TMComposites gate: per-gun confidence faithful for 3 guns; no pair composes

Adds a per-sample intrinsic-confidence field (GunPrediction.confidence,
threaded through FeedbackEvent/VirtualBullet, populated by Pattern, DecayGF,
KNN, GuessFactor, Tsetlin, TMHorizon) and an offline recorder + analyzer that
reproduce the paper's Figure 2 per gun and its Eq-8 composite.

Measured on 3 held-out tr-bridge DrussGT battles (33k ticks, ~133k samples/gun):
- FAITHFUL: DecayGF (rho +0.133), KNN (+0.090), Pattern (+0.064, weak).
- GuessFactor is ANTI-faithful (rho -0.067); Tsetlin c_max is useless (0.001).
- No pair of guns specialises complementarily: the same gun dominates both
  high-confidence slices in every pair.
- Eq-8 alpha-normalised confidence-weighted composite: 18.41% vs Pattern
  20.45% (McNemar p=3.1e-126). Faithful-only variant 18.68%, still loses.
  Shuffle control passes weakly (composite > shuffle, p=4e-14) so ~0.7pp of
  competence is real but ~2pp short. Offline veto: design is dead.

See docs/tmcomposites_gate.md.

common_libs/gun_harness/gun_interface.nim

@@ -35,6 +35,14 @@ type
   GunPrediction* = object
     ## Absolute (x, y) where the gun predicts the enemy will be.
     x*, y*: float
+    confidence*: float
+      ## The gun's INTRINSIC, per-sample confidence in THIS prediction, on any
+      ## positive scale the gun likes (the TMComposites gate normalises it by
+      ## the sample range, Eq 7). 0.0 means "this gun has no confidence signal"
+      ## (every deterministic geometric gun): such a gun casts no composite
+      ## vote. Deliberately NOT a rolling hit-rate: that is the selector
+      ## mechanism already measured negative (docs/gun_rack_analysis.md).
+      ## See docs/tmcomposites_gate.md and common_libs/tests/measure_tmcomposites.nim.
 
   FeedbackEvent* = object
     ## Outcome of a resolved virtual bullet.
@@ -45,6 +53,7 @@ type
     powerBin*:    int    ## index into PowerBins the bullet belongs to
     missDistance*: float  ## px; < BotRadius = hit
     hit*:         bool
+    confidence*:  float  ## copied from the spawn-time GunPrediction (above)
 
 proc bulletSpeed*(power: float): float {.inline.} =
   20.0 - 3.0 * power

common_libs/gun_harness/virtual_bullets.nim

@@ -348,6 +348,7 @@ type
     bulletSpeed*:  float
     travelDist*:   float   ## accumulated px so far
     fireDist*:     float   ## distance to target at fire time
+    confidence*:   float   ## spawn-time GunPrediction.confidence (TMComposites gate)
     active*:       bool
     pointScored*:  bool    ## the parallel point-model score for this flight has
                            ## been recorded (tie-break mode only)
@@ -461,6 +462,7 @@ proc spawnBullets*(t: var VirtualTracker, gunId: GunId,
       bulletSpeed: speed,
       travelDist: 0.0,
       fireDist:   fireDist,
+      confidence: pred.confidence,
       active:     true,
       pointScored: false,
       hitSeen:      false,
@@ -643,7 +645,7 @@ proc tickBullets*(t: var VirtualTracker, state: WorldState,
         t.fitness[b.targetId][b.gunId].bins[b.powerBin].record(hit)
 
       let fe = FeedbackEvent(
-        prediction:  GunPrediction(x: b.aimX, y: b.aimY),
+        prediction:  GunPrediction(x: b.aimX, y: b.aimY, confidence: b.confidence),
         actualX:     ex,
         actualY:     ey,
         bulletPower: PowerBins[b.powerBin],
@@ -651,6 +653,7 @@ proc tickBullets*(t: var VirtualTracker, state: WorldState,
         powerBin:    b.powerBin,
         missDistance: missDist,
         hit:         hit,
+        confidence:  b.confidence,
       )
       onResolved(b.gunId, b.powerBin, fe)
       b.active = false
@@ -720,7 +723,7 @@ proc tickBullets*(t: var VirtualTracker, state: WorldState,
         if b.targetId in t.fitness:
           t.fitness[b.targetId][b.gunId].bins[b.powerBin].record(b.hitSeen)
         let fe = FeedbackEvent(
-          prediction:  GunPrediction(x: b.aimX, y: b.aimY),
+          prediction:  GunPrediction(x: b.aimX, y: b.aimY, confidence: b.confidence),
           actualX:     rx,
           actualY:     ry,
           bulletPower: PowerBins[b.powerBin],
@@ -728,6 +731,7 @@ proc tickBullets*(t: var VirtualTracker, state: WorldState,
           powerBin:    b.powerBin,
           missDistance: missDist,
           hit:         b.hitSeen,
+          confidence:  b.confidence,
         )
         onResolved(b.gunId, b.powerBin, fe)
         b.active = false