« back to all articles
« back to all articles

Introducing: Value Over Stream (VOS)

Andrew Luck Quarterback Colts

A new metric is now available on PlayerProfiler: Value Over Stream (VOS). VOS is a helpful guidepost to be utilized during fantasy football drafts. The metric calculates each player’s fantasy points per game above the average waiver wire replacement in standard fantasy league formats during either the current or previous season, depending on the time of year. VOS also takes positional bust rates into account.

Now, let’s break down precisely how the metric is derived and how it may impact fantasy draft strategy.

Replacements In The Stream

First, what is an average waiver wire replacement? For QBs, the “stream QB” is represented by the 19th best quarterback (fantasy points per game) in 2014, because quarterback No. 19 most likely represents the best available replacement on the waiver wire.

This assumption was determined after some crude market research, after which we concluded that approximately 50-percent of fantasy gamers carry 2 QBs in standard single-quarterback leagues at any point during a season, which equates to 1.5 QBs per team. The most common number of teams in a league is 12, and 12 multiplied by 1.5 is 18. Therefore, in an average standard fantasy league, the 19th best quarterback at any given time is likely to be available for streaming via the waiver wire.

In fact, the 19th highest ranked quarterback in fantasy points per game in 2014 was Mark Sanchez, who was the quintessential option that many fantasy gamers, who were otherwise weak at the quarterback position, may recall turning to on the waiver wire last season. Mark Sanchez is the stream, and the stream is Mark Sanchez.

For RBs, the “stream RB” is represented by the 49th best running back (fantasy points per game) in 2014. This player represents the likely best available option on the waiver wire. We assume this because our crude market research suggests that the most common league format includes two dedicated running back slots and one flex. If we also assume that the fantasy gamers use a running back at flex 50-percent of the the time, then the average standard fantasy league team will have 2.5 active running backs. Our research also found that most fantasy leagues have approximately five bench spots, which are most often dedicated to the following positional mix:

  • QB: .5
  • RB: 1.5
  • WR: 1.5
  • TE: .5
  • DEF: .5
  • K: .5

Given this, most fantasy teams carry four running backs. In a twelve-team league, that equates to approximately forty-eight rostered running backs throughout the season — a few more at the beginning of the season, and a few less during the height of the NFL bye weeks when more team roster a second QB/TE/K/DEF. Therefore, in an average standard fantasy league, the 49th best running back at any given time is likely to be available on the waiver wire.

The “stream WR” who represented the prototypical wide receiver waiver wire pickup in 2014 was also the 49th best wide receiver (fantasy points per game). In a twelve-team league, just like with running backs, about 48 wide receivers are rostered. A handful more WRs are typically rostered early in the season, and a few less are rostered during the peak of the NFL bye weeks as fantasy teams are often forced to add more quarterbacks, tight ends, kickers, and defenses. Therefore, in an average standard fantasy league, the WR49 is likely to be available on the waiver wire at any given moment during the NFL regular season.

For tight ends, the “stream TE” is represented by the 19th best tight end (fantasy points per game) in 2014. Just as with quarterbacks, we concluded that about 50% of fantasy gamers carry 2 TEs in single-tight end leagues, which equates to 1.5 TEs per team. Therefore, in an average, standard fantasy league, the 19th best tight end at any given time is likely to be available for streaming.

Bust Rates

After calculating each player’s value over the likely waiver wire replacement player, we took the additional step of including a bust rate coefficient. Why do bust rates matter? In his groundbreaking piece, Zero RB, Antifragility, and the Myth of Value-based Drafting, Shawn Siegele explained that the inherent fragility and subsequent volatility of the running back position leads to less accurate RB fantasy output forecasting. Given this additional uncertainty, a discount must be built into any value-based player valuation schema. The bust rate coefficient serve as a throttling component of the VOS metric, which discounts for prediction uncertainty.

Bust rate by its very nature is an arbitrary device with no universal definition. Therefore, it necessitates a comprehensive, contextual explanation. We determined bust rates for the Value Over Stream (VOS) metric using the following approach.

Games Played

We first discount a player’s fantasy points per game based on his number of games played. To accomplish this, we multiply the players fantasy points per game by the percentage of games missed in 2014.

Cost-discounted Bust Threshold

We narrow our focus to only players drafted in the first twelve rounds when players drafted with the intention of being fantasy starters are typically selected. We then divide players into cohorts (QB1, RB1, RB2, WR1, WR2, TE1) and average their cohort’s fantasy points per game. For instance, the QB1 (fantasy QB No. 1 through QB No. 12) cohort’s average fantasy points per game in 2014 was 19.6.

To create a cost-discounted bust indicator, we take each quarterback who was drafted it the first 12 rounds of 2014, and multiply his fantasy points per game by a “draft capital premium,” and then subtract that number from 19.6. The round premiums for quarterbacks are as follows:

  • Round 1 draft capital premium: .85
  • Round 2: .80
  • Round 3: .75
  • Round 4: .70
  • Round 5: .70
  • Round 6: .70
  • Round 7: .65
  • Round 8: .65
  • Round 9: .65
  • Round 10: .60
  • Round 11: .60
  • Round 12: .60

Any negative values indicate busts. Yet, despite only averaging 16.3 fantasy points per game in 2014, because Andy Dalton played all 16 games and was not drafted until in the 10th round (on average), he met the baseline threshold required to not be classified as a bust. Because most fantasy teams start only one quarterback and a single tight end, the tight end draft capital premium allocation and subsequent bust rate threshold determination uses precisely the same logic as the quarterback calculation outlined above.

For running backs, we categorize players as busts by first accounting for the number of games played, then setting a cost-discounted bust threshold by first dividing running back drafted in rounds 1 through 12 into three groups:

  • Cohort 1: Running backs drafted with the intention of possibly achieving RB1 status (RB1’s averaged 17.5 fantasy points per game in 2014).
  • Cohort 2: A blend of running backs drafted with the intention of being RB2’s while possibly achieving RB1 status (RB1’s and RB2’s together averaged 14.3 fantasy points per game in 2014).
  • Cohort 3: Running backs drafted with the intention of achieving RB2 status (RB2’s averaged 11.7 fantasy points per game in 2014).

Given that most fantasy teams draft a running back within the first four rounds, Cohort 1 is comprised exclusively of running backs drafted in rounds 1 through 4, Cohort 2 is a blend of RB1 and RB2 average fantasy points per game, and Cohort 3 consists of running backs drafted in rounds 9 through 12 with the likely intention of filling the role of starting RB2 on a fantasy team. By separating bust rates into tiers, we ensure that a second round running back who produces like a low-end RB2 will be considered a bust, even if that running back is still measurably better than the best available waiver wire running back.

Cohort 1 (RB1 – 17.5 ppg average)

  • Round 1 draft capital premium: .80
  • Round 2: .75
  • Round 3: .70
  • Round 4: .65

Cohort 2 (RB1/2 – 14.3 ppg average)

  • Round 5: .80
  • Round 6: .75
  • Round 7: 70
  • Round 8: .65

Cohort 3 (RB2 – 11.7 ppg average)

  • Round 9: .80
  • Round 10: .75
  • Round 11: .70
  • Round 12: .65

Because most fantasy teams also start multiple wide receivers, the WR cohort and related draft capital premium allocation, as well as bust rate threshold determination, uses precisely the same general logic as the RB calculation outlined above.

The top five VOS players in 2014 were as follows:

  1. Antonio Brown: +8.8
  2. Le’Veon Bell: +7.8
  3. Andrew Luck: +7.7
  4. DeMarco Murray: +7.1
  5. Matt Forte: +6.8

Pass on Antonio Brown this season at your own risk.

Draft Strategy

How does VOS impact draft strategy? Should you draft Andrew Luck in the first round, because he finished with a top-3 VOS in 2014? No. VOS is primarily a guidepost for analyzing player valuations across positions in each round. For example, if your fourth round pick is forthcoming, the VOS metrics makes decision making easier. When deciding between Aaron Rodgers (+5.6 VOS) and Kelvin Benjamin (+2.3 VOS) with the No. 40 pick, VOS demonstrates that Aaron Rodgers will likely return a higher value over the available streaming options. While the Rodgers vs. Benjamin 2-to-1 gap illustrates straightforward decision making, when the gap is tighter, disregarding round and draft position by making selections based strictly on VOS score alone may be sub-optimal.

For instance, there is usually a significantly steeper decline in the quality of RBs and WRs than QBs as the draft goes along. Rather than drafting Andrew Luck early, because he possesses the highest VOS of the players available, a more optimal roster construction may be achieved by locking up high VOS RBs and WRs early, because once you get to round 8, the average VOS for that running back cohort is less than +1.0. On the other hand, the average VOS for quarterbacks like Ben Roethlisberger and Tony Romo, who were available in round 8 of most fantasy leagues in 2014, was over +3.0.

For background, QBs busted at a rate of 16.7-percent in 2014, which was down substantially from 2013 when QBs busted 23.8-percent of the time. Given the low year-to-year volatility of the quarterback position overall, and the low draft capital-adjusted bust rate of quarterbacks, fantasy gamers can be more confident that their 10th round quarterback will produce at a higher level “above stream” than other positions. How confident?

From 2013 to 2014, 13 of the 18 quarterbacks (72.2-percent recurrence) from the 2013 “above stream” cohort (top-18 QBs) also showed up on the 2014 above stream list. However, during that same period, only 24 of 48 running backs (50-percent recurrence) from the 2013 “above-the-stream” list also showed up on the 2014 above stream list. Therefore, you can be 20 to 25-percent more confident that any given quarterback will hold his value above stream from year to year.

The annual turnover at the running back position is relatively high, which is expressed both by the low above-stream-recurrence outline above, and by the position’s ~50-percent bust rate over the last two seasons. However, year-to-year instability of the RB position is skewed by the backs drafted in rounds five through twelve. For example, a 10th round running back had a higher than 50-percent likelihood of busting during the last two seasons, while running backs drafted in the first four rounds busted at a 41.5-percent rate over that same span. Paying an “above VOS” premium for the additional stability offered by the early round (1-4) running back may be logical even if the particular running back’s raw VOS is less than the best available quarterback in the early rounds.

What about the tight end position? Rob Gronkowski is being routinely picked in the first round by value-based drafters because of his extraordinary Value Over Replacement Player (VORP) at a position that features a scarce number of productive fantasy assets. However, Gronkowski’s VOS checks in at No. 15 overall. Gronk’s VOS would be higher if it were not driven down by a relatively high overall TE bust rate, which rivals that of the running back position. Therefore, when taking stream replacement value and positional volatility into account, selecting Gronkowski before a top-6 running back or top-6 wide receiver is likely sub-optimal. Drafting Gronk in 2015, similar to selecting Jimmy Graham in years past, likely only makes sense if he is available in the second round or later. In most formats, the first round should be reserved exclusively for running backs and wide receivers.

Further making the case against drafting a quarterback or tight end in the first few rounds, RB1s and WR1s drafted in the first 12 rounds also had a significantly higher Value Over Stream (VOS) than TE1s and QB1s that were drafted in rounds 1-12 over the last two seasons.

  • RB1 average VOS: +5.1
  • WR1: +4.0
  • TE1: +3.3
  • QB1: +3.0

Drafting running backs and wide receivers early, while saving tight end and quarterback selections for later in the draft, more often results in an optimal roster structure.

To summarize, PlayerProfiler‘s VOS player valuation metric is a helpful guidepost to be utilized during fantasy football drafts. However, selecting players based strictly VOS may not be optimal. A nuanced approach that takes into account an individual player’s raw VOS, while also factoring in the downward slope of the positional round-by-round VOS curve, will often lead to more optimal roster construction.