(Note: this article was originally written in August of 2016). 

I'm relatively new to MFLs, the My Fantasy League best-ball format where you draft 20 players, knowing you'll start 1-QB, 2-RB, 3-WR, 1-FLEX, 1-TE and 1-D each week. It's "best ball" because you don't set your lineup (or make any moves) - the software just takes your highest possible scoring lineup each week. The other 11 players remain on your bench. It's also PPR scoring.

But after drafting four of these leagues so far - I might sign up for a couple more at higher stakes - I have a better idea about what I want to do: Use at least three of my top four picks on running backs. In one of my leagues I did just that, taking David Johnson, Devonta Freeman and Mark Ingram, before taking receivers with six of my next seven picks.

In a normal league where you have to set your lineup each week, not having star receivers is a big problem. Not only do they score the most points in a PPR format, but you need more of them. Moreover, once you get outside the top 15-20 or so, it's hard to predict which week a WR will go off. It's actually hard to predict which week even a Dez Bryant or an A.J. Green will go off, but because they're so good, and you never remove them from your lineup, you don't have to worry about that. But for mid-level receivers like DeSean Jackson or Tyler Lockett, you might sit them after a few games with low-target totals and production for someone who put up a couple 7-for-70 lines and watch them catch a 50-yard TD or two from your bench. Simply put, if you're messing around with mid-level WR, it's hard to time them, and they're not good enough simply to keep active no matter what.

But in best ball this problem goes away. You don't have to guess when Jackson or Lockett will go off because the software will make sure they're in your lineup on those days. Suddenly receiver volatility isn't a liability but an asset. You can make up for not having a star on whom you can count by having 7-8 volatile types that collectively will blow up enough to carry you.

For running backs in best ball, the opposite dynamic is at play. In regular leagues, drafting mid-level running backs isn't a problem because you usually know in advance whether they're getting touches in the short term. If you have Gio Bernard and DeMarco Murray, you can be fairly sure what their roles are most weeks, so long as they're healthy. Same with Danny Woodhead and Charles Sims. There's not a ton of guessing whether they'll catch a few passes and get a handful of carries.

But while running backs are more stable game to game, they're less stable year to year. Running backs are bigger injury risks, and few mid-level ones are good or established enough not to get replaced when their production suffers. In standard leagues, you can deal with that problem by picking up their backups, working the wire and making trades. In best ball, you're stuck with the ones you picked all year. For that reason, you need to have the safest of the bunch - they're still riskier than top WR, but they're much less risky than mid- and later-round backs.

So the "best-ball" mechanism reverses the risk calculus for the two positions. Instead of guessing when your mid-level WR will go off, you just grab a bunch with stable roles but erratic game-to-game production and find safety in numbers. And instead of punting RBs which can be mined from the wire and figured on a case by case basis all year, you lock in the ones most likely to keep their jobs and see heavy workloads, barring injuries.

In both league types you're finding stability in the volatile positions by getting the most stable of the volatile (WR in standard, RB in best ball), and increasing upside in the safe position by embracing variance (RB in standard, WR in best ball.) In some ways, regular fantasy football is more like a hybrid of season-long and DFS, and the best-ball format is the true season-long where your roster is stuck for the entire year.

As for TE, QB and defenses, I'm starting to think you just take three of each, mostly in the second half of your draft. Ideally, I'd have three star RBs, 8-9 receivers with an excellent shot at 90-plus targets, and 2-3 QB, TE and defenses (three where the players are worse, two where they're better.) While I've read articles from people far more experienced than me arguing for taking a star TE (or two) early, I don't see myself pulling the trigger on Rob Gronkowski this year (no Tom Brady for four weeks, only 72 receptions in PPR last year) over a David Johnson or Todd Gurley.