Hockey results Friday 30 October 2026: AI predictions review

Friday's NHL slate produced eight games and a modest return for the machines. The ten-model consensus finished the night 4 of 8, a coin-flip record that flatters the process less than it first appears: three of the four misses came on calls backed by every model in the panel.

The standout surprise was in St. Louis, where 80 per cent of the models had Boston winning and the Blues instead ran out 4-1 winners. That was not even the most emphatic misread. Minnesota and Dallas both went down despite 100 per cent consensus support, and Anaheim's 6-7 home defeat to Buffalo added a fourth blemish to the card.

Where the AI got it right

The panel's clearest hit came in Nashville, where Washington's 4-3 win was called by 60 per cent of models with an average confidence of 56 per cent. The most common scoreline projected was 3-4, which is exactly how the night ended. A narrow majority, correctly read.

Los Angeles delivered for the unanimous crowd. All ten models backed the Kings, and the 6-5 win over Ottawa — far wilder than the projected 3-2 — still landed on the right side of the result. Average confidence was 58 per cent, the sort of figure that suggests the models saw a close game and got one.

San Jose's 3-2 win over Vancouver was a 60 per cent call at 55 per cent average confidence, another case of a split panel landing on the correct outcome. The most common score was 4-3, so the margin was tighter than projected, but the verdict held.

Florida provided the fourth correct call and the highest confidence of the night's winners. All ten models picked the Panthers at an average confidence of 60 per cent, and a 5-2 win over Winnipeg made that the most comfortable correct read on the card. See the full Florida Panthers vs Winnipeg Jets prediction for the pre-match reasoning.

Where it got it wrong

St. Louis was the headline miss. Boston drew 80 per cent of the panel and a projected 2-4 scoreline at 57 per cent average confidence, and the Blues won 4-1. The models were not merely wrong on the winner; they were wrong on the scale of it. Full detail sits on the St. Louis Blues vs Boston Bruins prediction page.

Minnesota was unanimous and wrong. Ten of ten models backed the Wild at 56 per cent average confidence with a 3-2 projection; the New York Islanders won 4-3. A unanimous panel carries no extra weight when the result goes the other way, and this one did.

Dallas was the most expensive error of the night. Every model picked the Stars, the most common score was 4-2, and average confidence was 59 per cent. Montreal won 5-2. That is a three-goal swing against the consensus and the clearest single failure on the card.

Anaheim completed the set of misses. Sixty per cent of models backed the Ducks at 55 per cent average confidence with a 4-3 projection, and Buffalo took a 7-6 goalfest. The panel read the winner incorrectly and undersold the scoring by a wide margin.

Models of the day

Three models shared the top of the table. GPT-6 Luna, MiMo V2.6 Pro and Kimi K3 each finished 5 of 8, one clear of the consensus and the best individual returns of the day. Claude Sonnet 5.5, Mistral Medium 3.5 and Qwen 3.8 Max each matched the panel at 4 of 8.

Below the line, DeepSeek V4 Pro, Grok 4.7 and Gemini 3.8 Flash each managed 3 of 8. GLM 5.3 had the roughest full slate at 2 of 8. Meta Muse Spark 1.3 recorded 1 of 1, a perfect return on a single fixture that says little about its overall form.

Full results

• NHL: St. Louis Blues 4–1 Boston Bruins — consensus Boston Bruins win (80%) WRONG
• NHL: Nashville Predators 3–4 Washington Capitals — consensus Washington Capitals win (60%) RIGHT
• NHL: Minnesota Wild 3–4 New York Islanders — consensus Minnesota Wild win (100%) WRONG
• NHL: Dallas Stars 2–5 Montreal Canadiens — consensus Dallas Stars win (100%) WRONG
• NHL: Anaheim Ducks 6–7 Buffalo Sabres — consensus Anaheim Ducks win (60%) WRONG
• NHL: Los Angeles Kings 6–5 Ottawa Senators — consensus Los Angeles Kings win (100%) RIGHT
• NHL: San Jose Sharks 3–2 Vancouver Canucks — consensus San Jose Sharks win (60%) RIGHT
• NHL: Florida Panthers 5–2 Winnipeg Jets — consensus Florida Panthers win (100%) RIGHT

The lesson from Friday is straightforward: unanimity was not a signal. Three of the four unanimous calls landed, but the one that did not — Dallas — was the night's heaviest defeat relative to expectation. The panel's split verdicts went 1 of 3, so the edge, such as it was, came from the games where the models agreed. Saturday's card is already processed; the full set of calls is on the today's hockey predictions page.

Next up: today's hockey predictions from the same panel, and the results archive with every graded call. Forecasts are informational only.