The Table Says One Thing, the Model Another
Serie A's sixth round brings a fixture where the standings and the badge point in different directions. Cagliari arrive fourth on 12 points from five games, with four wins and one defeat; Juventus sit seventh on 10 points from five, with three wins, a draw and a loss. At the Unipol Domus, that gap is the storyline. Nutmegly's AI read, generated on 14 September, leans home: 47.7 percent for a Cagliari win, 24.5 percent for a draw and 27.8 percent for Juventus. That is a lean, not a guarantee — fewer than half the model's probability mass lands on the home side, and the away win is not far behind the draw. The most interesting tension is not that the model likes Cagliari; it is that the model likes Cagliari only moderately, against a Juventus side that has scored more goals in the stored sample.
Two Very Different Five-Game Sketches
Cagliari's stored results are tight and low-scoring: 1-0 at Udinese, 2-1 at Atalanta, 1-0 at home to Lecce, a 0-1 home defeat to Inter, and 1-0 at Parma. Five goals scored, two conceded. Three of the four wins came by a single goal, and the only defeat was also by one. Juventus' five stored results read louder: 2-0 at home to Atalanta, a 2-3 defeat at Sassuolo, a 1-1 home draw with AC Milan, 2-0 at home to Parma and 1-0 at Frosinone. Eight scored, four conceded. The away side has more goals in the sample; the home side has more points. The table's venue splits sharpen the contrast: Cagliari have played two home games, winning one and losing one, with one goal for and one against; Juventus have played two away games, winning one and losing one, with three for and three against.
Why the Model Sides With Cagliari
The stored-sample rates support the home lean. Across the eight most recent stored matches for each side, Cagliari average 2.00 points per game and Juventus 1.50; within the Serie A sample, the split is 2.40 against 1.60. Those are stored-sample averages and may include older seasons, so they are context rather than a current-season total. The venue split adds a second layer: across the stored same-venue sample, Cagliari average 1.20 goals per home game, while Juventus concede 1.00 per away game. The model also puts over 2.5 goals at 53.7 percent — a slight lean toward goals, though Cagliari's 1-0 habit sits awkwardly against it. Put together, the case is not that Cagliari blow Juventus away; it is that the home side's points rate, venue scoring rate and the away side's away concession rate all tilt the same way.
The Strongest Case Against the Read
Juventus' scoring volume is the clearest counterweight. Eight goals in five stored results is more than Cagliari's five, and the 2-0 win over Atalanta shows the away side can produce a clean, multi-goal afternoon. The model's read was generated on 14 September, before Cagliari's win at Udinese and Juventus' win over Atalanta, so it does not incorporate the most recent stored evidence. The data also carries no odds summary, no external forecast, no confirmed lineups and no injury reports, which means the result probabilities are uncalibrated against the market and against team news. The prediction's own counterpoint is that lineups have not fully landed and late tactical or personnel choices could still change the outcome picture. The 29-day and 28-day gaps since the last stored matches are not evidence of rest; the data notes explicitly say such gaps should not be read as an advantage, and a missing record does not prove a team has not played.
The Watchpoint
The match's shape may matter more than its winner. If Cagliari's low-scoring pattern holds, the game could turn on a single goal — the kind of 1-0 or 1-1 afternoon their stored results keep producing. If Juventus' higher goal volume translates, the over 2.5 lean gets its support and the home lean looks less comfortable. The specific thing to watch is whether the game stays in that one-goal band or breaks open early. The model's 47.7 percent home lean is the starting point, not the conclusion.