Weight deep Ashtakavarga support in finance scoring
This commit is contained in:
@@ -0,0 +1,43 @@
|
||||
# Ashtakavarga Deep Support Weighting v1
|
||||
|
||||
Date: 2026-06-28
|
||||
|
||||
## Scope
|
||||
|
||||
This pass upgrades the newly bridged finance evidence layers from visibility-only into controlled low-weight score modifiers:
|
||||
|
||||
- `PAV / Prastara AV`
|
||||
- `Sodhita AV`
|
||||
- `Kakshya`
|
||||
|
||||
## Weighting
|
||||
|
||||
The v1 weights are intentionally small:
|
||||
|
||||
- `PAV supportive`: `+2`
|
||||
- `Kakshya supportive`: `+2`
|
||||
- `Sodhita obstructive`: `-2`
|
||||
- `Kakshya obstructive`: `-2`
|
||||
|
||||
## Boundary
|
||||
|
||||
These weights are not allowed to directly mint or revoke:
|
||||
|
||||
- `dominant_label`
|
||||
- `payout_label`
|
||||
|
||||
They only adjust the route score through the existing threshold system.
|
||||
|
||||
## Why This Matters
|
||||
|
||||
The project already computed these layers, but they were previously dead weight inside strict adjudication.
|
||||
|
||||
This pass keeps the scoring conservative while letting deeper Ashtakavarga and degree-trigger signals actually influence finance ranking.
|
||||
|
||||
## Still Open
|
||||
|
||||
- calibrating whether `PAV` should distinguish contributor identity more sharply
|
||||
- deciding whether `Sodhita` should also have a positive support path
|
||||
- tying `Kakshya` more directly to transit timing instead of route-average usage
|
||||
- validating the weights against external oracle cases before claiming mature timing precision
|
||||
|
||||
Reference in New Issue
Block a user