Editor's Note: This piece initially covered only the 2019 NFL season but has since been updated to include the 2020 and 2021 seasons.

I won't lie to anyone here. When FanDuel launched single-game daily fantasy NFL contests, I didn't really know what I thought. I liked it, of course. I always like more fantasy sports to play.

But I wasn't really sure what to expect. I had never examined a single NFL game to such depths before. Sure, I've pored over matchups in painstaking detail, but I hadn't had reason to treat a single game as its own, self-contained fantasy entity.

I quickly learned that I high-key love single-game DFS because it requires you to think about the fantasy football a little differently than what I have learned over the past decade-plus. It's usually pretty easy by Friday in a week to figure out which running back has the best chance at a big game out of 20-plus teams. But in single-game daily fantasy, you really have to dig in and think about the way a game can flow and how it's all connected.

It's fun. Really fun. But I still wanted to find a way to dive into the numbers and see the best ways to play single-game daily fantasy football on FanDuel. So I gathered a whole boatload of data, and now I'm here.

Organization proved to be a bit hard because there are so many things to cover, so I'm just going to start, go over various questions I have about the format, and then wrap it all up at the end.

What Is Single-Game NFL?

It's just what it sounds like: daily fantasy centered on a single NFL game. You have to roster at least one player from each side, choosing between quarterbacks, running backs, wide receivers, tight ends, and kickers. One of your players is deemed your MVP, and his FanDuel points are multiplied by 1.5.

You still have to adhere to a $60,000 salary cap.

What's the Sample?

I have information and optimal lineups from more than 600 single-game offerings since the 2019 NFL season through the 2021 season.

To clarify, "optimal lineup" refers to the best possible lineup we could've built within the salary cap. I'll also call this the "perfect lineup" throughout.

We're usually dealing with small samples in examining the NFL, but by expanding this study, we have three seasons' worth of single-game slates to examine.

There are kind of unlimited questions surrounding single-game strategy, so I'm just going to hop right in from what's probably most basic and then expand from there.

Should You Use All the Salary Cap?

Leaving salary on the table (i.e. not spending all $60,000 of your salary cap) always has some merit in traditional daily fantasy contests because not a lot of people do it to a relevant degree. It's one thing to spend $59,500 of the cap, but that's not really wasting salary. It's another to spend $52,000.

But in a single-game format, there's gotta be some leverage by eating some cash, right? We're picking five players from the same game, so it'd help to think differently, yeah?

Let's find out. Here's the distribution of salary from the perfect lineups in the sample.

It looks like there is some leverage. Occasionally it's a good thing to save a significant chunk of salary while seeking the perfect lineup.

A solid number of perfect lineups (17.4%) used $55,000 or less of the salary cap, and only (10.6%) used every dollar possible. That being said, 74.4% of these perfect lineups utilized at least $56,500 of the salary.

Be okay being different, but make sure there's a reason for it.

Are 3-2 Lineups or 4-1 Lineups More Common?

You have to draft five players, and you have to use at least one player from each team. That means your lineup combinations are binary: three from one team and two from the opponent or four from one team and one from the opponent.

You'll often hear the term "onslaught" for the 4-1 lineup, something that (anecdotally at least), makes sense in a blowout.

Was that a more common construction over these optimal lineups?

Nope. Of course, 4-1 lineups are still relatively prevalent all things considered, but the data shows that a balanced approach is more common in the optimal.

I think this qualifies as a takeaway: lean toward a 3-2 lineup more often than not. For four players to outscore opponents, we need a very specific game flow to occur, and even then, we have to adhere to a salary cap.

What Does the Optimal Lineup Distribution Usually Look Like?

I'm not going to lie, this was probably the thing I was most excited to study, but it was quite a disappointment.

What I wanted to find here is the answer to the question: "How many quarterbacks, running backs, receivers, tight ends, and kickers is best to roster?"

Well, among these lineups, there isn't a whole lot of consensus. I didn't want to split this section into winning/losing players because then we're getting too granular and nitpicking between a QB-RB-WR-WR-TE lineup based on which team each of the five played on that game.

I straight up just wanted to see how frequently a certain distribution of positions occurred.

Even while keeping this more general and bucketed based solely on position, the most common lineup combination appeared just 8.9% of the time: QB-RB-RB-WR-WR (not adjusting for MVP selection).

I want to pause here, at least, and let it sink in that the most common perfect lineup composition actually had two running backs, which I never would've expected in a single-game format because we usually see opposing running backs have quite a mild or even negative correlation with one another.

Here are the combinations that occurred in at least 3.0% of the lineups.

There's no clearcut answer of which positions to roster.

I think we just need to build the best lineups we can without getting too cute and also not shy away from a double-running back lineup when the game permits.

What Should We Do With That MVP Slot?

FanDuel single-game NFL lineups do not require you to roster any specific positions, but one of your five players must be designated as your MVP at no extra salary.

Usually, that'll be a quarterback, as nearly half of all single-game lineups last year used a quarterback at MVP.

Makes sense. Quarterbacks are usually the highest-salaried players in a single-game contest, and they can easily rack up a high floor of fantasy points.

It also tracks that running backs are next on the list in the half-PPR format and that receivers are the only other realistic options for an MVP choice when players are constructing lineups.

But what do the optimal lineups have to say? Here's the breakdown of MVPs by position in the 600-plus optimal lineups.

There is leverage in rostering running backs and receivers at MVP: they're more likely to be in the optimal lineup than they are drafted as such.

Still, along with this, quarterbacks have nearly a 45% chance to finish as the MVP, so it's not like we should fade them outright as the MVP selection. It's more that we should be open to drafting running backs and receivers as the MVP more than the overall sample shows that we actually do.

How valuable is it to get weird with your MVP by rostering someone off the radar? When it hits, it'll be great, but the average salary of the optimal MVPs was $13,664 last season.

Only 18.9% of the optimal MVPs had a salary of $11,000 or lower, and 13.2% of the MVPs had a salary below $10,000. The long-term winning approach isn't really to plug in a random play at MVP, though there are definitely instances during which a low-salaried MVP can pay dividends.

I'm going to stick with the MVP conversation for now and look at correlations and trends when each of the three primary positions are the MVP. Logic would suggest that when a receiver is the MVP, it's a different type of game flow than it is when a running back is the MVP.

Does that hold water?

How Do the Positions Correlate According to the MVP Position?

We know enough by now to have some thoughts forming in our minds. Running backs and quarterbacks are the most common MVP choices.

But we can look at what the average lineup looks like when a given position is the MVP. That's what this table shows.

The correct way to read this table is as such: on average, when a quarterback is the MVP of the perfect lineup, 0.0 quarterback teammates are in it, 0.55 running back teammates are in it, 1.0 receiver teammates are in it, and 0.43 quarterback opponents are in it.

Again, these are averages, so they really just indicate trends and interconnectedness.

When a Quarterback Is the MVP
Okay, so, the second strongest pair on the whole table is a quarterback MVP with his receivers. A quarterback can have multiple pass-catchers in the perfect lineup with him, and when a quarterback scores enough to be the MVP, it's likely going to happen when his receivers fare well.

On average, when we have a quarterback as the MVP, we get more than 1.0 receivers of his making the optimal lineup.

An MVP quarterback's team averaged 29.7 points and won 75.9% of the time.

When a Running Back Is the MVP
When we have a running back at the MVP spot, the most frequent pair is actually a kicker (1.14). Optimals with running backs at MVP and at least one kicker had an average over/under of 45.7 points. Lineups with a running back MVP but no kickers had an average over/under of 46.1.

The actual points scored were 41.7 and 47.4, respectively, so that tracks: kickers are more viable in lower-scoring games.

In the sample, 59.0% of the time when a running back is the MVP, his quarterback is also in the perfect lineup. That gives credence to an offensive stack but with the back as the MVP rather than the quarterback.

An MVP running back's team won 77.0% of the time and was favored, on average, by 2.0 points. The MVP's team averaged 27.1 points.

When a Wide Receiver Is the MVP
Receiver MVPs come with opposing receivers in the perfect lineup quite often: there are an average of 0.75 opposing receivers in the perfect lineup when a receiver is the MVP choice, and 85.3% of receiver-led perfect lineups had at least one opposing receiver in it as well.

Naturally, we see quarterbacks in the perfect lineup frequently when a receiver is the MVP -- but not always: 77.8% of receiver-MVP lineups had that player's quarterback in it as well.

While it's pretty common to see a receiver make the perfect lineup when his running back is the MVP, it's far less common to get a running back picked as a flex when it's the receiver who had an MVP-worthy game.

In the sample of lineups with a receiver at MVP, only 32.5% also had a running back from the same team in that lineup.

Conversely, 52.8% of perfect lineups with a running back at MVP had a receiver from the same team also make it.

Basically, consider fading the running back of your wide receiver MVP choice when you do make that kind of lineup.

Do You Have to Stack and If So, How Much?

Stacking, or pairing teammates together, is a common daily fantasy philosophy.

The easiest stack to make is a quarterback with one of his receivers. If a quarterback throws a touchdown pass to a receiver and you drafted both, then you have 10 FanDuel points from the touchdown alone (and then 0.5 for the catch and whatever boost for the yardage).

For a quarterback to put up a huge game, at least one of his pass-catchers has to have a touchdown (or two), unless the quarterback ran for all the scores. That's why stacking in a single-game situation makes sense in theory.

Here's the frequency of stacks (quarterback plus pass-catcher [receiver or tight end] only) in these optimal lineups, split by winning and losing teams.

This shows the percentage of perfect lineups with a quarterback and pass-catcher stack of any amount of pass-catchers followed by specific number of pass-catchers.

So it's evident that stacking is common in optimal lineups.

Of these 600-plus lineups, 87.3% featured a stack from either the winning or losing team.

Should we be pairing both quarterbacks with a pass-catcher to vie for the two-stack lineup? Eh.

Only 18.8% of the optimal lineups had a quarterback-receiver/tight end stack from both teams. It's more likely that a quarterback and two of his pass-catchers will make the optimal than it is that both quarterbacks and a pass-catcher from each make the optimal.

Stacking both quarterbacks together makes a lot of sense, but the primary issue with that is salary, as quarterbacks are almost always the highest-salaried picks in a game. A two-quarterback optimal happened 34.3% of the time.

Stack often, but don't overstack.

What's the Deal With Kickers?

Well, kickers have been an MVP in 1.2% of the optimals, so it's not a great play. However, they're low-salaried fillers who do have a path to a floor.

In fact, there are actually nearly as many kickers in the sample as there are tight ends if we look at all individual players across the lineups.

Only 8.1% of the players in the optimal lineups were kickers, and again, very few were MVPs. It makes sense, as there isn't significant upside in a kicker, who can score only through attrition. Tight ends may actually be better punt plays.

That being said, 37.2% of the lineups featured at least one kicker, compared to 42.0% that had a tight end. Only 3.5% featured double kickers.

When there was a kicker in the lineup, the average over/under was 46.6 points but the average point total was just 42.8.

By contrast, lineups without any kickers had an average point total of 49.7 despite an average over/under of 47.2.

Games that we think will hit the under should be when we target kickers as lower-salaried differentiation plays.

Overall and MVP Draft Percentages

I don't know if anything presented thus far has been groundbreaking, but I do think that -- if nothing else -- we're about to get to a big, important takeaway.

So far, we've figured out that running backs and quarterbacks are the most common MVP picks and that you don't roster three tight ends and a kicker to get a huge leg up or anything wacky like that.

But we can easily get a boost by being willing to think differently just in how we construct our lineups.

Here are the draft percentage rates for the MVPs in the perfect lineups -- broken down in buckets of popularity.

This is the good stuff.

Roughly 40.0% of MVPs from the perfect lineups were drafted as the MVP in fewer than 10.0% of all lineups. Basically, unpopular MVP selections are frequently the optimal lineup's MVP.

That being said, 27.9% of the optimal MVP choices were very popular as MVPs (rostered on at least 40.1% of all lineups).

We should look for situations where there are natural MVP pivots (i.e. targeting receivers and running backs if everyone is in on the quarterback for a given game).

Summary and Important Notes

There's a lot to try to remember, so I wanted to throw some of the biggest takeaways into a tidy list and also throw in some random things I came across that didn't make the cut above.

- 74.4% of the perfect lineups used at least $56,500 of the salary cap.
- 17.4% used $55,000 or less of the salary cap.
- 67.4% of the perfect lineups featured a 3-2 lineup split, compared to 32.6% for a 4-1 lineup.
- The most common lineup combination (disregarding MVP choice) was QB-RB-RB-WR-WR, but that occurred in only 8.1% of the lineups.
- 54.9% of all lineups rostered a quarterback at the MVP spot, but quarterbacks were MVPs in only 44.5% of actual perfect lineups.
- Running backs are more likely to be the optimal MVP (29.3%) than they are to be rostered as such (23.3%).
- The same goes for receivers with a 20.8% optimal MVP rate and a 17.5% MVP roster rate.
- We really shouldn't consider tight ends for MVP excluding the rare outlier -- and especially not kickers.
- 87.3% of the perfect lineups had a quarterback-receiver/tight end stack of any type, but it's rare for there to be three pass-catchers from the same team in the perfect lineup.
- Only 34.3% of the perfect lineups had two quarterbacks (i.e. one from each team).
- 52.7% of MVPs came from the home team.
- 74.6% of MVPs came from the winning team.
- 60.6% of MVPs came from the favored team entering the game.
- 37.2% of the perfect lineups featured at least one kicker.