The Giants opened 2026 by beating Dallas 28-20 at home in week 1, then lost 28-6 at the Rams on Monday night in week 2. Jaxson Dart was hurt on his fifth dropback of the second game and did not return; Jameis Winston finished it. That's the whole factual story through two games. Everything else — where this team is headed, whether the quarterback is the ceiling, whether John Harbaugh's first year in East Rutherford is actually different from the Giants' recent baseline — is a range, and the honest version of the range is wide.

Table of Contents

1. What actually happened
2. Three baselines the season should be measured against
3. What the rest of 2026 plausibly looks like
4. Is the quarterback the binding constraint?

1. What actually happened

Through two games, Dart's two starts look nothing alike:

Dart's 5th dropback of the week-2 game is the one the play-by-play flags: "NYG-6-J.Dart was injured during the play."

Winston finished the game and had the workload the rest of the way:

At the team level the two weeks are equally divergent: NYG offense posted +0.369 EPA/play in week 1 (65 plays) and -0.429 EPA/play in week 2 (55 plays). The defense allowed +0.221 EPA/play in week 1 and +0.290 in week 2 — worse in both games than any full-season NYG defensive mark of the last four years (see below). That already tells you the week-2 collapse was not only about the quarterback change.

2. Three baselines the season should be measured against

Giants' own record, 2021-2025

Five seasons, one winning record, an average of about 5-12. That is the base rate any 2026 projection should be argued against, not away from.

John Harbaugh's Baltimore, same five years

Here is the comparison the angle points at, and I'll let it sit without editorializing: Harbaugh's worst Baltimore season in this window was 8-9. The Giants' best in the same window was 9-7-1. They overlap by roughly one game. The floor of one team is essentially the ceiling of the other over half a decade. Whether Harbaugh personally closes that gap in year one is exactly the kind of question two games cannot answer.

Jameis Winston, EPA per dropback by season

The reason to look at this by season, with dropbacks alongside, is that Winston's samples vary by an order of magnitude and his numbers vary with them:

The largest recent sample is 2024 in Cleveland: 327 dropbacks at -0.093 EPA/dropback. That is the fairer read of what he is as a passer over real work, and it is meaningfully below league-average starter quality. His Tampa Bay years are the large positive sample and they are also the old one: 2015-2019 is 2,733 dropbacks at +0.073 EPA per dropback, the most recent of them seven years ago. Everything from 2021 on — 773 dropbacks — is -0.051. His 2025 NYG cameo (+0.223 in 69 dropbacks) sits somewhere between "encouraging" and "small enough to not update on."

If Dart misses time, the honest expectation for Winston is closer to his 2024 season than to either his week-2 blowup or his 2025 cameo.

3. What the rest of 2026 plausibly looks like

Here are three scenarios with the assumption on each one exposed:

• If the week-2 offense is closer to the real one — i.e., NYG's -0.429 EPA/play on offense and +0.29 allowed on defense in that game are directionally correct — this team is worse than its 2024 (3-14) and 2025 (4-13) versions and finishes with the fewest wins of the Harbaugh/Giants baselines above. I don't think one game against a good Rams defense justifies that assumption on its own, but it's the low-end scenario and it exists.
• The blended, most-likely middle — Dart's true early-career level is somewhere between +0.707 and -0.223 EPA/dropback, and any extended Winston stretch pulls the offense toward his 2024 baseline of about -0.093 EPA/dropback over 327 dropbacks. That path lands the Giants inside the 4-13-to-9-7-1 band the franchise has actually lived in for five years, with the Harbaugh hire and Dart's ceiling as the reasons to bet on the top of that band rather than the bottom.

4. Is the quarterback the binding constraint?

Worth taking seriously, because the week-2 collapse was not confined to the passer. NYG's defense allowed +0.290 EPA/play in that game and +0.221 in the win over Dallas — 56 and 52 plays respectively. For context, here is the Giants' defensive EPA/play by full season:

Both 2026 games are worse than any of those four full-season marks — but that comparison flatters the point, because a single game swings far more than a season average. Measured game-to-game, which is the fair test: 41% of the Giants’ individual defensive games from 2022-2025 were worse than their worst season average, so clearing that bar means little. The Week 1 performance (+0.221) was worse than 87% of those games, and Week 2 (+0.290) worse than 89%. That is the honest version: two poor games, in the worst decile of recent Giants defense, on 108 combined plays. It is enough to say the Week 2 collapse was not confined to the quarterback. It is not enough, on its own, to say no quarterback could save this roster.

My read, stated as opinion: the QB is a constraint, not clearly the constraint. Dart's week-1 tape is the most encouraging Giants passing performance in the sample I pulled; Winston's 2024 body of work is the realistic downside. But the roster around them, particularly on defense in the games played, has not looked like a group one quarterback change lifts to Harbaugh's Baltimore range. Two games is early. It is also what we have.