The table says Leeds; the model says Bournemouth
The most interesting thing about Bournemouth vs Leeds is not the league table. Leeds sit seventh on 11 points, Bournemouth 13th on nine. Leeds also have the better recent points-per-game rate in the stored sample: 1.38 from eight matches against Bournemouth's 1.12. Yet Nutmegly's AI read makes Bournemouth the lean, with a 45.3% home-win probability. That gap between standings and projection is the storyline.
This is the ninth-round Premier League fixture at Vitality Stadium, and it pits a host coming off two strong results against a visitor whose season has been built on draws. Bournemouth beat Manchester United 3-0 away and Sunderland 2-0 at home in their last two recorded matches. Leeds have drawn five of their eight league games, including 0-0 at Sunderland and 1-1 at home to Manchester United in their last two. The contrast is sharp: Bournemouth's recent run has higher peaks; Leeds' has more stability.
Bournemouth's home case is narrower than the last two scores suggest
The case for the home lean starts with Bournemouth's current-season record: 12 goals scored and 13 conceded in eight games, with a home split of one win, two draws and one defeat, five goals for and four against. Their home average of 1.20 goals per game meets a Leeds away average of 1.20 conceded per game. In the Premier League sample, Bournemouth have taken 1.40 points per game against Leeds' 1.20. Those are small edges, but they point in the same direction as the model.
The caveat is that Bournemouth's last two results sit beside a 1-5 defeat at Chelsea and a 0-1 home loss to Liverpool, plus a 2-2 home draw with Brentford. That is not a smooth upward curve. The two wins are real and recent, but they are part of a five-match run that also includes two defeats and a draw. The model's home lean should be read as a marginal edge, not a verdict on form.
Leeds' draw habit is the strongest counterargument
Leeds' strongest counterevidence is not just their league position. It is the shape of their results. They have two wins, five draws and one defeat in eight league games, with nine goals scored and six conceded. On the road they are one win, two draws and one defeat, three goals for and three against. In the recent sample they have drawn 0-0 at Sunderland, drawn 1-1 with Manchester United, lost 1-2 at Arsenal, drawn 0-0 with Crystal Palace and beaten Newcastle 4-1. That is a team that has been hard to separate from opponents, not one that has been swept aside.
The home/away splits are almost mirror images: both teams have one win, two draws and one defeat in four games in the relevant role. Bournemouth's home goals are 5-4; Leeds' away goals are 3-3. If the match follows those splits, a draw is not a niche outcome. The model gives it 27.9%, while Leeds' away win is 26.9%. The home win is 45.3%, so the leading outcome is not a majority. That matters for how the preview should be framed: Bournemouth are favourites, not certainties.
What the AI read is weighing, and what it cannot see
Nutmegly's AI read gives Bournemouth 45.3%, the draw 27.9% and Leeds 26.9%. It also puts over 2.5 goals at 49.1%, which is close to a coin flip. The model's home lean is supported by the same-competition points-per-game split, the home attack versus Leeds' away goals-against average, and the one-day difference in the recorded gap between matches. But the data also carries clear warnings. An external forecast points in a different direction from the model, and the two are not using identical samples. No lineups are confirmed for either side. Seven potential absentees are reported, without names or roles in the data. Those are real limits on any pre-match read.
The injury and lineup note should not be overread. The data does not say who is missing, whether they will definitely miss out, or how either side will replace them. It does say the starting lineups are not confirmed. That means the model's probabilities describe the match before the final selection is known. If the team sheets change the balance of the sides, the pre-match edge could look different by kickoff. The correct conclusion is not that Bournemouth's advantage disappears; it is that the advantage is conditional.
The watchpoint: can Bournemouth turn peaks into control?
The match question is whether Bournemouth can carry the scoring output of their last two recorded results into a home game against a Leeds side that has drawn five of eight. If they can, the 45.3% home lean has a clear path: Bournemouth's home goals per game and Leeds' away goals-against per game both sit at 1.20, and the hosts have the better same-competition points rate. If they cannot, Leeds' draw pattern and their 3-3 away split make a tight, low-scoring game the most plausible alternative. The over 2.5 probability of 49.1% does not promise goals.
The specific watchpoint is the first team sheet, not the first shot. With no confirmed lineups and seven potential absences reported, the final selection is the biggest variable the model could not settle. After that, watch whether Bournemouth's recent wins look like a repeatable home performance or a short spike. Leeds have already shown they can hold a 0-0 and a 1-1; Bournemouth have shown they can win 3-0 and 2-0. The question to watch: can Bournemouth turn those two clean-sheet wins into a home performance, or will Leeds' draw habit pull the game back toward the middle?