Win Your Snake Draft

In this post, we will help you determine the best possible players to pick in a fantasy football snake draft through the FFA Webapp.  This is where you will find the most accurate fantasy football projections available (source accuracy evaluation article here). There are plenty of customizable features, allowing you to tailor a draft strategy to your league’s unique settings and own personal preferences. Now let’s dive deeper into how it truly operates, below is a more thorough description.

How it Works:

• To begin, a script is used to scrape the projections for each player from the provided sources. You can choose which projection sources to include, modify the weights, and choose what type of mean to calculate:
  • Mean: Straight average of the projections for a player with each source weighted evenly.
  • Weighted Average: Every projection from each source is given different levels of importance (weights), based on their historical accuracy or user input. The projections are then averaged according to the weights, producing a tailored result.
  • Robust average: A robust average takes into account potential outliers, minimizing the impact of extreme values and providing a more stable estimate when sources might vary widely.

• Second, based on the user’s scoring settings, we calculate players’ projections using an average of the analysts’ projections (by default, the sources are weighted according to historical accuracy).

• Third, we calculate the value of each player over a typical replacement player at the given position (VOR) and determine player rankings (article here for more on VOR).

• In addition to VOR, we calculate a player’s uncertainty level using two values: the range of projections and the standard deviation, weighted equally (50/50). The combined result is adjusted to fit a scale from 1 to 99. On this scale, a higher number means the player is less predictable (more uncertain), while a lower number means their performance is more consistent and reliable.

• For further analysis, the app provides a column for dropoff value. This column displays values for the “dropoff” in projected points from one player to the following two players at the same position. If a player has a high dropoff, you should consider targeting him because that position drops in value very quickly. Identifying dropoffs helps you avoid overpaying for a player when there are comparable options available later in the draft. It can also highlight positions where securing a top-tier player can significantly improve your team’s chances of success.

Strategy

• The uncertainty value can be a useful tool if understood correctly. It is generally best to draft starters with low uncertainty values, while maintaining high projected points. After those top tier players come off the board and bench players start to be drafted, the risk associated with each pick becomes manageable if the starting lineup is strong. Thus, identifying productive players with higher uncertainty values, potentially high ceilings, but low floors can pay off big time in later rounds.

• To better understand dropoff and VOR, let’s assume that Allen, Lamar, and Mahomes are projected as the top three fantasy QBs and to score the following points:
  • Allen: 300 points | ADP: 30
  • Lamar: 298 points | ADP: 42
  • Mahomes: 250 points | ADP: 47

• A decision purely based on rankings would mean drafting Allen. If we look at the projected points, Allen is only projected to score 2 more points than Lamar, who is projected to score 48 more points than Mahomes. Based on the small dropoff from Allen to Lamar, and the larger dropoff from Lamar to Mahomes in this scenario. There is more potential value to be gained by skipping Allen, drafting another position instead and still being able to draft Lamar a round later. Manipulating ADP with dropoff and VOR is an invaluable skill to a successful drafter.

To form a complete value-based draft strategy, the key takeaways should be the following. Use VOR to find the highest value players regardless of position, examine those players and the drop offs to the players immediately below them. Assess the draft board and roster construction to effectively use uncertainty, decide if you’re in the early or late portion of the draft and adjust your strategy accordingly. Use the uncertainty value to decide whether you should select a relatively safe option or try to hit on the high upside pick.

Snake Draft Settings

Sidebar Controls

• Season: Select the NFL season for projections
• Week: Select ‘Season’ for the full season projections

Main Settings

General

• Roster Size: Select the number of roster spots for each team
• Number of Teams: Select the number of teams in your league
• Positions: Select the position types available in your league
• Maximum Players: Maximum number of players displayed in projections

Number of Starters

• Select the number of starters for each position

Data Aggregation

• Impute Replacement Level Points for Missed Games: Check box to add points of the “typical replacement player” for each game a player is expected to miss
• Average Type: Choose between three projection calculation methods:
  • Weighted Average (Default): Uses analyst weights based on historical accuracy
  • Mean: Equal weights for all analysts
  • Robust Average: Reduces impact of outlier projections
• Analysts Selection
  • Choose which analysts to include in projections
  • Set custom weights for each analyst when using weighted average
  • FantasyPros and FantasyFootballNerd, by default, are excluded from the weighting to avoid double counting of sources. The user can weight them in the data aggregation settings if they would like to

Draft Settings

• ADP Source: Select the source of Average Draft Position for each player
• VOR Baseline: Value Over Replacement Baseline
  • This will determine the replacement-level players based on the projections position rank
  • The replacement-level players help us quantify how much better other players are compared to this baseline
  • Learn more about VOR

Cohen’s d Values

• These settings are used to determine whether two players fall in the same or a different tier
• Cohen’s d estimates the magnitude of difference between two player’s distributions of projected points
• Default threshold is set to 1.0
• Larger values will create fewer tiers and more players in the same tier
• Smaller values will create more tiers and fewer players in the same tier

Scoring Settings

• Configure point values for each statistical category

How to Use the Projections

Projected points and range

• Filter for position at the top
• The graph shows the projected points and range of projections for the selected position
• The larger the range, the more uncertainty in a player’s projections
• The colors indicate the tiers

Using the Projection Details table

Projection details

• Floor: 10th percentile projection (downside scenario)
• Points: Average projected points across selected analysts
• Ceiling: 90th percentile projection (upside scenario)
• SD: Standard deviation of projections
• Uncertainty: Variability and risk associated with player projections (1 = little uncertainty, 99 = large degree of uncertainty). Learn more about uncertainty
• VOR: Value Over Replacement. Learn more about VOR
• Drop Off: Decline in projected points for the next two players of the same position. Learn more about drop off
• Tier: Position-based performance tier
• ADP: Average draft position

Download Data