We take a look back at simulated preseason predictions for the first half of the Bills’ season
The Buffalo Bills are 6-2 and in firm command of the AFC East with a 3.5-game lead on their nearest competition in the Miami Dolphins. Given that we’re essentially at the midway point of the 2024 NFL season, it felt like a good opportunity to look back at preseason predictions AI (ChatGPT) made for each of the Bills’ first eight games, compare them to the real outcomes, and see how close they were to what actually happened.
To look back at AI’s original season predictions check out that article here. Below I break down ChatGPT’s predictions game by game, and provide grades for the accuracy.
Week 1: Bills vs. Cardinals
Prediction: Bills 27, Cardinals 17
Actual Result: Bills 34, Cardinals 28
Original Analysis: Expected a strong start from quarterback Josh Allen and solid defense to hold the Cardinals’ offense in check.
Grade: B
Explanation: The prediction was close in terms of Buffalo’s offensive output, but Arizona scored more than expected after the Bills got off to a slow start. Not quite as dominant a performance as anticipated.
Week 2: Bills at Dolphins
Prediction: Dolphins 31, Bills 28
Actual Result: Bills 31, Dolphins 10
Original Analysis: Projected a high-scoring game with Miami squeaking by Buffalo due to their offensive firepower.
Grade: D
Explanation: AI got this one wrong, though it did pick one team to score 31 points — it was just the wrong one. The Bills dominated this entire game and the Dolphins couldn’t establish anything on offense, especially after quarterback Tua Tagovailoa went down with an injury. Though this wasn’t the anticipated shootout, Buffalo easily handled business.
Week 3: Bills vs. Jaguars
Prediction: Bills 24, Jaguars 21
Actual Result: Bills 47, Jaguars 10
Original Analysis: A close game was anticipated, with nickel cornerback Taron Johnson making a pivotal defensive play to secure the win.
Grade: C
Explanation: While the Bills did win, this wasn’t the close matchup expected — rather a blowout for the ages ensued. Buffalo’s offense was far more explosive than predicted, and while Johnson’s standout play didn’t happen because he didn’t suit up for this game due to injury, the defense held up superbly.
Week 4: Bills at Ravens
Prediction: Ravens 30, Bills 24
Actual Result: Ravens 35, Bills 10
Original Analysis: Expected a challenging matchup, with Baltimore’s defense containing Buffalo’s offense, while quarterback Lamar Jackson’s dual-threat play proved decisive.
Grade: B+
Explanation: This was a largely accurate forecast, capturing the Bills’ struggle against the Ravens’ strong defense and Jackson’s impact.
Week 5: Bills at Texans
Prediction: Bills 34, Texans 21
Actual Result: Texans 23, Bills 20
Original Analysis: Expected Buffalo to bounce back with a strong game on offense, featuring Allen and multiple defensive turnovers.
Grade: C-
Explanation: The outcome was off, as the Texans emerged with the win on a last-second field goal. The game was closer than AI’s prediction, and it overestimated the Bills’ offense. Buffalo didn’t find its rhythm for the second game in a row, struggling against Houston’s defense.
Week 6: Bills at Jets
Prediction: Bills 27, Jets 24
Actual Result: Bills 23, Jets 20
Original Analysis: Expected a close divisional matchup, with Buffalo narrowly winning thanks to defensive pressure on quarterback Aaron Rodgers.
Grade: A
Explanation: This prediction was spot-on in both score and game dynamics. The Bills pulled out a win in a close game where their defense managed to disrupt Rodgers, resulting in an accurate call.
Week 7: Bills vs. Titans
Prediction: Bills 28, Titans 13
Actual Result: Bills 34, Titans 10
Original Analysis: Predicted a controlled game by Buffalo, with a strong showing from running back James Cook and a defensive clampdown on Tennessee’s passing game.
Grade: A+
Explanation: An almost perfect prediction here. The Bills’ offense performed slightly above the score prediction, while the defense controlled the Titans exactly as anticipated. James Cook didn’t break 100 rushing yards because he was held out due to injury. Instead, rookie running back Ray Davis had a fantastic game. The score and game flow of this prediction were very close.
Week 8: Bills at Seahawks
Prediction: Seahawks 30, Bills 27
Actual Result: Bills 31, Seahawks 10
Original Analysis: Expected Seattle to leverage home-field advantage, winning a close game.
Grade: D
Explanation: The actual result was quite different. Buffalo’s offense and defense both excelled, while Seattle’s offense struggled far more than anticipated. The Bills controlled the game from start to finish, and Bills Mafia nullified any home-field advantage the Seahawks had.
Overall midseason simulation grade: B
ChatGPT predicted a couple of games pretty close to the actual outcome, like the Jets and Titans games. However, it missed severely on others like the Dolphins and Seahawks results.
Overall, AI predicted the Bills to have a 5-3 record, which was off by only one game. It’s a bit uncanny how close to reality the predictions were. We’ll revisit the overall simulated predictions once the season finishes.