Multi-Dimensional Input
Beyond xG, Momentum and Tempo analyses, contextual variables like scoreline, minute, half boundary, league character and team form history are evaluated together.
Our Toprak model's central decision component. Combines all in-match signals, weights them contextually, and continuously learns — triggers live goal signals and feeds match insights.
MFT Composite is how Toprak's AI engine produces a single decision score from all in-match data. Not just an average — a contextual reasoning, continuous learning, and dynamic decision component calibrated with thousands of historical matches.
Shots exist but pressure isn't sustained — xG confidence low, MFT more conservative.
Sustained pressure + controlled rhythm — goal is imminent, MFT more aggressive.
Passive match, all signals weak — MFT gives no signal.
How does Toprak's central decision engine work?
Beyond xG, Momentum and Tempo analyses, contextual variables like scoreline, minute, half boundary, league character and team form history are evaluated together.
How much to trust each signal? Our Toprak model applies dynamic priority based on match state — when data is missing, remaining inputs receive automatic weight reallocation.
Trained on thousands of historical match data. Every completed match outcome is used as feedback to calibrate weights and thresholds. The model becomes smarter over time.
MFT Composite is not a standalone indicator — it is the main threshold that triggers our Live Goal Signals and the core input feeding our Pre-Match Insights. The heart of the entire Toprak product ecosystem.
Central role in the Toprak product ecosystem — 3 different output paths.
When the high MFT threshold is exceeded, the system automatically publishes a goal signal. MFT trend direction and confidence determine signal quality.
Our Toprak 1.5 model's pre-match predictions are based on MFT pattern analysis. Similar historical matches' MFT trend → confidence score.
Detailed MFT trend, component breakdown, contextual analysis and in-match snapshot history are available to Premium members.
6 dimensions output by the Toprak engine — transparent data layer.
Overall decision score. Toprak's intelligent rating for the current match state.
Contextual contribution of xG data to MFT. Varies according to match state.
Contribution of momentum data to MFT. Reflects pressure state.
Contribution of tempo data to MFT. Reflects match rhythm.
Toprak model's total confidence in its own forecast. 80+ = solid data.
Number of inputs currently used (xG / Momentum / Tempo). Auto-adapts when data is missing.
Auto-selected: highest MFT contrast finished match from the last 7 days.
Toprak interpretation: Both teams produced similar MFT scores in this match.
Clear MFT gap — Toprak model produced a definitive decision.