The Giants are 3-1 with Jameis Winston throwing three touchdowns against Arizona, and our model lifted New York to eight wins with the playoffs in view
Updated
The Giants beat Arizona 36-24 at home, their third win in four games, and did it with Jameis Winston throwing three touchdown passes and Deonte Banks returning an interception for the score that sealed it. New York is 3-1, its best start in years, and the model has moved. When we ran the rest of the season the Giants came out at 8-9 with a 16.5 percent playoff chance, up from 6.2 wins and two and a half percent after Week 3 and nearly double the 4.7 wins we projected in July. A win and a half in a week is a large move for a team the model has graded conservatively all year, and the reason is simple: the Giants keep winning the games the model gives them a coin flip in, and the defense keeps being better than its grade.
The quarterback situation is settled for now and unsettled for later. Jaxson Dart is out for the rest of the regular season after knee surgery. J.J. McCarthy arrived from Minnesota for a 2027 fifth-round pick and is the backup for now, with Winston keeping the job after the Arizona game. Winston is 43 of 78 for 479 yards with three touchdowns and three interceptions across his three appearances, a 70 rating, and the Arizona game accounted for all three scores and two of the picks. Our model projects him for 3,409 yards and 16 touchdowns against 13 interceptions behind an 83 rating, up from 13 touchdowns and 81 a week ago, and the model has not priced McCarthy as a starter. Malik Nabers has 18 catches for 208 yards and projects for 84 catches and 1,139 yards. Isaiah Likely has two touchdowns on 22 catches and projects for 62. Cam Skattebo has 235 rushing yards and projects for 696. The offense projects for 305 points, 26th in the league, up from 280.
The defense is the reason the Giants are 3-1 and the model has been upgrading it weekly. New York projects to allow 362 points, 20th in the league, up from 387 and 24th a week ago, and the gap between that and the 79 points actually allowed in four games is where the next move will come from if it comes. Jevon Holland has three interceptions and projects for nearly five. Kayvon Thibodeaux projects for seven and a half sacks and Abdul Carter for seven. DJ Reader for six from the interior. Arvell Reese's interception projection has gone from near zero to two as the linebacker has taken on coverage duties, and he projects for 113 tackles. The team projects for 36 sacks, 24th, and 13 interceptions, ninth. The unit held Dallas to 20, Tennessee to seven and Arizona to 24 with a pick-six, and the Rams game is the only one it has been beaten in.
The division is a four-team race and the Giants are a real part of it. Dallas is 2-2 and projects for ten wins, with the head-to-head already lost to New York. Philadelphia is 2-2 and projects for nine. Washington is 1-3 and projects for four and a half. The Giants lead the NFC East at 3-1 and project to finish third in it. When we ran the rest of the season New York won the division in a small minority of simulated years and finished third most often, which is still two spots higher than the model had it in July.
The next month is where the projection gets its answer. A trip to Washington at 60 percent, New Orleans at home at 56, a trip to Houston at 34, a bye in Week 8, and a trip to Philadelphia at 22. When we ran that stretch the Giants went 2-2 more often than anything else, and the Washington and New Orleans games are the two the model expects them to win. A 5-3 record entering November would push the projection to nine and make the 16.5 percent playoff chance look conservative.
The player lines are a quarterback the model has repriced and a defense it keeps upgrading. Winston's projection climbed from 3,295 yards to 3,409 and from 13 touchdowns to 16, with his interception line up from eleven to 13 and his rushing projection from 92 yards to 176, which is the model noticing the scrambles. Dart's remaining line is the 250 yards and three touchdowns he had. Nabers' projection rose from 82 catches to 84 and from 1,053 yards to 1,139. Likely's climbed from 55 catches to 62. Fields' from 50 to 49 with the yardage up. Skattebo's rushing projection dropped from 781 to 696, and Najee Harris' from 267 to 240. On defense, Thibodeaux's sack projection rose from 7.3 to 7.6, Carter's from 6.7 to 7.1, and Holland's interception line dropped from 4.8 to 4.6. Reese's interception projection went from near zero to 2.1 and his tackle line from 111 to 113. The number to watch is McCarthy. The model has him at zero attempts for the Giants. The day he starts, the eight-win projection gets rerun from scratch, and the direction depends on a quarterback the model has not seen in this offense.
The matchup record is one of the four largest overperformances in the league. We gave New York 20 percent against Dallas and the Giants won by eight. We gave them four percent in Los Angeles and they lost by 22. We gave them 54 against Tennessee and they won by five. We gave them 52 against Arizona and they won by twelve. Three wins from about one and a third expected is the gap, and the Cowboys game is the one that started it. The market opened New York around six wins. Our model sat well under it in July and sits two over it now. The lean is the defense, which has allowed 79 points in a month, and a quarterback situation the model will reprice the moment McCarthy starts. Three wins in four weeks earned eight. The next five opponents will say whether it was earned for keeps.
