One model. A clear view of our approach.

MatchPoints uses specialized models to predict different kinds of breakthroughs. This page showcases one example—and how its predictions compare with what happened.

Explore more modelsUnlock the platform to explore more specialized breakthrough models.

This example covers 1,855 athletes. The athletes and results shown here belong to this example model. We test across many more athletes and model populations throughout the platform. Today we serve a calibrated MPI for 2,437 athletes on the live board: 1,254 ATP and 1,183 WTA.

The evidence behind the Index

Every score on this page was tested against what already happened

We scored 1,855 ATP athletes ranked 101 to 800 at past points in their careers, then waited to see what happened.

Development evidence: these are the model’s development results, not an independent assessment.

The commercial consequence follows from the timing. Every player on the board is cheaper today than they will be if the Index is right about them.

The closer a dot sits to the line, the closer the index is to reality.

Compare predictions with what actually happened

Each dot compares predicted chances with actual results for its MPI range. Choose a dot or range below to see up to 30 real examples, selected to roughly match the summary group’s success rate; scores 70+ are combined in the summary. Times are approximate from selected weeks to recorded target weeks; a dash means unavailable.

MPI range

Loading the selected calibration interval and named outcomes.

Named outcomes unavailable for this selection.