AI NFL Predictions · Week 11 Daggy vs. the Market

We’re testing our AI forecast agents against the hardest possible benchmark: a real-time, liquid market. Before each game, we post our predictions along with those from Kalshi’s prediction market and score them as they settle. Why we’re doing this.

Through 150 graded games this season, Daggy's combined AI forecast has picked the winner 98 times and missed 52 (65.3%), with a Brier score of 0.219. Kalshi's prediction market, the benchmark to beat, scores 0.216, so the market is slightly ahead. Lower Brier is better: 0 is a perfect forecast and 0.25 is calling every game 50-50. Week 11 predictions for every game → · Week 10 recap →

Daggy, drawn as the saloon's robot bartender
Daggy
W-L Record
98-52
65.3%
Brier score0.219
VS
Kalshi, drawn as the saloon's cash register — the market price
Kalshi
W-L Record
100-50
66.7%
Brier score0.216
00.25
Daggy 0.219Kalshi 0.216
Lower Brier = better forecasts. 0 is a crystal ball, 0.25 is a coin flip.

Week by Week

WeekMethodDaggy W-LKalshi W-LDaggy BrierKalshi Brier
Supervisor9-59-50.1940.196
Supervisor14-114-10.1250.117
Supervisor9-510-40.1880.177
Supervisor8-69-50.2490.242
Supervisor7-79-50.2670.267
Supervisor8-79-60.2400.239
Supervisor11-59-70.2140.223
Supervisor9-78-80.2430.258
Extremized11-511-50.2480.231
Mean12-412-40.2160.211
SeasonOur call98-52100-500.2190.216
Pick a week to see its games below.
Weeks 9–13

Week 11

INDatHOU
THU, NOV 19 · 8:15 PM ET · Week 11
Daggy picks
HOU
70%
Kalshi prices
HOU
84¢
MIAatBUF
SUN, NOV 22 · 1:00 PM ET · Week 11
Daggy picks
BUF
88%
Kalshi prices
BUF
92¢
BALatCAR
SUN, NOV 22 · 1:00 PM ET · Week 11
Daggy picks
CAR
59%
Kalshi prices
CAR
58¢
NOatCHI
SUN, NOV 22 · 1:00 PM ET · Week 11
Daggy picks
CHI
79%
Kalshi prices
CHI
85¢
TENatDAL
SUN, NOV 22 · 1:00 PM ET · Week 11
Daggy picks
DAL
75%
Kalshi prices
DAL
77¢
TBatDET
SUN, NOV 22 · 1:00 PM ET · Week 11
Daggy picks
DET
69%
Kalshi prices
DET
72¢
ARIatKC
SUN, NOV 22 · 1:00 PM ET · Week 11
Daggy picks
KC
86%
Kalshi prices
KC
92¢
JAXatNYG
SUN, NOV 22 · 1:00 PM ET · Week 11
Daggy picks
JAX
70%
Kalshi prices
JAX
70¢
NYJatLAC
SUN, NOV 22 · 4:05 PM ET · Week 11
Daggy picks
LAC
57%
Kalshi prices
LAC
61¢
LVatDEN
SUN, NOV 22 · 4:25 PM ET · Week 11
Daggy picks
DEN
80%
Kalshi prices
DEN
83¢
PITatPHI
SUN, NOV 22 · 4:25 PM ET · Week 11
Daggy picks
PIT
70%
Kalshi prices
PIT
67¢
MINatSF
SUN, NOV 22 · 8:20 PM ET · Week 11
Daggy picks
SF
59%
Kalshi prices
SF
61¢
CINatWAS
MON, NOV 23 · 8:15 PM ET · Week 11
Daggy picks
CIN
73%
Kalshi prices
CIN
72¢

Agent Breakdown

Over 150 graded games, Moonshot's Kimi with Daggy's statistical engine has the best (lowest) Brier score, 0.212, ahead of Kalshi's 0.216. Then Google's Gemini at 0.215, OpenAI's GPT at 0.217, and Anthropic's Claude at 0.217. Without the engine, the same models score 0.218 to 0.226. Averaged into one forecast, the agents with Daggy's engine scored 0.215 and the same models without it 0.220 (Kalshi 0.216); all four providers did better with it.

Models against Kalshi, over time

Click a provider to compare its engine-on line (solid) with its engine-off line (dashed). You can also click a line or its name.

Kalshi’s Brier minus the agent’s, summed over every game so far · above zero = beating Kalshi−1.00Kalshi+1.00Week 1Week 2Week 3new models · market-redacted from hereWeek 4Week 5Week 6Week 7Week 8Week 9Week 10Moonshot · engine on +.65Google · engine on +.16OpenAI · engine on −.11Anthropic · engine on −.14OpenAI · engine off −.18Google · engine off −.49Moonshot · engine off −.85Anthropic · engine off −1.47
Daggy engine on Engine offColor is the provider. Kalshi is the zero line.

Each line is a running total: game by game, it adds how much lower (up) or higher (down) that agent’s Brier was than Kalshi’s on the same game, so a rising line is beating the market. For example, Moonshot · engine on ends at +.65 over 150 games: +.004 per game, the gap between its season Brier and Kalshi’s in the table below. Click any Brier there to see the game-by-game sum. Totals run all season per provider; the week-3 rule marks the model upgrade.

Season and week by week

Engine on means the agent ran with the Daggy stats engine; engine off is the same model without it. Each cell shows Brier (lower is better) over W–L; a forecast of exactly 50% counts as picking the home team. Kalshi’s row is the market to beat, and the best Brier in each column is bold. Click a column to sort, a provider to highlight it on the chart, or a Brier to see its games.

Market-redactedMarket-aware
market-redactedmarket-redactedmarket-redactedmarket-redactedmarket-redactedmarket-redactedmarket-redactedmarket-redactedmarket-awaremarket-aware
Kalshi100–509–514–110–49–59–59–69–78–811–512–4
Average of 4
Engine on100–509–514–110–49–57–78–711–59–711–512–4
Engine off100–5010–414–110–47–77–79–611–59–711–512–4
Moonshot
Engine on98–529–514–111–37–76–88–711–510–610–612–4
Engine off99–5110–414–110–47–78–68–711–59–710–612–4
Google
Engine on100–509–514–110–48–67–78–711–510–611–512–4
Engine off98–5210–414–110–47–77–79–69–79–711–512–4
OpenAI
Engine on99–519–514–110–49–58–68–79–710–611–511–5
Engine off97–539–514–18–67–77–79–611–59–711–512–4
Anthropic
Engine on98–529–514–111–37–77–78–710–69–711–512–4
Engine off96–5410–414–19–58–66–89–69–78–811–512–4

Questions About AI NFL Predictions

How accurate are AI NFL predictions?

Through 150 graded games this season, Daggy's combined AI forecast has picked the winner 98 times and missed 52 (65.3%), with a Brier score of 0.219. Kalshi's prediction market, the benchmark to beat, scores 0.216, so the market is slightly ahead. Lower Brier is better: 0 is a perfect forecast and 0.25 is calling every game 50-50.

Which AI model predicts NFL games best?

Over 150 graded games, Moonshot's Kimi with Daggy's statistical engine has the best (lowest) Brier score, 0.212, ahead of Kalshi's 0.216. Then Google's Gemini at 0.215, OpenAI's GPT at 0.217, and Anthropic's Claude at 0.217. Without the engine, the same models score 0.218 to 0.226.

Does a statistics engine make AI forecasters better?

Averaged into one forecast, the agents with Daggy's engine scored 0.215 and the same models without it 0.220 (Kalshi 0.216); all four providers did better with it.

Can AI beat a prediction market like Kalshi?

Two of the four agents with Daggy's engine score better than Kalshi's 0.216 this season: Kimi 0.212 and Gemini 0.215. Daggy's combined forecast, at 0.219, is slightly behind.

How This Whole Racket Works

Why We're Running This Game

Superforecasters typically turn qualitative signals into probabilities via judgment. In a football game, these signals could be player-coach friction. But probabilistic modeling and machine learning generally beat qualitative judgment, especially when we have a well-defined quantitative target. This page documents our journey to build an AI forecasting system that mixes the judgment of frontier LLMs with Daggy’s statistical capabilities.

Every forecast is posted before kickoff and scored in public, so if the approach doesn’t work, it shows here first.

1

The Two Players

Daggy is not one model. It's 8 agents forecasting every game — 4 of them wired to the Daggy statistical engine, the rest working from what they can reason out — combined into a single pick per game, before kickoff.

Kalshi isn't a rival forecaster. It's the live price real money is trading at, on a market where anyone can bet on the outcome. When we say “Daggy vs. Kalshi,” we mean model vs. market — not two people guessing.

Claude Opus 5.5 · GPT-6 Sol · Gemini 3.8 Flash · Kimi K3
2

The Score

In addition to win-loss, we use the Brier score, which judges based on confidence:

Said 90%, team won
0.01
Said 55%, team won
0.20
Said 55%, team lost
0.30
Said 90%, team lost
0.81
Lower Brier score = better forecasts. 0 is a crystal ball, 0.25 is a coin flip.
3

Aggregation Methods

We have 8 agents predict each game, but what's left is to aggregate all 8 predictions to one forecast. We use the following methods, but commit to one before each week.

  • Mean: A simple average of each agent's prediction
  • Extremized: Agents often overly hedge their bets toward 50/50. Extremizing pushes the predictions toward zero-one by using Platt scaling
  • Supervisor: An AI agent looks at all 8 predictions, thinks them through and then comes up with a final forecast.

For information and entertainment only. Not betting, trading, or investment advice.