2023 WeeklyWinners

Reconstructed the entire 248,000 field and pulled learnings from the 2023 Weekly Winners contest.

User Portfolio Search Player Expected Value Full Data Download

UnderdogFantasy's weekly winners contest was first offered last year (2023). $3.5M prize fund with a max of 261,000 entries. (248,748 entries actually entered), with a 10.6% rake. For your $15 entry fee, you got a chance to win $20k every week. $205k was given out every week.

I reconstructed all quarter of a million lineups with their weeks 1-17 results.

Here's what I found:

TL;DR:

  • 3 QB Stacks are optimal (Both 2QB/3QB stacks were +EV)
  • If a RB/WR scored 6 points. 50% of the time that score would be used in the lineup's score. If a WR scored 10 points, that score would be used 90% of the time.
  • Stacking byes can be extremely effective. In week 13 (6 bye week), Only 65% of the field, was able to construct a full roster (every position scored non-zero)
  • Only 14 of the 32 stacks were +EV. Only 5 were extremely profitable (SF, MIA, HOU, DET, DAL)
  • Elite WR/TE are over valued. This is seemingly a remenant of WeeklyWinners ADPs being derived from BBM. (Typically in best ball, Elite QB/TE value is compounded over the course of the regular season)
  • Hero RB was the best strategy (this is probably skewed due to CMC starts)

Player EV / Methodology:

Positional Value:

In the above section, we looked at Player EV's. i.e how valuable was a specific player over the course of a season in the WW format. In this section we'll look at player scores by position instead to answer, What weekly point thresholds should we care about for our players? It is intuitive to the community that WeeklyWinners benefits players who's score spike more than normal BestBall formats. We'd prefer a WR who scores 30 points 1 week and 6 points the next vs. a WR who scores 18 in back to back weeks.

A standard Underdog Bestball lineup is made up of the following players:

  • 1 QB
  • 2 RB
  • 3 WR
  • 1 TE
  • 1 FLEX (RB/WR/TE)

The lineup having asymmetric constraints based on position in addition to positional values being different. We perform a similar analysis to the Player EV one.

  • For each player, we take their score and floor it. note their position, if the player was used and how much money that lineup made.
    • Example: DK Metcalf scored 18.4 points, his score was used in the lineup and the lineup cashed for $200
    • Data: {points: 18.0, position: WR, payout: 200, used: true}
  • Calculate the PnL and how often that score/position pair was used in the lineup:

PnL and Usage by QB Score



There are 2 important things to note from the chart above:

1. Breakeven point for each position. This is the score at which the average lineup with that player breaks even. For example, on average a QB will break even at 25 points. This means that in a given week, if you roster a QB who scores 25 points, you can expect to cash $0.88 (the cost of the lineup).

2. Percent used. You can see in the table to the right, QBs have the lowest percent used at their break even score (89.94%). This means that QBs are much more likely to double up at their break even score than other positions. 10.06% of the time, a lineup has 2 QBs above the break even score of 25 points.
- Note that: If an entry has a WR that scores 10 points, 90% of the time that score will be used in the entry. Even though that score has a expected value of -$0.68. It's even more wild to realize that if a WR/RB scores 6 points, that score will be used 50% of the time.

PositionBreak Even ScorePercent Used
QB2589.94%
RB1698.68%
WR1599.44%
TE1097.91%




PnL by QB ADP





PositionRound 1-5 BreakEven ScoreRound 6+ BreakEven Score
QB22.525
RB1616
WR14.515.2
TE1218

These charts are a function of positional point distribution, lineup constraints (how many of each position can go in a lineup), the cost of each player and conventional wisdom. The chart above shows the expected value of a player based on their ADP. The chart is broken down into 2 sections: Round 1-5 and Round 6+. The break even score is the score at which the average lineup with that player breaks even. For example, on average a QB drafted in rounds 1-5 will break even at 22.5 points. This means that in a given week, if you roster a QB who scores 22.5 points, you can expect to cash $0.88 (the cost of the lineup).

RB/WR have very similar break even points, regardless of where they are drafted. This is due to the quantity of WR/RB you need in your lineup. If a single late round WR scores 12 points, that score is not as valuable since its use in your lineup limits the number of high scoring WRs you can have in your lineup. (For a 12 point late round WR to be used, that would mean it is at least the 3rd highest scoring WR on your team, which makes it unlikely that your lineup cashed.)

Compare this to a onesie position (QB/TE). If a late round QB scores 22 points, that score is very valuable since it is likely your lineups top QB score of the week and close enough to the top scoring QBs to help you cash in addition to any positive correlation you have with your stack pieces and the additional value you got from punting an early position. To overcome these other factors, an early round QB would need to score 3 more points to break even.

Choose the number of stacks a lineup had.