From data to a prediction
What comes out
4 min
FinalSkore turns those inputs into a few related outputs. Each answers a different question, so it helps to keep them straight.
The projected score
The model produces a projected final score — its best estimate of where the game lands if it played out from the current state. As the game progresses and uncertainty shrinks, that projection naturally tightens toward the real result.
The win probability
It also produces a win probability — the chance each team has of winning from here. This is the number most people misread, so be careful: a 70% win probability does not mean the team is "going to win." It means that, across many games in this exact situation, the team would win about 70 times out of 100. The other 30 are real.
Value versus the line
Finally, FinalSkore compares its own projection to the bookmaker line to flag potential value — moments where it thinks the market has mispriced a side. A pick isn't surfaced just because a team is likely to win; it's surfaced when the model's view and the odds disagree in a way that could favor the bettor.
Putting it together
- Projected score — where the model thinks the game ends.
- Win probability — how confident it is, expressed as a percentage.
- Value — whether the odds make that confidence worth acting on.
These three can point in different directions, and that's normal — the next chapter is all about reading them honestly.