Notre Dame Took Seven from the Portal. One Thousand Two Hundred Other Players Have Nowhere to Go.
The NCAA voted last October to consolidate college football’s transfer windows. The old system had two: a long December window after the regular season ended, and a shorter April window after most programs finished spring ball. Coaches argued the spring window disrupted practice and rewarded last-minute roster shopping over patient development. The April window is gone. The new single window, January 2 through January 16, is the only path for an FBS player to change schools. Two CFP championship-game participants get a five-day extension after the title game. Everyone else — coaches, players, athletic departments — lives with whatever they signed by January 17.
The first round of consequences arrived on the players’ side. By the close of the window, an estimated 1,200 FBS-experienced players who had entered the portal had not found a new program. Under the old rules, most of them would have waited for the April window, watched spring depth charts shake out, and tried again. Under the new rules, they are stuck. Some will return to their original programs. Some will drop a level — FCS, Division II, junior college. Some will leave football. None of them will play 2026 college football for a different FBS team unless their current school grants a release through some narrow medical or hardship channel.
The Notre Dame Class, in One Paragraph
Notre Dame closed the window with seven commitments. Two are wide receivers who left Ohio State after the playoff run: Quincy Porter and Mylan Graham, both former four-star recruits with two seasons of eligibility remaining. The defensive front added Tionne Gray, a defensive tackle from Oregon, and Francis Brewu, a defensive lineman from Pittsburgh who had seven tackles for loss in 2025. Three additional players filled depth needs at safety, linebacker, and place-kicker. None of the departures were starters. Marcus Freeman’s class was rated a top-five portal class by the major recruiting services, with a particular emphasis on plug-and-play immediate contributors rather than developmental projects.
This is the kind of class that was, under the old system, supplemented by spring shopping. A coaching staff would assess April practice, identify two or three weaknesses, and pick from the spring portal. That tool is gone. Whatever Notre Dame committed to in mid-January is the team that will line up against Miami on August 30.
Why a Locked Roster Looks Like the NFL Now
The Roster-Construction Calendar Has Compressed
The closest analog is NFL roster construction, where a team has roughly one major spending window each spring (free agency in March + the draft in late April), followed by minicamps and training camp where the roster is essentially set. NFL teams cannot, under most circumstances, sign a frontline starter in July. The 2026 college football calendar now resembles that compressed structure. The portal opens, the portal closes, the team is the team.
This has predictable second-order effects. Front offices — or whatever the college equivalent now is — are forced to do their entire roster math in two weeks rather than five months. Decisions made under that pressure are noisier than decisions made with a spring re-look option. The variance of outcomes goes up. Some teams will get the January window very right; others will get it very wrong; and the ones who got it wrong will find out without a corrective tool.
The Empirical Question
The methodological piece this newsletter has been developing — that small samples lie, that stabilization takes longer than fans want it to, that one window does not give you enough information to know if you got it right — applies to the new portal calendar more than it applied to the old one. Under the two-window system, a January class that underperformed in spring practice could be patched. Under the single-window system, the January class is the bet. By the third game of the season we will have early signal. By Halloween we will know which power-conference programs picked well. By December we will have a quantifiable answer to a question college football has never previously had to answer in this form: given exactly one shot at the roster, who got it right?
| What Will Be Visible By | What We’ll Know |
|---|---|
| Week 4 (mid-Sep) | Which January-window starters are real, which were class ranking artifacts. About 30% of true talent is visible by here per Issue #37’s stabilization curve. |
| Week 8 (late Oct) | About 60% of true talent visible. Conference standings start to mean something. Programs with portal misses start to look conspicuously thin. |
| Bowl Season | Full-season grade. The first single-window experiment is judged. NCAA decides whether to keep the format or split the window again. |
“The new calendar punishes teams that needed two looks. The old one let those teams hide their first mistake. We are about to find out which group is bigger.”
— The Columnist, on the cost of one swingThe Unsigned
The 1,200 number deserves a closer look. These are not all marginal players. The portal in any year contains a top tier (returning starters seeking better opportunities), a middle tier (rotation players whose roles changed), and a bottom tier (depth pieces and young players who lost their position battle). Under the old system, the bottom tier disproportionately found homes in the spring window, when programs needed late depth. Under the new system, the bottom tier is the part that does not, on average, get signed at all. Some of those players will be back at their original schools. Many will not be playing FBS football in 2026.
This is not a tragedy. It is a structural change with predictable distributional consequences, and it will produce its own data. We will revisit this piece after the season — both to grade Notre Dame’s seven additions against their underlying-talent metrics, and to ask whether the NCAA’s consolidation move did what it was advertised to do, which was “bring stability to the sport.” The hypothesis under test is whether stability and one-window roster locks are the same thing. The early read, six months in, is that they are not.
