Devin Roche
Old Dominion (2–6, Sun Belt). Running back. 2026 season, games through Oct 24, 2026.
Devin Roche is a running back for Old Dominion (2–6). Among running backs this season he's 36th in carries (102), 76th in runs of 12+ yards (8), 89th in rushing yards (387), in 8 games. This week: Marshall at Old Dominion.
The season
How does Devin Roche rank among FBS running backs?
| Stat | Season | National rank |
|---|---|---|
| Rushing yards | 387 | 89th of 417 |
| Carries | 102 | 36th of 418 |
| Yards per carry | 3.8 | 306th of 360 |
| Rushing success rate | 32.4% | 324th of 360 |
| Runs of 12+ yards | 8 | 76th of 388 |
| Catches | 8 | 147th of 385 |
| Receiving yards | 76 | 116th of 382 |
Ranks are among every FBS running back with the stat this season. Rates need a minimum sample (50 dropbacks for passing rates, 20 carries, 8 catches). From the play-by-play.
Charts
How does Devin Roche compare, in charts?
Game log
Devin Roche's games
| Week | Opponent | Result | Rushing yards | Carries | Rushing TDs | Runs of 12+ yards | Catches |
|---|---|---|---|---|---|---|---|
| 1 | vs. Norfolk State | W 31–10 | 51 | 13 | 0 | 1 | 0 |
| 2 | at Virginia Tech | L 21–44 | 45 | 9 | 0 | 2 | 0 |
| 3 | vs. East Carolina | L 17–20 | 101 | 17 | 0 | 3 | 1 |
| 4 | vs. James Madison | L 20–46 | 18 | 8 | 0 | 0 | 2 |
| 5 | at Georgia State | L 10–42 | 37 | 9 | 0 | 1 | 0 |
| 6 | at App State | W 10–3 | 101 | 23 | 0 | 1 | 1 |
| 7 | vs. Georgia Southern | L 15–23 | 25 | 15 | 0 | 0 | 2 |
| 8 | at Louisiana Tech | L 0–26 | 9 | 8 | 0 | 0 | 2 |
FAQ
Questions about Devin Roche
What position does Devin Roche play?
Devin Roche plays running back for Old Dominion, which is 2–6 this season.
What are Devin Roche's stats this season?
387 rushing yards, 102 carries, 3.8 yards per carry, 32.4% rushing success rate, 8 runs of 12+ yards in 8 games, through Oct 24, 2026.
How does Devin Roche rank nationally?
Among FBS running backs: 36th of 418 in carries, 76th of 388 in runs of 12+ yards, 89th of 417 in rushing yards.
When does Devin Roche play next?
Old Dominion hosts Marshall Sat, Oct 31, 1:00 PM ET on ESPN+. Big Al's pick: Marshall.
Related
College football, 2026 season
Nerd shit for
ball knowers.
Data-driven picks for every FBS game + interactive charts you can customize to support any narrative (except Big 12 supremacy).
Week 9
Game of the week
Picks
Picks and projected scores for every FBS game this week.
Parlay at your own risk
Moonshot of the week.
Nerd shit
This week's games in charts, on the full season's numbers. Screenshot to end (or start) an argument in the group chat.
Week 8
Retro
Every pick, graded.
Rankings
Our Top 25 after last week, and where we split from the AP and Coaches polls.
Playoff field
The 12-team field if the season ended today.
The week in charts
Last week's story, told with that week's games alone. Screenshot to end (or start) an argument in the group chat.
2026
Superlatives
The best, worst, luckiest and most overrated so far. Tap a tile to learn more.
Hot and cold
Who's streaking and who's shitting the bed.
Make your own case
Everything you build can be saved as an image for the group chat.
Betting lines are off
The parlay builder needs the lines, and you've turned them off for this browser. Turn them back on with the Show betting lines switch at the bottom of any page.
Up to 10 legs, one per game: a moneyline, a side against the spread or an over/under. Lines and prices are DraftKings'. When we don't have DraftKings' price on a spread or total, it shows the standard −110. Each leg shows our call on it, from our number (SP+ and PRIME): a Lock, Take or Lean on the side we like, Other side when we'd bet the opposite. Random parlay grabs 3 to 6 games from the list and picks a leg in each, no thinking involved. It's for fun. If gambling stops being fun, call 1-800-GAMBLER.
This week's games
Drag a team by its handle (⋮⋮) to move it; with a keyboard, focus the handle and use the up and down arrow keys. To put a different team in a spot, pick it from that row's menu; if it's already ranked somewhere else, the two trade places. Your ranking is saved in this browser.
Your Top 25
Compare
About
Sources and methods
Everything on this site starts with someone else's data, and we credit all of it. The projections, picks, rankings and tiers are our own math on top.
Where the data comes from
| Source | What we use | How often |
|---|---|---|
| CollegeFootballData.com, via the public cfbfastR files | Schedules, final scores, pregame Elo, play-by-play and box score stats, rosters | Usually posted Sunday or Monday morning after games |
| DraftKings lines (via The Odds API and CollegeFootballData.com) | The line and total on each game, and where each opened | Every morning, and before kickoffs Thursday to Saturday |
| Team logos and colors (ESPN's, via CollegeFootballData.com) | The logos on charts, game cards and team pages. They belong to the schools and are used only to identify the teams. | Preseason |
| SP+ by Bill Connelly (CFB Update) | Overall, offense, defense and special teams ratings | Weekly |
| PRIME (primecfb.com) | Net, offense, defense and special teams ratings, strength of schedule, strength of record | Weekly |
| AP Top 25 (via NCAA.com) | Rankings and points, including others receiving votes | Sunday afternoons |
| AFCA Coaches Poll (via Yahoo Sports, CBS Sports, SI) | Rankings and points, including others receiving votes | Sundays |
| 247Sports | Team talent composite, star counts, 2026 transfer portal rankings | Preseason |
| The NIL Standard | Estimated NIL roster values | Preseason |
| us-atlas (U.S. Census Bureau boundaries) | State shapes for the map | Once |
How we turn it into projections
- Projected score. Each team's SP+ offense, plus how far the opponent's SP+ defense is from average, plus special teams. The home team gets half of a 2.5-point home edge added and the visitor half subtracted.
- Win chance. The projected margin turned into a probability, assuming games usually land within about 14 points of the projection. A 0-point margin is 50%; 14 points is about 84%.
- The three ratings. SP+ is built from every play this season. PRIME is a possession-efficiency rating. Elo uses only final scores and carries over from year to year, with 55 Elo points of home edge.
- Big Al's pick. The team at least two of the three ratings favor, with our confidence the average of their three win chances for that team. On the game cards, a pick is a lock when all three favor the same team and the average win chance is 80% or more, a coin flip when it's under 55%, and a lean otherwise. Injuries can drop the tier: each key player out counts by what he does for his team (a starting quarterback is 2 points, the feature back or top receiver half a point, doubtful counts half), the other side's injuries offset, and 1 point drops a tier, 2 points two. The game page shows it as an injury alert. On the Top 25 games, the game of the week and every coin flip, Big Al makes the call himself, sometimes against the numbers, and says why.
- Totals and moneylines. Over/unders are graded on how far the SP+ projected total is from the book's: a take at 6 points or more, a lean at 3, a pass under that. There's no lock, because only SP+ projects totals. On moneylines we say who we think wins. Everywhere we bet, we show the book's price and our call with the reason, not odds of our own.
- Big Al's Top 25. SP+, PRIME and strength of record, each converted to a common scale and averaged.
- Playoff picture. The 2026-27 rules: automatic bids for the ACC, Big Ten, Big 12 and SEC champions (the current conference leader stands in until the title games) and for the highest-ranked team from the American, Conference USA, MAC, Mountain West, Pac-12 or Sun Belt. Notre Dame is in if it ranks in the top 12. The rest are at-large teams by the chosen rating, and all 12 are seeded straight by that rating, so the top four get the byes. Once the committee's rankings are out (first release Tuesday, Nov. 3), they become the default ranking.
- Missing data. When a team's play-by-play is clearly incomplete (fewer than 45 offensive plays a game), we hide its offense stats instead of showing bad numbers.
Rushing yards here exclude sacks, so they run higher than official totals. Ranks are out of the 138 FBS teams.
About
Glossary
Every stat and term on the site, explained like you're five. Or like you're a grown adult who never asked what EPA means and is now too embarrassed to.
Betting
The words on the betting side of the site, explained like you've never placed a bet. Good for you.
- Spread
- The handicap the book puts on the favorite to make a game a fair bet. Texas −21 means Texas has to win by 22 or more for a Texas bet to cash. Take Mississippi State +21 and you win if they lose by 20 or fewer, or win outright.
- Favorite and underdog
- The favorite is the team the book expects to win, and it carries the minus sign (−7). The underdog carries the plus sign (+7).
- Cover
- Beat the spread. A favorite covers by winning by more than the spread. An underdog covers by losing by less than the spread, or by winning.
- Against the spread (ATS)
- A record of covers instead of wins. A team can be 7–0 and 1–6 against the spread if it keeps winning by less than Vegas expected.
- Push
- The game lands exactly on the number, say a 21-point win with a 21-point spread. Nobody wins and you get your money back.
- Pick'em
- A game with no favorite: the spread is zero, so you just pick the winner.
- Moneyline
- A bet on who wins, no spread involved. Favorites have a minus price: at −200 you bet $200 to win $100. Underdogs have a plus price: at +300 you bet $100 to win $300.
- Total (over/under)
- The book's number for the two teams' combined points. Bet the over if you think they'll score more, the under if you think they'll score less.
- −110, the juice and the vig
- −110 is the standard price on spreads and totals: you bet $110 to win $100. The extra $10 is the book's cut, called the juice or the vig. It's why you need to win about 52.4% of your bets just to break even.
- Implied odds
- The win chance a price works out to. −110 is about 52.4%, +300 is 25% and −200 is about 66.7%. Add up both sides of a game and you get more than 100%. The extra is the book's cut.
- Parlay
- Several bets stacked into one. Every leg has to hit, and the payout multiplies with each one. Five coin flips at −110 pay about 24 to 1 and hit about 3% of the time.
- Leg
- One bet inside a parlay.
- Opening and closing line
- The opening line is the first number posted for a game. It moves as money comes in and news breaks. The closing line is the last number before kickoff. Getting a better number than the close is the best sign a bettor knows what they're doing.
- Sprinkle
- A small bet you can afford to lose. The right size for anything you read here.
- Parlay at your own risk
- Our parlay of the week on the home page: two real edges, one total, one underdog we kind of like and one moonshot. Our numbers pick it. Your bank account pays for it.
Ratings and advanced stats
What the numbers on team pages, charts and head to head mean.
- SP+
- Bill Connelly's rating, built from play-by-play efficiency and explosiveness and adjusted for who each team played. The number is how many points better than an average FBS team a team is on a neutral field. SP+ 20 beats SP+ 10 by about 10.
- PRIME
- A rating built possession by possession: how well a team turns drives into points and keeps opponents from doing the same, adjusted for opponents. Zero is an average FBS team.
- Elo
- A results-only rating. Every team has a number, 1500 is average, and the winner takes points from the loser after every game. Beat a great team and you take a lot. Beat a bad one and you take a few. It carries part of last season over, so it's slow to catch a team that changed.
- Strength of record
- How impressive a record is, given who a team played: how many more games it won than an average team would have against the same schedule. It grades the resume. SP+ and PRIME grade how good the team is.
- Strength of schedule
- How hard a team's opponents have been so far.
- Opponent-adjusted
- Corrected for who a team played. Hanging 50 on an FCS team counts for less than hanging 30 on Georgia.
- EPA (expected points added)
- Every down, distance and spot on the field is worth some number of points on average. 1st and 10 at the opponent's 20 is worth a lot more than 3rd and 15 at your own 10. EPA is how much one play changed that number. A 50-yard bomb is a big plus. A pick-six is a very big minus.
- EPA per play
- A team's average EPA across all its plays. For an offense, above zero means it adds points and higher is better. For a defense, lower is better.
- Success rate
- The share of plays that keep an offense on schedule: half the yards needed on 1st down, 70% of them on 2nd, all of them on 3rd and 4th. A 4-yard run on 1st and 10 is a success. A 4-yard run on 3rd and 6 isn't.
- Explosive play rate
- The share of plays that go for a run of 12 or more yards or a completion of 16 or more.
- Yards per play
- Total yards divided by plays: runs, passes and sacks.
- Sack rate
- Sacks divided by dropbacks. For an offense, lower is better. For a pass rush, higher is better.
- Turnover margin
- Takeaways (interceptions and fumbles recovered) minus giveaways (interceptions thrown and fumbles lost).
- BAWA (Big Al Wins Added)
- Where a team's wins came from. Every team starts as a .500 team. The record above or below .500 gets split four ways that add up to the real record: schedule (what an average team would win against that schedule), roster (the wins its talent and returning starters are worth on paper), coach (wins from playing better or worse than that roster) and luck (actual wins against the wins its SP+ says it should have). On team pages and in two charts.
- Schedule, Roster, Coach and Luck BAWA
- The four parts of BAWA. Schedule: what an average team would win against that schedule, minus .500. Roster: the extra wins from a roster this good on paper (247Sports talent and returning production). Coach: wins from playing better or worse than that roster should. Luck: actual wins against the wins SP+ expects. Recruiting rankings undersell the service academies, so their coaches get extra credit.
- Luck watch
- A tag for a team whose Luck BAWA is 1.5 wins or more either way: "Lucky +2.4" or "Unlucky −1.6." Its record runs ahead of (or behind) how it's played, and lines and polls often follow records. It doesn't change our calls; the record page tracks how fading luck does against the spread.
- Expected wins
- The wins a team should have, adding up its win chance in every game it played. Winning more than that usually means a lot of close wins, and close wins are mostly luck.
- Talent composite and blue-chip ratio
- 247Sports' score for a roster, built from every player's recruiting rating, with the best players counting most. The blue-chip ratio is the share of the roster that were 4- or 5-star recruits.
How we pick
How Big Al's picks, tiers and projections work. The full method is on Sources and methods.
- Projected score
- Each team's SP+ offense, adjusted for how good the other team's defense is, plus special teams. The home team gets 2.5 points of home edge, split between the two scores.
- Win chance
- The projected margin turned into a percentage, assuming games usually land within about 14 points of the projection. A tie projection is 50%. A 14-point favorite is about 84%.
- Big Al's pick
- The team at least two of SP+, PRIME and Elo favor. The confidence is the average of their three win chances.
- Lock, Lean, Coin flip and Stay away
- How sure Big Al is about a straight-up pick. Lock: SP+, PRIME and Elo all favor the same team and the average win chance is 80% or more. Lean: the ratings favor the team, but not that strongly or not all three. Coin flip: the average win chance is under 55%. Stay away: injuries take the pick down. Key players count by what they do for the team: a starting quarterback out is 2 points, the feature back, top receiver or top pass rusher half a point, the No. 2 receiver or playmaker a quarter, and a short-handed line or secondary half a point. Doubtful counts half. When the picked team's points, minus the other side's, reach 1, the pick drops a tier (a Lock to a Lean, a Lean or Coin flip to Stay away); at 2 it drops two.
- Lock, Take, Lean and Pass
- Our confidence in a pick against the spread. Lock: all three ratings have the team covering and our number beats the line by 3 or more. Take: all three, by 1.5 or more. Lean: at least two, by 1 or more. Pass: too close to call.
- Take, Lean and Pass on totals
- Our confidence in an over or under. Take: the SP+ projected total is 6 or more points from the book's. Lean: 3 or more. Pass: closer than that. There's no Lock, because only SP+ projects totals and a Lock needs all three ratings to agree.
- Our number
- Our projected margin for betting: the average of the SP+ and PRIME margins. If our number has Texas by 15 and the line is Texas −21, we like the underdog.
- Edge
- How far our number is from the line, in points. Texas by 15 against Texas −21 is a 6-point edge on the underdog.
- Game of the week
- Big Al's pick of the one game to watch on Saturday, usually two good teams in a game we expect to be close.
- Big Al's Top 25
- Allen's own ballot, set every Sunday. The numbers-only version (SP+, PRIME and strength of record, averaged) is on Rankings.
- Retro
- Monday's look back: every final score against what we projected, what we got right and what we got wrong.
Conferences and the playoff
Who's who and how the 12-team field works.
- FBS and FCS
- FBS is the top level of college football: the 138 teams on this site. FCS is the level below. When an FBS team plays an FCS team, it's usually a paid tune-up.
- Power 4
- The SEC, Big Ten, Big 12 and ACC. Notre Dame isn't in a conference, but we count it with the Power 4 wherever that matters.
- Group of 6
- Everyone else in FBS: the American, Conference USA, MAC, Mountain West, Pac-12 and Sun Belt, plus the independents other than Notre Dame. The highest-ranked team from those six conferences gets an automatic playoff bid.
- Automatic bid
- A guaranteed playoff spot. The ACC, Big Ten, Big 12 and SEC champions get one, and so does the highest-ranked team from the American, Conference USA, MAC, Mountain West, Pac-12 and Sun Belt.
- At-large
- A playoff spot earned by ranking high enough without an automatic bid. Notre Dame is in if it's in the top 12.
- First-round bye
- Seeds 1 through 4 skip the first round. Seeds 5 through 12 play first-round games at the higher seed's stadium.
Nothing matches that. Try a shorter word.
For where the numbers come from and the exact math, see Sources and methods. If gambling stops being fun, call 1-800-GAMBLER.
About
Terms of use
Last updated October 26, 2026.
This is a free site for fun and information. Use it, share it, argue with it, but don't treat it as a promise about Saturday.
What this site is
Big Al's CFB Lab publishes college football statistics, projections and opinions. Everything here is for information and entertainment. Projections and picks are estimates from public data and can be wrong, sometimes badly.
Not betting or financial advice
Nothing on this site is advice to place a bet. If you choose to bet, you're responsible for following the laws where you live, and you must be of legal age. If gambling stops being fun, call or text 1-800-GAMBLER.
Data and accuracy
Our numbers come from the sources listed on the Sources and methods page and from our own calculations. We try to keep them accurate and current, but we can't guarantee that every number is complete, correct or up to date.
Sharing
You're welcome to share images you make with the share tools and to link to any page. Please keep the Big Al's CFB Lab credit on shared images. Team and school names belong to their owners; we use them only to identify teams.
Using the site fairly
Please don't overload the site, try to break it, or copy the whole thing to republish as your own.
No warranty
The site is provided as is. To the extent the law allows, we aren't liable for losses that come from using it or relying on it.
Changes
We may update these terms as the site grows. The date at the top shows the latest version.
About
Privacy policy
Last updated October 26, 2026.
There are no accounts and no ad trackers. The only personal information we collect is your email address, and only if you sign up for the newsletter.
What stays on your device
The site remembers a few choices in your browser's local storage so they're there next time: the teams you last compared, your filters and chart settings, and any Top 25 you build. That information stays in your browser. It isn't sent to us, and you can clear it any time by clearing your browser's site data.
Share images
Images you make with the share tools are created in your browser. We don't upload or keep them.
The newsletter form
When you sign up for the newsletter, your email address goes to Beehiiv, the service that sends it. We use it only to send the newsletter. Every issue has an unsubscribe link, and unsubscribing removes you from the list.
Hosting
Our host, Cloudflare, keeps standard server logs, such as IP addresses and the pages requested, to run and protect the site. We count visits with Cloudflare Web Analytics, which doesn't use cookies or track you across sites.
Children
The site isn't directed at children under 13, and we don't knowingly collect information from them.
Changes
If this policy changes, we'll update the date at the top.
- 12 teams, under the 2026-27 rules. Automatic bids go to the ACC, Big Ten, Big 12 and SEC champions, and to the highest-ranked team from the American, Conference USA, MAC, Mountain West, Pac-12 or Sun Belt. Until the title games, each conference's current leader stands in as its champion.
- Notre Dame is in if you rank it in your top 12. Everyone else is at-large.
- Seeds go straight by your order. Seeds 1 to 4 get first-round byes; 5 to 12 play first-round games at the higher seed.
- The builder won't let you remove a Power 4 champion or your last team from outside the Power 4.
Drag a team by its handle (⋮⋮) to reseed it; with a keyboard, focus the handle and use the up and down arrow keys. Pick a different team from a row's menu to swap it in. Your field is saved in this browser.
Your 12 seeds
Your bracket
What's in it
- Thursday: the picks. The week's best games, the upset we like, projected scores for the big ones, and one chart worth arguing about.
- Monday: the retro. How the picks did, the record so far, what we got wrong and why.
- Every issue, on the site. Past issues will live here, so you can link anyone to them.
The first issue is coming soon.
What this shows
Beating the trend
Trailing the trend
Projected scores come from SP+ offense, defense and special teams, plus 2.5 points for the home team. Cards also show DraftKings' line and total as facts: we pick winners, not bets. Open Details for the full read, or Head to head for every stat.
This week
Final scores against what we said before kickoff. Big Al's pick is the team at least two of SP+, PRIME and Elo favored, unless he made the call himself on a toss-up or one of the week's biggest games. Ranked means in that week's AP poll. Open a game for the full preview as it stood going into the week.
Last week
All-time picks record
Straight-up picks: did the pick win? Big Al's pick is the team at least two of SP+, PRIME and Elo favor. Big Al and SP+ picks start in Week 8; Elo picks go back to Week 1.
Week by week
Each column is that week's ranking. Arrows compare the latest poll to the week before. NR means not ranked.
Rankings come from cfbtrack.com. Week 1 is the preseason poll; each later week is the poll released before that week's games.
Over time
What this shows
Rank
What this shows
Compare
What this shows
SP+ vs. payroll is the average, across a conference's teams with an NIL valuation (positive means its teams play better than their payrolls predict). Blue-chip ratio is the share of all its rostered players who were 4- or 5-star recruits.
Ranks are national, out of 138 FBS teams. A top-10 team also counts toward the top 25 and top 40, and everyone outside the last cutoff counts as weak. Open a conference to see its teams.
Conference depth
Projected bracket
Seeds 1–4 get first-round byes. Seeds 5–8 host the first round.
The field and the bubble
The columns on the right show where each team would be seeded under the other ratings. A dash means out.
How the field is built
- 12 teams. Five get in automatically: the ACC, Big Ten, Big 12 and SEC champions, plus the highest-ranked team from the American, Conference USA, MAC, Mountain West, Pac-12 or Sun Belt (champion or not).
- Notre Dame is in if it ranks in the top 12.
- At-large. The remaining spots go to the highest-ranked teams left. If an automatic qualifier ranks outside the top 12, it takes the last at-large spot.
- Seeding follows the ranking, as it has since 2025. A conference champion isn't guaranteed a bye.
- Champions mid-season. "Current conference leader" uses the best conference record so far, with ties broken by the chosen rating. "Highest-rated team" picks the best-rated team in each conference instead.
- Composite averages SP+, PRIME and strength of record after putting each on a common scale.
- This is a projection from today's numbers. The real field is chosen by a committee on Dec. 6, after conference title games.
How we picked it
How we picked it
Three ratings look at a team in three different ways. The numbers' pick is the team at least two of them favor. On the toss-ups and the biggest games, Big Al makes the call himself and says why.
Where the game is won
Where the game is won
Each row puts one team's unit against the unit it faces, by national rank out of all FBS teams. An edge is a gap of 25 spots or more.
Injuries and availability
Statuses come from the conference availability reports and team news, as of the date shown. "Hasn't played" means a regular (the main passer, a quarter of the carries or 15% of the catches) with no pass, carry or catch in his team's last two games, from the play-by-play. That can be an injury, a benching or a suspension. Linemen and defenders don't show up in that check.
The teams in charts
The players in charts
FAQ
Keep going
Ratings: SP+ via CFB Update; PRIME via primecfb.com. Box and play stats computed from CollegeFootballData.com play data (via the public cfbfastR files), FBS games only for scoring. Rushing yards here exclude sacks, so they run higher than official totals. Ranks are out of 138 FBS teams.