← College Football

College Football · Weekly Recap

Week 3

Michael Hawkins led the drafted pool with 45.9, JC French led it on usage. 44 ranked players scored nothing.

286 ranked players · top-250 ADP · 20260917, 20260918, 20260919, 20260920, 20260921

Everything the drafted pool did in Week 3: what they scored, what their usage says about next week, and who is worth adding. The one habit worth picking up from this page is reading the Opportunity Points column before the points column — it is the one that travels. Short, and in a fixed order, every week.

Opportunity Points: the leaderboard that travels

Opportunity Points price a player's volume at what that volume returns on average across college football: 0.40 a pass attempt, 0.71 a carry, 2.17 a catch, fitted by least squares over all 21,361 player-games of the 2025 season. It knows nothing about whether anything reached the end zone.

It leads the page because it predicts better than the box score does. Over 16,929 consecutive game pairs in 2025, a player's Opportunity Points correlated 0.507 with his points the FOLLOWING week; the points he actually scored correlated 0.481. A +0.026 edge is small, and it has pointed the same way in every sample we have measured. Volume carries; efficiency mostly does not.

Michael Hawkins threw 10 of 17 for 190 and 2 scores, and added 123 on the ground with 3 more — 45.9 points. Behind him, Kendrick Raphael carried 24 times for 197 and 3 scores (39.8), and LJ Martin carried 15 times for 176 and 3 scores (37.0).

Share of offence is his Opportunity Points as a fraction of everything his team generated that day, which is the version of the number that survives a blowout.

1. 286 ranked players, 3,548 points, 44 of them scoreless

286 ranked players were in action across 3 game days, and scored 3,548 points between them — an average of 12.4 each. Ranked players in action by day: Saturday 252, Friday 29, Thursday 5. Their combined usage priced at 2,925 Opportunity Points, so as a group they came in 623 points above what their volume alone was worth.

44 of them scored nothing. That number is in here on purpose: a player whose team played and who never reached a box score is a zero, and the zeros are what a recap usually hides.

Don't overreact. The median ranked player in this table has 3 games on his 2026 record. Nothing here is a season, and anybody selling you a conclusion off a sample that size is guessing louder than you are.

2. The early round: who paid, and who quietly cost you the week

49 players inside the top 50 of ADP were in action. They averaged 19.5 points against 12.4 for the pool as a whole, which is the number an early pick is supposed to buy.

The three who returned it:

And the silent killers — no disaster on the stat sheet, just a first-round price for a bench-grade afternoon, which is the more expensive kind of miss because nobody drops them:

3. Our preseason board is behind the market, and here is the number

Our college board was frozen on 3 September and has not been touched since, which is what makes it gradeable at all. Ordering the 278 board players whose teams have played, scored against their 2026 points to date, our rank gets a Spearman of 0.216 and the market's ADP 0.501. We are 0.285 behind.

That is not the result we wanted and it is the result we have. On the evidence so far the crowd's draft order has read this season better than ours, and the honest read of a two-week sample is that neither number is settled. We will keep printing it either way; a board nobody grades is a board nobody has to be right about.

4. Buy low: the usage is already there

Every one of these earned starter-grade volume and did not score like it. The volume is the half that repeats, so this is the list to trade for, not the one to drop.

5. Late-ADP adds, in the order I would take them

Ordered by usage rather than by Saturday's points, because usage is what you are actually buying. Each of these went after pick 150 and is playing a starter's share of his offence.

No bid percentages here, and that is a data gap rather than a style choice: college best ball has no public FAAB market and no rostered-percentage feed we can read, so any number in that column would be invented. You know your league better than we do — a two-week starter is a different add in a shallow ten-team pool than in a deep one.

Quarterbacks

Ranked by Opportunity Points rather than by what they scored. The bar for this position this week is 16.5, the median among ranked QBs who played; above it is starter-grade usage.

JC French, Cincinnati (ADP 168). He threw 31 of 43 for 309 and 2 scores, and added 39 on the ground with 1 more, worth 28.3 points. His touches priced out at 28.5 Opportunity Points, 26.4% of everything his offence generated. Season to date: 27.2 a game across 3 games. Start with confidence. Nothing here needed luck — the touches and the scoring agree, which is the combination that holds.

Byrum Brown, Auburn (ADP 24). He threw 20 of 35 for 271 and 3 scores, and added 65 on the ground with 1 more, worth 34.3 points. His touches priced out at 26.7 Opportunity Points, 33.6% of everything his offence generated. Season to date: 27.7 a game across 3 games. Start with confidence. Nothing here needed luck — the touches and the scoring agree, which is the combination that holds.

Marcel Reed, Texas A&M (ADP 32). He threw 26 of 49 for 236 and 1 score, and added 80 on the ground with 1 more, worth 25.4 points. His touches priced out at 26.6 Opportunity Points, 26.4% of everything his offence generated. Season to date: 21.3 a game across 3 games. Start with confidence. The volume earned the points, so both should repeat.

Running backs

Ranked by Opportunity Points rather than by what they scored. The bar for this position this week is 11.5, the median among ranked RBs who played; above it is starter-grade usage.

Matt Fuller, South Carolina (ADP 197). He caught 7 for 85 and 1 score, worth 31.3 points. His touches priced out at 27.2 Opportunity Points, 27.8% of everything his offence generated. Season to date: 20.8 a game across 3 games. Start with confidence. The volume earned the points, so both should repeat.

Cam Cook, West Virginia (ADP 9). He carried 30 times for 123, worth 13.8 points. His touches priced out at 25.5 Opportunity Points, 39.5% of everything his offence generated. Season to date: 17.8 a game across 3 games. Start him. He is being handed a starter's workload and has not been paid for it yet; the workload is the part that shows up again next week.

Antwan Raymond, Rutgers (ADP 9). He carried 21 times for 122 and 1 score and caught 3 more for 39, worth 23.6 points. His touches priced out at 21.4 Opportunity Points, 27.2% of everything his offence generated. Season to date: 17.3 a game across 3 games. Start with confidence. Nothing here needed luck — the touches and the scoring agree, which is the combination that holds.

Receivers

Ranked by Opportunity Points rather than by what they scored. The bar for this position this week is 9.3, the median among ranked WRs who played; above it is starter-grade usage.

Malachi Toney, Miami (ADP 3). He caught 12 for 114 and 1 score, worth 23.4 points. His touches priced out at 26.1 Opportunity Points, 25.4% of everything his offence generated. Season to date: 29.5 a game across 3 games. Start with confidence. The volume earned the points, so both should repeat.

Chris Durr, Maryland (ADP 215). He caught 12 for 109 and 2 scores, worth 28.9 points. His touches priced out at 26.1 Opportunity Points, 26.8% of everything his offence generated. Season to date: 20.2 a game across 3 games. Start with confidence. The volume earned the points, so both should repeat.

Mario Craver, Texas A&M (ADP 45). He caught 11 for 119, worth 17.4 points. His touches priced out at 23.9 Opportunity Points, 23.7% of everything his offence generated. Season to date: 13.6 a game across 3 games. Start him. He is being handed a starter's workload and has not been paid for it yet; the workload is the part that shows up again next week.

Tight ends

Ranked by Opportunity Points rather than by what they scored. The bar for this position this week is 6.5, the median among ranked TEs who played; above it is starter-grade usage.

Luke Reynolds, Virginia Tech (ADP 160). He caught 12 for 107 and 1 score, worth 22.7 points. His touches priced out at 26.1 Opportunity Points, 24.1% of everything his offence generated. Season to date: 20.9 a game across 3 games. Start with confidence. Nothing here needed luck — the touches and the scoring agree, which is the combination that holds.

Patrick Overmyer, Houston (ADP 146). He caught 8 for 72 and 1 score, worth 17.2 points. His touches priced out at 17.4 Opportunity Points, 17.4% of everything his offence generated. Season to date: 12.4 a game across 3 games. Start with confidence. Nothing here needed luck — the touches and the scoring agree, which is the combination that holds.

Dorian Fleming, Maryland (ADP 167). He caught 6 for 72 and 1 score, worth 16.2 points. His touches priced out at 13.0 Opportunity Points, 13.4% of everything his offence generated. Season to date: 7.1 a game across 3 games. Start with confidence. Nothing here needed luck — the touches and the scoring agree, which is the combination that holds.

Paid for, did not show

Read the Opportunity Points column before you cut anyone. A quiet week on real volume is a different problem from a quiet week with no touches, and only one of them is fixed by waiting.

Efficiency: points per Opportunity Point

Points scored for every point their volume was worth, among the 173 ranked players who cleared 8 Opportunity Points. Above 1.00 is a player who beat his usage. This is the column that does NOT predict next week — it is here to show you which of this week's big numbers were bought rather than earned.

Negative regression candidates

The mirror of the buy-low list: they scored well past what their volume was worth. Not a sell signal on its own, but if somebody offers you a real price this is the week to listen.

Best against what he cost

Points per unit of draft cost. A player taken at 200 scoring 20 is a different event from a player taken at 5 scoring 20, and this is the column that says so.

Best once you account for who they played

Raw points reward whoever drew the weakest opponent. This is points above what a player at that position averages against a defence of that strength, which is a different list and usually a more useful one.

Next up

Opponent strength is our power rating, where zero is an average FBS team. The line is the market's, captured before kickoff, and it is context rather than a recommendation.

The hindsight ceiling

Season-long best ball: QB, 2 RB, 2 WR, TE, flex and superflex. If you'd had every pick, the best lineup from this pool scored 281.1, with a QB in the superflex. No single drafter could have rostered all of them, so treat it as the week's ceiling, not a team anyone could have drafted.

Built purely on ADP instead, taking the highest-drafted player available at every slot, the same lineup scored 148.1. That is a gap of +133.0 points. Nobody drafts the ceiling, but the distance between those two numbers is what a roster is playing for.

Worth noting who made it from late: Justin Bowick (ADP 204), Collin Dixon (ADP 194), Luke Reynolds (ADP 160). A ceiling full of players nobody wanted is a board problem, not a luck story.

More from ATO

The college board · Schedule and matchups · Every weekly recap · NFL weekly report

Points are half-PPR, computed from ESPN box scores. A ranked player whose team played and who never reached a box score is counted as a zero rather than dropped, which is the number most recaps leave out. Opponent ratings are ours; lines are the market's, captured before kickoff.

← All recaps

Hear when the next recap is up

Leave your email and we will tell you when a new weekly recap or free college tool goes up. No set schedule, no spam, unsubscribe whenever.

Website

Send it