The quarterback audit asked whether FSN identified usable value before kickoff. The running back and wide receiver audit asks a different question: did the model understand role and ceiling?
That matters because skill-position outcomes are much more sensitive to touchdowns, targets and game script. A median projection can be structurally sound and still miss badly in one game.
Report card: five prominent Week 2 projections
Amon-Ra St. Brown — FSN 20.4, actual about 38.2, difference plus 17.8.
Christian McCaffrey — FSN 19.5, actual about 22.6, difference plus 3.1.
Jahmyr Gibbs — FSN 21.7, actual about 23.3, difference plus 1.6.
Derrick Henry — FSN 23.7, actual about 17.7, difference minus 6.0.
Bijan Robinson — FSN 20.3, actual about 11.1, difference minus 9.2.
That gives this sample one massive ceiling hit, two results close to the median and two clear misses below projection.
Ceiling smash: Amon-Ra St. Brown
St. Brown is the most important positive result in the sample because he did not merely beat the projection. He blew through it.
FSN projected 20.4 DraftKings points. St. Brown caught nine passes for 142 yards and two touchdowns against Buffalo.
Under standard DraftKings scoring, that is approximately 38.2 points: nine points for receptions, 14.2 for receiving yards, 12 for touchdowns and a three-point 100-yard bonus.
The raw miss was plus 17.8 points.
Lesson: the median did not predict a 38-point game, but the role gave St. Brown access to one. For tournament DFS, identifying that ceiling environment matters as much as minimizing median error.
On target: Jahmyr Gibbs
FSN projected 21.7 points. Gibbs finished with 52 rushing yards, six receptions for 61 yards and one receiving touchdown.
That works out to approximately 23.3 DraftKings points, only 1.6 above the FSN median.
This is the least dramatic result and arguably the cleanest model success. The role, receiving work and final production landed close to expectation without needing an extreme touchdown outcome.
Lesson: boring accuracy is still accuracy.
Close enough, but for a different reason: Christian McCaffrey
McCaffrey was projected at 19.5 and scored about 22.6.
He rushed 10 times for 23 yards and two touchdowns and added four catches for 43 yards.
The final score was close to the FSN median, but the route there was unusual. The rushing efficiency was poor. Two touchdowns carried the fantasy output.
Lesson: a good final projection does not always mean the model predicted the game script correctly. Outcome accuracy and process accuracy should be audited separately.
Clear miss: Derrick Henry
Henry carried the highest FSN projection in this five-player group at 23.7.
He finished with 68 rushing yards, one rushing touchdown and three catches for 19 yards, approximately 17.7 DraftKings points.
The miss was about six points high.
The role did not disappear. Henry touched the ball 19 times and scored. The gap came from FSN expecting more production from that workload.
Lesson: when volume is present but output falls short, the next audit should separate opportunity forecasting from efficiency forecasting.
Largest miss: Bijan Robinson
FSN projected Robinson at 20.3. He finished with 72 rushing yards, three catches for nine yards and no touchdown.
That is approximately 11.1 DraftKings points, a miss of about 9.2.
Robinson still had 19 combined carries and receptions. The workload was usable. The fantasy result collapsed because the touchdown and receiving-yardage branches never arrived.
Lesson: this is exactly the kind of result FSN should track over time. If the model repeatedly overprojects players whose median relies heavily on touchdown expectation, that is a calibration problem. If these misses occur at the expected frequency inside a wide outcome distribution, the model may be functioning normally.
What the five results actually say
It would be misleading to summarize this as “three hits and two misses.” Projections are not binary picks.
St. Brown was a ceiling event. Gibbs was close to median. McCaffrey’s final score was close, but touchdown variance did much of the work. Henry had the expected role with weaker-than-expected output. Robinson had usable workload but landed in a low-scoring branch.
Those are five different model outcomes, and they should not be graded the same way.
The next FSN DFS report card needs three grades
Median accuracy: how close was the final fantasy score to the projection?
Role accuracy: did FSN correctly anticipate carries, targets and overall opportunity?
Range accuracy: did the final result land inside the modeled floor-to-ceiling distribution?
That framework would turn the DFS audit from a weekly anecdote into something measurable across positions and slates.
Why the archive matters
Anyone can publish DFS recommendations before kickoff. The more valuable test is whether yesterday’s projections remain visible after the box score is known.
FSN already has the salaries, projections, floors, ceilings and post-slate infrastructure to build that record. The goal should not be to prove the model is always right. It should be to learn which parts of the model are right most often.
FSN takeaway: St. Brown showed the value of a high-ceiling role, Gibbs landed almost directly on the median, McCaffrey reached his number through touchdowns, and Henry and Robinson exposed two different forms of downside. That is a more useful report card than simply counting hits.
Sources
FSN Week 2 DFS Lineup Builder and projections: https://footballsportsnews.com/nfl/dfs/lineup-builder/week-2-2026/
NFL Derrick Henry Week 2 stats: https://www.nfl.com/players/derrick-henry/stats/logs/2026/
NFL Jahmyr Gibbs Week 2 stats: https://www.nfl.com/players/jahmyr-gibbs/stats/logs/2026/
NFL Amon-Ra St. Brown Week 2 stats: https://www.nfl.com/players/amon-ra-st-brown/stats/logs/2026/
NFL Bijan Robinson Week 2 stats: https://www.nfl.com/players/bijan-robinson/stats/logs/2026/
NFL Christian McCaffrey Week 2 stats: https://www.nfl.com/players/christian-mccaffrey/stats/logs/2026/
DraftKings NFL scoring reference: https://dknetwork.draftkings.com/2025/08/27/nfl-dfs-beginners-guide-draftkings/