Cole Keller vs. Kevin Roche Jr.: 2026 stats compared
James Madison vs. 
Georgia Tech
Cole Keller is ahead on 4 of 6 stats, including receiving yards, yards per catch and receiving TDs. Kevin Roche Jr. has the edge on catches. Both are tight ends: Cole Keller for James Madison and Kevin Roche Jr. for Georgia Tech, 2026 season through Oct 24, 2026.
Cole Keller vs. Kevin Roche Jr., stat by stat
| Stat | Cole Keller | Kevin Roche Jr. |
|---|---|---|
| Receiving yards | 353 148th | 273 253rd |
| Catches | 22 217th | 29 124th |
| Yards per catch | 16.0 122nd | 9.4 579th |
| Receiving TDs | 5 25th | 0 |
| Explosive catches (16+ yards) | 8 130th | 5 278th |
| EPA per play | — | 0.32 169th |
The better number on each row is marked. National ranks are among every FBS receivers and tight end with the stat; rates need a minimum sample (50 dropbacks, 20 carries, 8 catches). From the play-by-play, games through Oct 24, 2026.
Where do Cole Keller and Kevin Roche Jr. rank nationally?
Each row is one stat. The dots show where Cole Keller and Kevin Roche Jr. rank among FBS receivers and tight ends with enough snaps to qualify, with each player's rank printed by his dot: further left is better. Hover or tap a dot for the number behind it.
Cole Keller and Kevin Roche Jr., in charts
Cole Keller and Kevin Roche Jr.: career stats
Cole Keller
| Season | Team | Catches | Rec yds | Yds/catch | Rec TD | Rush yds |
|---|---|---|---|---|---|---|
| 2025 | East Tennessee State | 22 | 226 | 10.3 | 2 | — |
| 2026 | James Madison | 22 | 353 | 16.0 | 5 | 0 |
| Career | 44 | 579 | 13.2 | 7 | 0 |
Kevin Roche Jr.
| Season | Team | Catches | Rec yds | Yds/catch | Rec TD | Rush yds |
|---|---|---|---|---|---|---|
| 2026 | Georgia Tech | 29 | 273 | 9.4 | 0 | 0 |
Seasons since 2022 from CollegeFootballData.com (FBS games), plus this season from the play-by-play.
Their seasons, game by game
Cole Keller
TE, James Madison (7–1). Senior. 6-4, 250 pounds. Grain Valley, MO
| Week | Opponent | Receiving yards | Catches | Receiving TDs |
|---|---|---|---|---|
| 1 | Liberty | 30 | 1 | 0 |
| 2 | Wagner | 27 | 2 | 0 |
| 3 | San Diego State | 40 | 3 | 2 |
| 4 | Old Dominion | 21 | 3 | 1 |
| 5 | Marshall | 35 | 3 | 0 |
| 6 | Georgia Southern | 108 | 4 | 1 |
| 7 | Georgia State | 54 | 2 | 0 |
| 8 | James Madison | 25 | 2 | 1 |
| 9 | Troy | 13 | 2 | 0 |
Kevin Roche Jr.
TE, Georgia Tech (2–5). Freshman. 6-9, 265 pounds. Darien, CT
| Week | Opponent | Receiving yards | Catches | Receiving TDs |
|---|---|---|---|---|
| 1 | Colorado | 37 | 4 | 0 |
| 2 | Tennessee | 66 | 6 | 0 |
| 3 | Mercer | 38 | 2 | 0 |
| 4 | Stanford | 59 | 6 | 0 |
| 6 | Duke | 12 | 2 | 0 |
| 7 | Virginia Tech | 29 | 3 | 0 |
| 8 | Boston College | 32 | 6 | 0 |
Questions about Cole Keller vs. Kevin Roche Jr.
Who's better, Cole Keller or Kevin Roche Jr.?
Cole Keller is ahead on 4 of 6 stats, including receiving yards, yards per catch and receiving TDs. Kevin Roche Jr. has the edge on catches. Both are tight ends: Cole Keller for James Madison and Kevin Roche Jr. for Georgia Tech, 2026 season through Oct 24, 2026.
Who has more receiving yards, Cole Keller or Kevin Roche Jr.?
Cole Keller: 353 receiving yards to Kevin Roche Jr.'s 273, 148th among FBS receivers and tight ends.
Who has more catches, Cole Keller or Kevin Roche Jr.?
Kevin Roche Jr.: 29 catches to Cole Keller's 22, 124th among FBS receivers and tight ends.
Who has more receiving touchdowns, Cole Keller or Kevin Roche Jr.?
Cole Keller: 5 receiving TDs to Kevin Roche Jr.'s 0, 25th among FBS receivers and tight ends.
Who has more career receiving yards, Cole Keller or Kevin Roche Jr.?
Cole Keller: 579 to Kevin Roche Jr.'s 273, counting FBS seasons since 2022.
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.
Nerd shit
This week's games in charts, built on every game so far this season. They look a helluva lot better on their own pages, where you can tweak them and share them: copy the image, embed it or post it to X.
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.
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.
- 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.
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.
- 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.
- 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 30, 2026.
There are no accounts and no ad trackers. The only personal information we collect is what you choose to give us: your email address if you sign up for the newsletter, and whatever you put in the contact form.
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. If you use Post on X, we upload that one image so X can show it with your post's link. It stays at that link, with the chart's title and address and nothing about you, for 90 days, then it's deleted. Copy image, Save image and Embed upload nothing.
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.
The contact form
When you send a message through the contact page, we keep it, with the email address and name you give us, so we can answer it. It's emailed to us and stored on Cloudflare. We don't add you to the newsletter or share it with anyone. Ask us and we'll delete it.
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.
Ads
There are no ads on the site yet. When we add them, the ad network and its partners may use cookies and similar tools to show ads and measure them, and we'll name them here, update the date at the top and, where the law requires it, ask before any of that runs.
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.
About
About Big Al's CFB Lab
Big Al's CFB Lab is a college football stats site: a pick for every FBS game, head to heads across 75+ stats, the playoff picture, and charts you can bend to support any narrative (except Big 12 supremacy).
What's here
Every FBS team, player and game gets its own page, with national ranks, the ratings people argue about (SP+, PRIME and Elo) and charts you can share. The rivalry pages carry the series history. The tools let you build your own Top 25, your own 12-team playoff field and your own charts.
Who's Big Al
A college football fan who got tired of arguing without numbers. Big Al calls the Top 25 games, the Game of the week and every coin flip himself, and sometimes goes against the data. The rest of the picks are the ratings' call.
Where the numbers come from
We don't make the ratings; we put them side by side. SP+ is Bill Connelly's. SUPLEX is Nate Tice's. The play-by-play comes from CollegeFootballData.com and cfbfastR. The full list, and how we turn it into picks, is on Sources and methods.
What this isn't
Betting advice. The line shows up because it's part of the story of a game, not because we want you to bet it.
Get in touch
Questions, a stat you want to see, a bug, or a partnership idea: send us a note. The newsletter has the picks and the retro every week.
About
Contact
Questions, ideas, bugs, partnerships: they all come here, and Big Al reads them.
We use what you send only to answer it. See the privacy policy.
- 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 charts 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.
Edit corner labels
Up to 40 characters each. Leave one blank to keep ours. Your labels ride along in the link, the embed and the share image.
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
Edit corner labels
Up to 40 characters each. Leave one blank to keep ours. Your labels ride along in the link, the embed and the share image.
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.