Results
Hockey results Monday 9 November 2026: AI predictions review
Hockey results Monday 9 November 2026: the ten-model AI consensus went 2 of 3, with Calgary and Anaheim landing but Toronto's shutout loss a clean miss.
Three NHL fixtures, ten frontier models and one shared sheet of predictions made for a straightforward Monday night on paper. The consensus record finished at 2 of 3, with Calgary and Anaheim both delivered as called and Toronto the single blemish on an otherwise tidy card.
The surprise came in Ontario, where every one of the ten models backed the Maple Leafs and the Devils answered with a 4-0 shutout. It was the only genuine jolt of the night, and it arrived with the models sitting at an average confidence of 55%, a number that now reads as far too generous.

Full-time ice after a three-game NHL Monday that finished 2 of 3 for the ten-model consensus.
Where the AI got it right
The Calgary Flames vs Philadelphia Flyers prediction was the models' most assured call of the slate. All ten backed a Flames win at an average confidence of 59%, with a 4-2 scoreline the most common projection. Calgary delivered the win, though the Flyers made it a one-goal game at 4-3, so the result landed and the exact margin did not.
The Anaheim Ducks vs Pittsburgh Penguins prediction went the same way, a unanimous Ducks pick at 56% confidence and a 4-3 scoreline as the most popular forecast. Anaheim won by a wider margin than anyone projected at 5-2, meaning the models had the right team and underestimated the gap.
Two from two on the unanimous picks that held is the part of the night worth noting. The consensus was not cautious on either game, and it did not need to be.
Where it got it wrong
The Toronto Maple Leafs vs New Jersey Devils prediction was a clean miss. All ten models selected a Maple Leafs win, the most common score was 4-3, and the average confidence sat at 55%. New Jersey won 4-0. The models had the direction wrong and the margin wrong, and the confidence figure makes clear this was not a hedge but a genuine collective misread.
The lesson is the familiar one for a results desk: unanimity is not accuracy. Ten models agreeing on a home favourite is still ten models reading the same pre-match information, and when that information points the wrong way, the agreement simply amplifies the error. The Toronto result is the one entry on the card that will be remembered.
Models of the day
There was no separation at the top. All ten models finished on 2 of 3 correct: GPT-6 Luna, Claude Sonnet 5.5, DeepSeek V4 Pro, Qwen 3.8 Max, MiMo V2.6 Pro, Grok 4.7, Gemini 3.8 Flash, Kimi K3, GLM 5.3 and Meta Muse Spark 1.3. Every one of them backed Toronto, so every one of them carried the same loss.
The best and worst tallies are therefore identical, which is itself the story of the day. On a three-game slate with a unanimous wrong pick, the models had no room to differentiate themselves. A single contrarian read on the Devils would have produced a clear leader; none of the ten took it.
Full results
- NHL: Calgary Flames 4-3 Philadelphia Flyers — consensus Calgary Flames win (100%) RIGHT
- NHL: Toronto Maple Leafs 0-4 New Jersey Devils — consensus Toronto Maple Leafs win (100%) WRONG
- NHL: Anaheim Ducks 5-2 Pittsburgh Penguins — consensus Anaheim Ducks win (100%) RIGHT
The card closes at 2 of 3, with the Flames and Ducks calls holding and the Maple Leafs call falling flat in a 4-0 shutout. Today's fixtures and the next set of ten-model consensus picks are published on the hockey predictions for today page.
Next up: today's hockey predictions from the same panel, and the results archive with every graded call. Forecasts are informational only.