The table and the model disagree
Northern Ireland host Georgia at Windsor Park on Monday evening sitting top of their Nations League group with seven points from three games, two wins and a draw, four goals scored and none conceded. Georgia are bottom with a single point, no wins, and — the number that matters most here — no goals at all. On the raw standings this looks like a home side in control against a side that cannot score.
Nutmegly's AI read does not see it that way. The model leans draw, giving it 43.2 percent, with a Georgia win at 38.2 percent and a Northern Ireland win at just 18.6 percent. It also puts the chance of more than 2.5 goals at 8.7 percent, which is a forecast of a tight, low-scoring night rather than a comfortable home procession. That gap between the table and the model is the story of this fixture.
What Northern Ireland have actually built
The recent record is strong and, importantly, recent. Northern Ireland won 1-0 in Georgia on 25 September, drew 0-0 at home with Hungary on 28 September, then won 3-0 away in Ukraine on 2 October. Three group games, three clean sheets, and a goal difference of four. The two results before that window — a 1-3 defeat in France in June and a 0-1 home loss to Germany in October 2025 — belong to the stored historical sample rather than this campaign, and should be read that way.
There is one wrinkle in the home column. Northern Ireland's only stored home game in this competition is that 0-0 with Hungary, so the sample at Windsor Park contains no goals for them at all. One match is not a pattern, but it is the kind of thin evidence that keeps a draw lean alive.
Georgia's problem is at the wrong end
Georgia have played three group games and scored none, conceding twice. Their three most recent stored results read 0-1 in Hungary, 0-0 at home to Ukraine, and that 0-1 home defeat to Northern Ireland. Across the five stored results in the sample they have scored once, in a 1-4 loss in Türkiye in October 2025, with a 0-4 home defeat to Spain also in the file. The away sample is one competition match, a 0-1 loss.
That is the strongest argument against the model's 38.2 percent away win: a team that has not scored in the group is being given a better than one-in-three chance of winning away from home. The counterweight is that the model's own notes flag a sample of fewer than five recent matches for the home side, thin advanced-statistics coverage, no head-to-head data at all, and a direction that disagrees with an external forecast. It also predates the freshest evidence — it was generated on 24 September, before Northern Ireland's win in Georgia and before both sides' 2 October results.
The watchpoint
The conditional that decides the mood: if Georgia score first, the draw lean becomes genuinely live, because Northern Ireland's stored home sample in this competition is a goalless draw and the model already expects fewer than three goals. If Northern Ireland instead carry their away form home and score inside the opening half, the 18.6 percent home-win figure will look like the outlier.
Two caveats belong on the record. No lineups are confirmed, and no injury reports have been filed — which is not the same as two fully available squads, only an absence of information. With no odds available, the result probabilities are uncalibrated by any market. Treat the model's numbers as an estimate, not a result: the table says Northern Ireland, the model says draw, and Windsor Park gets to settle it.