limited range of outcomes, from a .398 wOBA in a great year to a .330 mark in a poor year, yet he stands alone in terms of reliable production. For someone like young catcher Jorge Alfaro, to choose just one name, it's possible he could be above average (.332 if everything breaks right) or a disaster (.237 in the worst case).
Up and down Philly's roster, you see these questions. Does Rhys Hoskins slug like Babe Ruth or "merely" an All-Star? How quickly will J.P. Crawford or Scott Kingery contribute? Will we get the good, healthy versions of Nick Pivetta, Jerad Eickhoff and Vince Velasquez, or the ones that helped put up a 4.55 team ERA last year? As you'd expect, young pitching is strongly correlated with uncertainty.
Athletics
We're already on the record as saying the A's roster is actually extremely interesting, despite their 75-87 record in 2017 and their 80-82 projection in '18, and that should tell you about the wide variance in possible outcomes here. This is, unsurprisingly, due to youth, as none of the top eight starters (the seven atop their depth chart, plus rising prospect A.J. Puk) are 30 years old; last year, no AL team had fewer starts coming from 30-plus pitchers.

This extends to Oakland's lineup, too. We know Matt Olson mashed like an elite star last year, but also that he did it in only 59 games. New addition Stephen Piscotty has been good in the past, yet he wasn't in 2017. There's a lot of interesting young talent that could come together quickly, but there's also an argument to be made that the A's are the fifth-best team in their own division. We know the Astros will be good, for example. We really don't know how good Oakland will be.
Rockies
You'll find this one surprising, probably; the Rockies made the National League Wild Card Game last year, and then made a lot of bullpen additions. Nolan Arenado and Charlie Blackmon were studs in 2017, and can be expected to be so again in '18.

But Colorado is relying on a ton of young starting pitching, and if you're noticing a trend here, it's that projecting young players without much of a track record is extremely difficult. The oldest potential Rockies starter this year, Chad Bettis, turns only 29 in April, which means that Colorado is counting on a lot from young arms like Antonio Senzatela, Kyle Freeland, Tyler Anderson and German Marquez, who all had decent but not dominant seasons last year. You can be as certain that at least one won't succeed as you can be that one will be very good.
Beyond the "big two" on the hitting side, it's more of the same. David Dahl has the talent to be an All-Star, yet he missed all of 2017 to injury. Ian Desmond has been an All-Star, yet he's had two poor seasons in the past three. Trevor Story followed up a smashing '16 debut with a disappointing '17 sophomore season, and young Ryan McMahon, in line to start at first base, has 24 Major League plate appearances. They all could succeed, certainly. But this isn't like looking at Michael Trout or Joey Votto and being confident in a great year to come, because there's as much uncertainty here as there is skill.