NFL DFS Tips: 2023 Guide

Preview

5 out of 6 DFS players essentially wing it and lose, according to McKinsey and Company. More experienced players with sophisticated lineup selection techniques swallow them up. We built DFS Hub to keep you from being a minnow. So leverage these NFL DFS tips, and improve your odds of cashing out.

The Basics

  • Don’t rely on player stats alone when putting together your squads. Look carefully at team matchups. Other things equal, you wouldn’t want to pick an offensive player, for example, if:
    • The defense tends to shut down his position.
    • The player’s offense has below-average productivity overall .
      Fun fact: The average team scores a TD in almost 1 out of 4 drives (23%). Ceteris paribus, you want players on teams that score on more than 1 in 4 drives.
  • Strategy differs depending on the site you play on: DraftKings or FanDuel. Whereas a lineup might score 150 on DraftKings, that same lineup might only score 120 FanDuel.
    • DraftKings awards:
      • a full point per reception (PPR), 1/2 point more than FanDuel
      • 3 points more for 300-yards of passing
        • That gives a little extra value to elite QBs who consistently top 300
      • 3 points for 100-yards of rushing
      • 3 points for 100 yards of receiving.
    • The most obvious effect is that wide receivers and tight ends hold more value. 
  • There’s a good correlation between the volume of touches and the average number of fantasy points scored by a player. The correlation is even stronger when it comes to red zone touches.
    • Look for players who are getting an increasing share of red zone targets over the last 2 of 3 games.
    • But keep in mind that red zone targets are relatively scarce. Top-10 pass catchers only get about 23 a year on average, and only catch about 62% of them.
    • If you see an under-radar player who suddenly starts earning his team’s confidence and gets 2+ red zone looks in 2 of his last 3 games, he’s worth investigating.

Salary

  • Salaries are set by DraftKings and FanDuel every Sunday night and do not change for the entire week.
  • Salary algorithms heavily factor in two things:
    • recent performance
    • vegas odds
  • When a player’s salary drops significantly (e.g., 500+ points), they tend to be better values—assuming their role hasn’t diminished. That’s because contestants undervalue them by avoiding weak recent performance. If a salary drops 1,500+ points, they are among the best values of all–other things equal.
  • Smart money is better at finding value. That’s why (not surprisingly) fewer contestants overpay in higher stakes tournaments.

Finding Value in DFS

“Bet Prices not Teams.”

Legendary Las Vegas oddsmaker, Roxy Roxborough
  • DFS is not unlike sports betting. Bettors prefer to bet on favorites, but favorites do not cover enough to consistently make profits. In DFS, you must find value from less-loved players.
  • A good value is a player who has a strong chance at 15+ points, with a floor foreseeably over 3x salary.
  • Winning starts with:
    1. Checking the over-unders and player prop bets to identify the 3 biggest production offenses
    2. Identifying three productive stacks
    3. Identifying 2-3 low-owned values (contrarian plays) with big upside
  • Many pros start with finding value players first so they know how much salary they have to work with for chalk players.
  • Most DFS players are adept at spotting value. That’s a problem because the more of your competitors that choose the same player, the less valuable that player is. (See contrarianism below.)
  • One of the easiest ways to find value is with replacement players. Check out the bottom of our NFL DFS Injury Report for related strategies.
  • As the DFS market becomes more efficient, one tactic that’s still underutilized is comparing offensive and defensive lines. You then want to assess how imbalances in those matchups could impact quarterback play, rushing and receptions. The best comparisons to make in this case are:
    • [PFFPBLK] versus [PFFPRSH]
    • [PFFRBLK] versus [PFFRUND]

      Look for significant imbalances that may not be fully priced into salaries of players who could benefit from these mismatches.
  • One of the top 10 rules to consider when looking for value is avoiding recency bias. That means not getting hung up on a string of bad performances, so long as a player has been getting opportunities and his role has not diminished.
    • A player who’s had eight targets in the last few games but only caught two a game, may just be unlucky. Other things equal, eventually he’s going to significantly exceed his projections.
    • Yet, the average DFS player isn’t thinking like this. They’re focusing instead of the players who are “on a roll.” That creates lower salaries for recent duds, and more potential value.

Vegas Lines

  • Vegas oddsmakers predictions are based on the best data available. Their livelihoods depend on the accuracy of their predictions. If they post lopsided lines, the sharks will eat them alive. It’s not surprising then that studies confirm that no team consistently outperforms or underperforms the spread. In other words, all teams beat the spread about 50% of the time, which is basically random chance.
    • Side note: Sportsbooks make their lines based on a median outcome—i.e., half the time the result will be over, and half the time it will be under, excluding ties. Most of DFS Hub’s average-based stats are also based on median performance. Compared to mean-based stats, ours:
      • better reflect expected player performance
      • avoid the misleading skew of mean average stats, and
      • better correspond to Vegas props.
  • Nevada sportsbooks win 6% to 8.5% a year, says Jimmy Vaccaro, and they never lose. Use their number-crunching to your advantage.
  • Vegas lines and totals are factored heavily into DraftKings and FanDuel pricing. So don’t expect to uncover tremendous hidden value analyzing this data.
  • As accurate as Vegas lines are (favorites win about 67% of the time), they’re not perfect representation of what should happen on the field. Setting the line is less about team/player analytics and more about attracting even two-way action from bettors.
    • Public perception is a huge factor in line setting, whether that perception is accurate or not.
    • Essentially, the pointspread represents oddsmakers’ prediction on what the betting public thinks is going to happen. Most of the time, this is not significantly different from what actually happens.
    • It is mistaken perception that creates inefficiencies in player pricing and ownership, and finding these inefficiencies is what DFS is all about.
  • That said, while the line is key, the most important Vegas data to consider is the over/under.
    • Fact: Bettors like shootouts. They choose the “over” instead of “under” by a 65%-35% margin.
    • Note: Games are getting more offensive. In 2020 the average total per game was 49.6 points, versus 45.4 five years prior (source). There’s a reason fantasy players target over/unders ( [OverUnder] ) of 50+. This trend is making ceiling even more important than floor in GPPs.
  • Ignore the moneyline. It won’t tell you anything that you can’t learn from the point spread.
  • If a game is projected to be among that week’s highest scoring, concentrate on players from that game. Just keep in mind that DFS Salaries will reflect that.
  • Player prop bets can be useful. When Vegas projects the Saints at 29 points and receiver Michael Thomas has an over/ under of 7.5 catches, 93.5 yards, and 1.5 touchdowns, pay attention.
  • There is a correlation between line movement and expected fantasy points. If a team total moves from 23.5 to 25 on a Sunday morning, you may get some value from players on that team.
  • Alternatively, if you’re debating between two defenses and one of their opponents team total moves down a point, that’s a good tie breaker.
  • Favorites by 4+ points win significantly more than favourites by 1 to 3.5 points. Teams favoured by a TD win an large percentage of time.
  • Roughly 35% of NFL games result in an upset. Upsets occur less often in the postseason and most often in September.
    • Upsets may occur more easily these days thanks to things like increased penalty calling and going for it on 4th down, which can favor underdogs (see this NYT article).

Contrarianism and Ownership

  • In sports betting, there is a clear relationship between the percentage of bets on a team and that team’s recent success. It’s well-known bias. Player ownership (popularity) works the same way.
  • Ownership is positively correlated to:
    • Recent player performance
    • Vegas implied totals
    • Perceived value
  • Following the masses and choosing high-value players with great recent performance, in a game with a high total, is a recipe for mediocrity—particularly in GPPs.
  • DFS is not just about picking the most productive players, it’s about picking the most productive players with low ownership. If you nail a player who’s little little owned, you gain that much more on your competitors, especially in GPP contests.
  • Ownership ranges from “low” (less than 7%) to moderate (7% to 15%) to “high” (over 15%). The latter are considered “chalk” plays.
  • The more competing players there are, the more it pays to be contrarian (if your goal is to win a tourney).
  • If projections have narrowed things down to two players and all else is near equal, the lower-owned player could be the difference between you winning a GPP contest or just cashing.
  • Don’t get too crazy with long-shot picks (3% ownership or less). Your chances of hitting on more than one little-owned player are slim.
  • Taking too many risks is a guaranteed way to lose, especially in cash games. Don’t avoid every obvious value just to prove you’re a contrarian. Pick some faves too. Most of DraftKings Millionaire winners have at least a few players in their lineup that were highly owned.
  • Don’t overfocus on ownership. Success requires finding a balance between ownership and value. Moderate ownership and above-average value beats low ownership and average value.
  • Contrarianism has its limits. If you’re building a lot of lineups, you still want high-owned chalk in some lineups so that you don’t miss out on blockbuster performances. But you want to be underexposed overall to players who are heavily owned. That way, if the player flops you gain an edge on the masses.

Field Surface and Weather

  • Playing surface has a statistically insignificant impact on player performance, according to research by Amundson et al. (2006). i.e., NFL players do not have trouble transitioning between field surfaces.
  • Borghesi (2007) found that teams that usually play in warmer weather struggle on the road against cold-weather opponents. As a result, and other things equal, betting on home teams in cold weather pays off when they’re facing a away team from the south.

Home Field Advantage

  • Home field advantage traditionally accounts for up to 3 points of a Vegas pointspread.
    • Not coincidentally, out of all win differentials, home teams most often win by three points.
  • But on average, teams that play on home turf win by just over two points more than if they play away.
  • Home field advantage, as measured by Vegas point spreads, has been falling since 2006, dropping from 3.3 points in 2006 to just 1.3 points in 2020 (when teams played in front of empty stadiums due to the pandemic). A good rule of thumb is about a 2.0 to 2.5 point edge for teams that play at home, other things equal.
  • Data shows that home field edge is more pronounced in the first half of the season.
  • Although a home-underdog bias used to be profitable, studies through 2011 suggest wagering on home underdogs is no longer profitable.
  • Teams known for their crowd noise (e.g., Seatttle and KC) have a stronger advantage at home. Here are the top 10 loudest NFL stadiums at the time of this writing.

Bankroll Management

  • Limit your daily spend to 20% of your entire bankroll. If you deposit $100, don’t risk more than $20 in entry fees each day.
  • If you’re a serious player and have time to do the research, play multiple lineups instead of betting more on a single lineup. Every Sunday, many of the best DFS players in the world (like maxdalury) diversify in this manner. They often creates hundreds of lineups, typically around a core group of high-potential players.

Reading the News

  • Of all the NFL DFS tips we can give you, this one is perhaps most important. Read the news!
  • We’re talking news on players and news on matchups. Without third party insights, DFS is just a gamble. New is how you find up-and-coming players, players whose role is changing, performance in practices, defenses who are exploitable, and so on. DFS Hub is unparalleled in this department because we deliver more news sources than any other DFS site, at least any that we’re aware of. And more importantly, we deliver more local sources. Local reporters are uniquely positioned to provide insights into reps in practices, performances in training camp and commentary from coaches.

Correlations (Stacking)

  • Correlation measures how much two things are related. A correlation over 40% is an indication that one thing is reasonably linked to another. A strong correlation is around 70%+.
  • Correlation doesn’t mean that one thing causes another, however. There are all too many spurious correlations in football. Rushing yards and wins are one example. They are correlated, but more rushing does not “cause” wins. Teams tend to rush more because they are already winning.
  • In DFS, “correlation” can tell you how one player’s performance moves with another’s.
  • In big tournaments, you want to roster positively correlated players. Stacking is not required in cash games, however. In cash games you mainly want the highest possible volume at the best possible price, regardless of the team a player is on.
  • Maximizing correlation means that if one player does well, other players in your lineup do too. You take advantage of this phenomenon by stacking.
    • By contrast, you would not want to stack a quarterback with the opposing defense, for example.
  • Stacking players from the same team has a related benefit. It’s easier to be right about one game matchup that multiple matchups.
  • Roughly 90% of perfect lineups have at least one stack, and as game totals trend higher, stacking matters even more.
  • Historically, the highest NFL DFS correlations have typically been:
    1. QB – WR1
      • The problem is, this is usually the highest owned stack. Owning the chalk QB – WR1 stack won’t give you much edge in a big tournament.
    2. QB – TE1
      • A far less owned stack than QB – WR1, unless we’re talking a cream of the crop duo, like Mahomes – Kelce
    3. QB – WR2
    4. QB – WR3
    5. QB – Opposing WR1
      • This one does especially well when the QB has a big day
      • If you’re expecting a blockbuster day from your QB, always ask if there’s an opposing receiver who could benefit while playing catch-up.
      • Taking a QB, one of his pass catchers and a WR from the opponent is a popular “game stacking” strategy.
      • Game stacks can be especially beneficial in GPPs if one of the receivers is inexpensive. That way, if a game is a shootout like you predicted (and why would you game stack it otherwise), you get more leverage from him being a less-owned player.
    6. RB1 – DST
      • Naturally this correlation does better when the RB is a proficient pass catcher. Otherwise, don’t go out of your way for this stack.
      • Sometimes when you can’t afford a stud RB1, taking his defense is a way to get exposure to him, indirectly.
    7. QB – RB1
      • Again, if you go this route, lean towards rushers who get multiple targets per game
  • We should point out that a large minority of perfect lineups last year had a RB stack. With offensive production trending up, if the passing game does well, so do the best DFS running backs.
  • Historically, 2-man stacks usually do best when a team is favored by 1.5 or more points and expected to score 24+ points (source: 4FOR4.com)
  • The strongest negatively correlated stack is QB – opposing DST
    • For a contrarian play in GPPs, and when appropriate, consider fielding a defense that’s facing a heavily owned QB
  • Game stacks, pairing a player on one team with a player on the other, can be effective (and somewhat undervalued) in big tournaments, particularly in games with high totals. An example would be QB – WR1 – Opposing WR1. In GPPs, game stacks raise the overall correlation of a lineup, the uniqueness of your lineup, and hence, the upside. Most perfect lineups have game stacks.
    • Our research suggests that the sweet spot—in terms of GPP edge and probability of winning—appears to be 3 + 1 games stacks — those where you choose 3 players on team A and one player on team B. The best expected returns for Millionaire Maker contests come from 4 + 2 lineups.
    • Pro Tip: It may be counter-intuitive but game stacks in lopsided games (e.g., 10+ point spreads) tend to underperform more often than not. (Source) Large spreads imply struggling underdogs.
  • Don’t get too crazy with same-team stacks like QB – WR1 – WR2. Unless it’s going to be an epic shootout and both receivers are top notch, the chances of hitting (with each player scoring 25 fantasy points or more) are less than 1 in 10.
    • Pro Tip: Avoid picking 3-man stacks without the QB-WR1. The odds of success are generally too poor.
  • Stacking works best on favorites in high-total games that are expected to have tight scoring. Stacking underdogs or players in lower-total games (less than 48 points) entails greater risk.

Do You Need to Use an Optimizer?

  • Unless you’re a full-time pro, certainly not. There are too many reasons not to use an NFL DFS optimizer.
  • Among the core NFL DFS tips, this is a big one that newbies ignore: Optimizers are of little value if you’re making a limited number of lineups. Professionals rely on foresight, quality projections and skill in exploiting correlations and leverage. You can do the same without an optimizer.
  • NFL coaches know all the data, and they know the opposition knows all the data. If they’re going to win they need to mix things up and run plays, and create player roles, that opponents may not expect. No optimizer can factor in these sorts of things based on past data alone. Offensive and defensive schemes change. They adapt to opponents and they change due to injuries and performance. Winning in DFS requires making manual adjustments to your expectations of a player’s performance, based on the circumstances surrounding a game that week.

Changing Roles

  • Be aware of players whose roles may have been downgraded. That can seriously reduce the player’s opportunities.
  • This happens when a team sours on a player. Clues that this has happened include:
    • A team signing other players of the same position.
    • A head coach praising other players of the same position, but not mentioning the player in question.

Other NFL DFS Tips

  • While stacking is a productive use of correlation, you must avoid drafting negatively correlated players at all cost. If you’ve drafted Patrick Mahomes and he’s playing Denver, for example, you wouldn’t want to draft Denver’s defense or running back. Reason being, if Mahomes has a big day, most likely the opposing defense and RB will not.
  • Good players who have bad performances often outperform in their next game if it’s against a weak opponent. Understanding positional matchups will be to your benefit.
  • Momentum can be a factor, especially early in the season. For example, the NFL says that “teams that start 3-0 can expect to make the postseason 76% of the time. The odds drop with each loss; a team that starts 2-1 sees its chances drop to 55%, a 1-2 team drops to 32%.”
Source: NFL Football Operations
  • Players who have big games are likely to see a scoring regression in their next game, especially receivers. Defenses (DSTs) are an exception to this rule.
  • Despite Tom Brady playing into his mid-40s, age is a factor in the NFL. Once running backs reach 28, for example, production generally tails off noticeably. The number is 29 for tight ends, 30 for wide receivers and 32 for quarterbacks, according to research from Doug Driven (former editor of Pro Football Reference). Players like Brady and Travis Kelce are outliers, not the rule.
  • How many fantasy points per game should you expect from an elite player over the course of a season? Well probably not more than 30.1. That was the highest fantasy points per game average over an entire year ever. And it’s a record set in 2006 and owned by running back LaDainian Tomlinson.
  • Don’t pay much attention to what coaches say about player roles. Instead, watch what they do with respect to player reps in practice, depth chart moves, etc.
  • Make sure to be aware of injuries and injury tags. For example, wide receivers who appear on injury reports produce materially less fantasy points. To go a step further WRs with concussions perform particularly badly, getting 20% fewer targets according to research cited by Fantasy Football Consultants. 
  • Don’t get too hung up on points and points again. Yards per play, adjusted for past opponent strength, are more predictive of future points.
  • Sample size has always been a problem when analyzing the NFL. “There aren’t enough games,” Roxy Roxborough says. “So trend analysis, which was one thing I was big on, is pointless.”
  • Points are not awarded for punt return yards, only punt return TDs. The DST is awarded six points for a punt return TD. If you draft a WR or RB and he scores a TD on a punt return, he gets no points for it.
  • Don’t get too blinded by depth chart rankings. “…Though Kenyan Drake was listed ahead of Chase Edmonds on Arizona’s depth chart, Edmonds had been getting double digit “opportunities” (either rushes or targets in the passing game) pretty much all season,” said David Bergman, who chose Edmonds and won $2.5 million at the December 20, 2020 Fantasy Football World Championship (source: ESPN).
  • Playoffs are a different animal. In the postseason, favorites tend to outperform underdogs by a bigger margin compared to the regular season. 
  • It’s dangerous to read too much into prior-season defensive rankings. This is a common mistake players make in the early weeks of a new season, but correlations between fantasy points allowed in one year and fantasy points allowed in the next year are low. The greatest season-to-season carry-through is with how defenses defend quarterbacks. According to 4for4 research, there was a 0.26 correlation between points allowed to quarterbacks in one year, and that same defense’s points allowed to quarterbacks in the following season.