AI prediction · locked November 5, 2026 8:05 pm
Model pick 1–0
Result after 90 minutes: 1–0. The model tipped 1–0 and scored 3 points.
Most likely scores
Why
The model expects PSG to score 1.60 goals: league home average 1.70 × PSG home attack 1.07 × Troyes away defence 0.88.
It expects Troyes to score 0.87 goals: league away average 1.33 × Troyes away attack 0.72 × PSG home defence 0.91.
Based on 3 home games for PSG, 4 away games for Troyes and 81 finished matches in the competition.
Attack above 1.00 means scoring more than the league average; defence above 1.00 means conceding more than average (weaker).
Most likely outcome: PSG (55%). This is an estimate, not a certainty.
Form & stats this competition season
| PSG | Troyes | |
|---|---|---|
| Table position | #1 · 23 pts | #17 · 8 pts |
| Last 5 (newest first) | WWWWW | LWDLL |
| Home W-D-L (goals) | 3-0-1 (7:3) | 1-1-3 (5:11) |
| Away W-D-L (goals) | 4-2-0 (15:7) | 1-1-3 (2:6) |
| Goals per game at home (for/against) | 1.75 / 0.75 | 1.00 / 2.20 |
| Goals per game away (for/against) | 2.50 / 1.17 | 0.40 / 1.20 |
| xG | n/a | n/a |
xG is not available from the current data source (football-data.org). Competition averages: 1.63 home and 1.32 away goals per match (90 finished matches).
Model assumptions & limitations
Assumptions
- Goals for each team follow independent Poisson distributions.
- Expected goals (λ) = league average home/away goals × attack strength × opponent defence strength.
- Strengths use only this competition and season: home form for the home side, away form for the away side. If either team has too few home/away games, overall form (all games) is used instead and this is stated in the explanation.
- Strengths are shrunk towards the league average when few games have been played.
- Predictions are for the score after regular time (90 minutes plus stoppage time).
- The model's tipped score is the one with the highest expected points under the game's scoring system, not necessarily the most likely single score.
Limitations
- No injuries, suspensions, line-ups, transfers, rotation or motivation are considered.
- No xG or shot data is used – the current data source does not provide it.
- Results in other competitions (e.g. domestic league for Champions League teams) are not included.
- Independent Poisson models tend to slightly underestimate draws.
- Probabilities are estimates with considerable uncertainty, not facts or guarantees.