Metrics and semantics¶
The recommended graph-based method is Discounted Future-Fate Propagation (DFFP). The public field names remain explicit, while concise terms are used throughout the documentation.
Conditional Fate Affinity¶
For supervised fate probabilities \(p_{if}\), Conditional Fate Affinity (CFA) is
It describes fate identity conditional on reaching a supervised fate. It is
exposed as future_fate_affinity and conditional_fate_affinity.
Discounted Fate Reach¶
Discounted Fate Reach (DFR) is exposed as future_fate_reach and
discounted_fate_reach. It is horizon-dependent and is interpreted as
time-bounded resolution or endpoint reach, not universal physical time.
Future-Fate Specificity¶
Future-Fate Specificity (FFS) is zero for a uniform fate distribution and one for a point mass.
Resolved Commitment¶
Resolved Commitment (RC) combines reach and specificity. It is exposed as
reach_supported_specificity and resolved_commitment. Always report DFR
and FFS separately alongside RC.
Unresolved Future Probability¶
Unresolved Future Probability (UFP) is the chance that geometric stopping occurs before any supplied selected or competing anchor is reached. It prevents unlimited random-walk mixing from assigning every cell to an outcome.
Signed Ordering Flux¶
Signed Ordering Flux (SOF) is the expected one-step change in the supplied
ordering, conditioned on transitions retained in the selected furcation. It is
exposed as signed_progression and signed_ordering_flux. Positive values
are forward, negative values are retrograde, and values near zero may indicate
stationary, mixed, turning, or loop-like motion.
Compatibility names¶
The established result properties remain available. In future-fate mode:
Established property |
DFFP meaning |
|---|---|
|
Conditional Fate Affinity |
|
Discounted Fate Reach |
|
future-fate entropy |
|
Future-Fate Specificity |
|
Resolved Commitment |
|
discounted selected-fate probability by fate |
|
Signed Ordering Flux |
|
one-step selected-path coverage |