clock menu more-arrow no yes mobile

Filed under:

2023 Transfer Rankings Review: Quarterbacks

Looking back at how well last year’s transfer rankings translated to production for the QBs

NCAA Football: Stanford at Colorado Ron Chenoy-USA TODAY Sports

Last year I put together my own formula to try to come up with a data-driven approach to ranking players in the transfer portal. I tried to use logic to include the elements that I thought had the most relevance when trying to decipher how a player performed. Now though we have a year’s worth of data to be able to compare how the players I ranked actually played on the field in 2023 and we’ll start that review with the quarterbacks.

Now, how well some of these transfers end up looking hasn’t been finalized yet. Washington only got a few games out of Ja’Lynn Polk the first season after he transferred to the Huskies from Texas Tech. Then in his second year he showed he was a very good #3 WR. And after his third season he became a likely 2nd round draft pick.

Still, I think it’s fair to take a first look at how my rankings compared to what actually transpired through year one. Some players had multiple years of eligibility remaining and some were grad transfers playing their final college season no matter what.

We’ll start with my ranking’s top-ten:

1. Shedeur Sanders, Jackson State to Colorado- 92.5 grade

2. DJ Uiagelelei, Clemson to Oregon State- 92.1

3. Spencer Sanders, Oklahoma State to Ole Miss- 91.4

4. Tanner Mordecai, SMU to Wisconsin- 91.1

5. Hank Bachmeier, Boise State to Louisiana Tech- 89.9

6. Graham Mertz, Wisconsin to Florida- 89.3

7. Emory Jones, Arizona State to Cincinnati- 88.9

8. JT Daniels, West Virginia to Rice- 86.9

9. Devin Leary, NC State to Kentucky- 85.5

10. Brennan Armstrong, Virginia to NC State- 85.3

Some of those players did better than others. 9 of the 10 started at least 8 games for their new team with only Spencer Sanders failing to make any starts. I think it’s very easy to make the case that Shedeur Sanders was correctly viewed as the #1 QB. He finished 6th among all transfer QBs in total yards (2nd among P5) and tied for 1st in highest PFF grade (min 15 snaps). That’s a pretty solid start.

We want to try to get a sense for the rankings as a whole though. To help answer that question I put together a productivity score which combines each player’s snaps played in 2023, their PFF grade, and total starts into one number for comparison across the spectrum. I broke down the players into buckets based on their transfer grades (90+, 80-89, 70-79, etc.) and looked at the average productivity score for each. Note that I rounded up for players with for example a 79.5 to be in the 80+ bucket.

2023 Quarterback Transfer Performance by Rating

Tranfer Rating Group Average Prod Score Average # Starts Average # Snaps Average PFF Grade Average Total Yards
Tranfer Rating Group Average Prod Score Average # Starts Average # Snaps Average PFF Grade Average Total Yards
90+ 53.7 8.2 550.6 82.9 2,302.4
80-89 45.3 8.6 512.4 64.7 1,995.4
70-79 48.8 8.4 560.9 66.4 2,257.8
60-69 29.0 4.4 330.6 60.6 1,200.8
50-59 28.5 4.8 326.6 59.6 1,287.0
40-49 16.3 2.8 199.3 56.7 816.5
30-39 8.3 1.5 115.1 47.2 289.6
20-29 12.0 2.1 146.1 34.9 503.2
10-19 19.7 4.1 261.0 51.0 869.0
0-9 1.6 0.5 26.5 20.7 53.0

For the most part there’s a pretty clear trend between the different groupings which is what we’d like to see. There are only 5 players that earned a 90+ grade but on average they started 8.2 games, took 551 snaps, had an 82.9 PFF grade, and 2,302 total yards for a 53.7 productivity score. The average PFF grade, total yards, and productivity score were higher than those of any other group as we would hope.

It was close to a toss-up between the 70+ and 80+ buckets. They were within 0.2 starts of one another and 50 snaps of one another. The playing time was nearly equal but the performance from that 70 bucket was just a little better. For all intents and purposes though, it was fair to say that any player that finished with a score between 70 and 90 had about the same chances of success in 2023 with their new school.

As we continue looking across the groups there was again almost identical scores between the 60-69 grouping and the 50-59 one. They were within 0.5 starts, 5 snaps, 90 yards, and 1.0 on their productivity score. Again, it’s fair to say there wasn’t much of a cutoff between that 60 demarcation and the next group down.

After you get below a 50.0 rating though there aren’t very many success stories. The 10-19 grouping has a bump up but those results took place largely at FCS programs with full-time starters at UT-Martin, Alabama St, Stony Brook, and Jackson State all in that category.

Biggest Overachievers

All 3 of these stand out as the clear examples to look at when trying to figure out some lessons learned to inform the formula in future years.

1. Jordan McCloud, Arizona to James Madison- 47.5 grade, 81.6 productivity

McCloud started 9 games as a low three-star true freshman at South Florida for a team that went 4-8 putting up meh numbers (12 TDs, 8 INTs, 6.3 YPA). As a sophomore he started 7 games in the weird Covid year for a 1-8 team with 9 TDs, 2 INTs, and 6.9 YPA. McCloud then transferred to Arizona and in 2 seasons threw more picks than touchdowns with once again a YPA mark in the 6’s.

Then McCloud transferred to James Madison and they almost went through the season undefeated while he had 43 total TDs and threw for 8.8 YPA. There was absolutely no way to predict that other than maybe that he was a 5th year player who was transferring down a level.

2. Chandler Rogers, ULM to North Texas- 60.3 grade to 85.5 productivity

Chandler Rogers also started as a true freshman (unrated in the 247 Sports Composite), beginning 7 games and putting up solid numbers at ULM with 9 TDs and 3 INTs on 7.3 YPA. PFF was much less optimistic about his play though as they credited him with 10 turnover-worthy plays despite the 3 actual interceptions. He improved as a true sophomore starting all 12 games with slightly better advanced numbers and a similar YPA total. Then he transferred to North Texas and shattered all his previous career bests with +0.6 YPA and +14 TDs on just 65 more dropbacks.

3. Haynes King, Texas A&M to Georgia Tech- 57.6 grade to 82.3 productivity

The last example is Haynes King who was a clear four-star prospect headed to a recruiting power in Texas A&M. He redshirted as a freshman but won the starting job out of camp in his sophomore year. Unfortunately, he fractured his leg in the first quarter of their 2nd game and missed the rest of the season. He regained the starting job in year 3 but didn’t play well coming off the injury and was benched. King transferred to Georgia Tech and threw a ton of picks but was one of the better dual threat QBs in the country now as a senior.


So maybe the lesson is look for quarterbacks that won starting jobs in their first two seasons but then ran into bad luck either with injuries or bad situations? Either way, it doesn’t seem super sustainable to carry forward into future formulas.

Biggest Underachievers

1. Spencer Sanders, Oklahoma State to Ole Miss- 91.4 grade, 5.8 productivity

Sanders was a four-year starter at Oklahoma State and while he had never come close to being an All-American type player, he was at least an average Big 12 quarterback with plenty of experience. He knew he wasn’t guaranteed a starting job and Ole Miss actually also added a highly rated developmental prospect in Walker Howard through the portal to go with incumbent starter Jaxson Dart. In the end, Dart won the job and had a very good season while Sanders was the backup and hardly played.

2. Chance Nolan, Oregon State to TCU- 84.2 grade, 0.0 productivity

Nolan showed a lot of promise at times playing for Oregon State but was a little too loose with the ball for Jonathan Smith’s liking. He suffered a concussion and never regained his job as OSU relied on their running game and defense formula with a game manager to keep it on track. Nolan transferred to TCU and was seemingly in a competition to win the starting job but abruptly left the program at the end of the first week of preseason camp for undisclosed reasons and didn’t return.

3. Tyler Buchner, Notre Dame to Alabama- 79.1 grade, 3.6 productivity

Buchner began the 2022 season as the starting quarterback for Notre Dame under new head coach Marcus Freeman but struggled. In their first two games he threw 0 TDs and 2 INTs as the Irish lost to both Ohio State and at home to Marshall. That was enough to get Buchner benched and he only got back into the game for their bowl once new starter Drew Pyne also transferred. Buchner followed his OC Tommy Rees to Alabama but it was clear he wasn’t expected to be the starter there and he in fact wasn’t.


If there’s any lesson to be learned here then maybe it’s don’t trust quarterbacks who played at major programs despite throwing a lot of interceptions. That might be a sign they are playing because there aren’t other options rather than their true talent level. But it seems clear that Sanders would’ve started at many programs, just not Ole Miss. And you can’t predict that Nolan will leave the team in August. Once again, not sure there are true takeaways that are usable.

Comparison to 247 Sports Rankings

Since 247 Sports is now the go-to resource in the recruiting business, we’ll look at their transfer portal rankings from last year. They gave 16 quarterbacks from the portal at least a 4-star rating in 2022-23. Here’s how many of that group actually finished in the top-16 QBs at FBS schools in each category for me versus 247 Sports:

Total Starts: 247- 6, Me- 6

Total Snaps: 247- 4, Me- 4

PFF Grade: 247- 6, Me- 7

Total Yards: 247- 4, Me- 5

Productivity Score: 247- 4, Me- 4

Overall, I think you an say that it is fairly close. There’s obviously a good amount of overlap between all of these categories. That means that each of us had our own group that “hit” at the top of the rankings with some overlap. Here’s a list of the actual top-16 in productivity based on which of us accurately forecast it.

We Both Got Right (3): Sheduer Sanders (Colorado), DJ Uiagelelei (Oregon St), Graham Mertz (Florida)

Only I Got Right (1): Emory Jones (Cincinnati)

Only 247 Got Right (1): Hudson Card (Purdue)

Neither Got Right (11): Chandler Rogers (UNT), Haynes King (Ga Tech), Jack Plummer (Louisville), Jordan McCloud (JMU), TJ Finley (Texas St), Alan Bowman (Ok State), Davis Brin (Ga Southern), Thomas Castellanos (Boston College), Donovan Smith (Houston), Mikey Keene (Fresno St), Kyron Drones (Virginia Tech)

Overall, I think you can say that we didn’t exactly do great in this exercise. Things would’ve improved for both if you opened things up a little farther. The next 5 players in productivity score included 4 players that were in the top-16 for both of us (Sam Hartman, Devin Leary, Tanner Mordecai, and Brennan Armstrong).

A lot of players from good G5 programs ended up making the list that neither of us exactly saw coming like Rogers, McCloud, Finley, Brin, and Keene. Would some of the quarterbacks we had higher in the rankings have beaten out those same players if they had transferred to a G5 level? Probably. But that’s how it goes.

In the end I think it’s fair to say that my system was slightly better than 247 Sports at accurately picking out the top-16 QB transfers from last cycle. It’s not exactly a surprise that the flaw in the 247 Sports methodology appears to be that they are overrating recruiting ranking given that’s kind of their thing.

There were 3 of their top-16 that ultimately played fewer than 15 snaps for their new team and all of them earned at least a 0.93 rating in the composite out of high school. Maybe they’ll win the job for those schools next year and prove the rankings right in the end but none appear to be a favorite to start next year. That compares to just 1 QBs in my top-16 that didn’t play at least 15 snaps (Chance Nolan).


Here you can see how each of the 127 quarterbacks that started fall camp on the roster for a new team after transferring fared with their portal rating in my system and then their actual productivity score.

The r-squared value that is on the trendline represents how much of the variation is due to the actual relationship between the two variables. So about 26% of the difference is explained by my portal formula.

If this were most things in a business/science environment, that wouldn’t be anywhere close to substantive. When you’re talking about trying to predict the results for individual college players who are about to change systems, depth charts, and potentially levels of competition...I think it’s not too shabby.

I tried slicing things a couple different ways to see if it would improve the reliability. If you only include players who transferred into power conference schools does it get better? What about if you exclude players who didn’t play this year in case that was due to injury or another fluke occurrence?

Going with just the FBS players did increase the r-squared value slightly up to 0.302. That makes sense since the FCS QBs are ones who almost certainly wouldn’t have seen playing time anywhere else and it takes away some of the dots you see in the above graph with a portal rating of around 20 and a production score greater than 30.

Trying to only look at the P5 actually lowers the r-squared. That was somewhat surprising to me because of the number of players like Chandler Rogers at North Texas who my system didn’t expect very much of but put up big numbers playing at that level. However, there are also a lot of players who did poorly or didn’t play at major schools that went down to the G5 level and still didn’t play very much which helped the overall performance.