Fantasy football is a volatile game.

Sometimes, a shoelace tackle is the difference between a 10-yard catch and a 70-yard touchdown, and sometimes goal-line carries go to backup players.

It happens. A lot.

And, don't get me wrong -- median projections are quite valuable and capture the most-likely scenario. Setting your lineups based on 90th-percentile projections isn't the right way to handle things for a head-to-head lineup, but if you want to figure out which players can bust a slate open, then you'll have to embrace some risk.

That's why I've started simulating NFL weeks and seeing what happens when the slate is played out 1,000 times. Here are some things that popped at each position this week, based on my simulations, which factor in numberFire's projections and my own tweaks.

Table Terms

FanDuel Salary: A player's main slate salary on FanDuel.
Median FDP: A player's median FanDuel projection across the 1,000 slate simulations.
Value: Projected median FanDuel points per $1,000 in salary. All quarterbacks generally have a 2.00 FanDuel-point-per-$1,000 rate at the low end, which implies 2-times value, or 2x value. On a full slate of 13 games, roughly 13 running backs tend to have a 2x value projection. On a full slate of 13 games, a small handful of receivers may get to a 2x median projection, and just more than 30 will be at 1.5x. On a full slate of 13 games, few tight ends will get to a 2x median value, and anything above 1.5x is generally a top-six projection. It's important to understand the different value expectations across positions.
25th Pct: The player's 25th-percentile FanDuel point projection, meant to show a low-end (or floor) outcome. Every player's true floor is zero.
75th Pct: The player's 75th-percentile FanDuel point projection, meant to show a somewhat attainable/projectable high-end (or ceiling) outcome without simply looking at true outlier performances.
FDP%: The frequency with which a player surpassed a certain raw projection threshold, meant to show a raw ceiling outcome. This doesn't adjust for salary and is a different value for each position.
Boom/Bust Ratio: The frequency with which a player had a "boom" game compared to a "bust" game based on historical, position-based value outcomes. For quarterbacks, this measures games with 2.75x value versus games with worse than 2x value. For running backs, wide receivers, and tight ends, it's 2x versus less than 1.5x. Think of it as a simple floor-versus-ceiling rating. Higher is better, and they should be compared only within the same position.

Quarterback

Simulation Results

Observations:
Lamar Jackson, Deshaun Watson, and Justin Herbert rate out a cut above the rest from a median standpoint. Jackson and Watson combined for more than 25.0% of the QB1 performances on the slate and are the only two with double-digit odds to lead the slate in fantasy points. Their floor/ceiling ratings are easily best at the position.

Somewhat interesting to me is that Cam Newton projected fourth-best if we look at the boom/bust ratios. Cam's passing efficiency has climbed the past three games, but the passing volume remains low. It's a promising overall game environment against Watson's Houston Texans.

The next-best boom/bust ratios belong to a trio of lower-salaried quarterbacks: Tua Tagovailoa, Matthew Stafford, P.J. Walker, and Alex Smith. Walker, of course, is a provisional play because the Carolina Panthers could be starting Teddy Bridgewater still or even Will Grier. For now, numberFire projects it to be the XFL standout, Walker.

Two passers at higher salaries have great ceiling projections and could go overlooked: Justin Herbert and Ben Roethlisberger. I really love me some Herbert this week, as I detailed in this week's Heat Check podcast.

Running Back

Simulation Results

Observations:
With the shift to Taysom Hill at quarterback (much more on that later), Alvin Kamara's projections took a dive. He initially rated out a little better than Dalvin Cook outright. With the assumption Kamara loses both offensive efficiency and high-leverage rushing work to Hill, Cook clearly becomes the best play on the board at the high end.

Kalen Ballage -- yes, Ballage -- rates out with the second-best floor/ceiling ratio. Ballage has been the Los Angeles Chargers' most efficient rusher by a long shot so far this season on a limited sample, and the team has expressed interest in keeping him involved after an 18-carry, 6-target game last week.

A few names that jump out with good floor/ceiling ratings are James Conner, Derrick Henry, and James Robinson. I have little interest in any of them. Conner has shown a lack of a ceiling, and his team is throwing the ball frequently the past four weeks. Henry has a narrow path to a ceiling due to a lack of receiving, and his matchup with the Baltimore Ravens isn't exactly pristine. Robinson runs against a top-tier defense. But, the math likes them, so perhaps they're better plays than I realize, especially with Kamara's projections lowered.

Wide Receiver

Simulation Results

Observations:
We see a salary-based cluster at the top of the receiver pool with Davante Adams, Michael Thomas, Julio Jones, Keenan Allen, Adam Thielen, and Calvin Ridley. Thielen is the outlier there in a bad way. He's due for his touchdown numbers to fall off for a bit. His 25th-percentile outcome is just not on par with what you need there.

The best floor/ceiling play is Terry McLaurin, followed by a lower-salaried duo of Jerry Jeudy and Marquise Brown. Jeudy runs into a tough matchup with the Miami Dolphins. Brown has significant volume concerns in a run-heavy offense.

Keenan Allen and Mike Williams rate out well, though. Williams has, well, a lot to love with his downfield work and elevated role in the Chargers' offense.

Curtis Samuel's data is a little lower here in the sims, but he's part of the Williams tier for me, given his red zone role (eight red zone carries over the past seven games with at least one in all seven) and apparent usurping of a significant portion of D.J. Moore's targets. Tee Higgins and DeVante Parker rate out well in the $6,000 range, and the value seems right for Denzel Mims at just $5,400.

Tight End

Simulation Results

Observations:
This position is usually bad, and was especially bad this week until we got news that Taysom Hill would be starting at quarterback for the New Orleans Saints. Listed as a tight end on FanDuel, Hill pretty much breaks the position -- and the slate -- open. He has 20 carries over the past three games with an average of 44.7 rushing yards and 6 total red zone carries in that span. Even if the passing production is low and even if he splits some time with Jameis Winston (which doesn't sound likely), you're getting a rushing quarterback in an expanded role for just $4,500 at tight end.

And when you look at the rest of the tight end position, there aren't many reasons to be overly concerned with going hard at Hill. There's no Travis Kelce or George Kittle with realistic 20-point upside, so the opportunity cost is low. The best odds to hit 20.0 FanDuel points belong to Mark Andrews, which he did only 10.6% the time in the simulations.

If you don't want to go entirely at Hill, Hunter Henry and Austin Hooper are the second tier at the position behind Andrews.