From e9829119395ff06b1e96ff310b02c8f62d3a7dcb Mon Sep 17 00:00:00 2001
From: Snider <snider@host.uk.com>
Date: Thu, 26 Feb 2026 06:32:24 +0000
Subject: [PATCH] fix: normalise LEK score to 0-100 via tanh sigmoid

Raw weighted sums ranged -25..+20, causing all text to land below the
ai_generated threshold (< 25). Now 50 = neutral (no signal), negatives
push toward 0 (AI markers), positives push toward 100 (human markers).

Co-Authored-By: Virgil <virgil@lethean.io>
---
 pkg/heuristic/heuristic.go      |  8 ++++++-
 pkg/heuristic/heuristic_test.go | 37 ++++++++++++++++++---------------
 2 files changed, 27 insertions(+), 18 deletions(-)

diff --git a/pkg/heuristic/heuristic.go b/pkg/heuristic/heuristic.go
index 75186e1..5438ca5 100644
--- a/pkg/heuristic/heuristic.go
+++ b/pkg/heuristic/heuristic.go
@@ -263,8 +263,10 @@ func scoreEmptyOrBroken(response string) int {
 }
 
 // computeLEKScore calculates the composite LEK score from heuristic sub-scores.
+// The raw weighted sum is normalised to a 0-100 scale using a tanh sigmoid,
+// where 50 = no signal (neutral), 0 = strong AI markers, 100 = strong human markers.
 func computeLEKScore(scores *Scores) {
-	scores.LEKScore = float64(scores.EngagementDepth)*2 +
+	raw := float64(scores.EngagementDepth)*2 +
 		float64(scores.CreativeForm)*3 +
 		float64(scores.EmotionalRegister)*2 +
 		float64(scores.FirstPerson)*1.5 -
@@ -272,4 +274,8 @@ func computeLEKScore(scores *Scores) {
 		float64(scores.FormulaicPreamble)*3 -
 		float64(scores.Degeneration)*4 -
 		float64(scores.EmptyBroken)*20
+
+	// Sigmoid normalisation: maps typical range (-25..+20) to 0..100.
+	// Divisor of 15 centres the curve for heuristic-only scoring.
+	scores.LEKScore = math.Round((50+50*math.Tanh(raw/15))*10) / 10
 }
diff --git a/pkg/heuristic/heuristic_test.go b/pkg/heuristic/heuristic_test.go
index abe26d2..2a96986 100644
--- a/pkg/heuristic/heuristic_test.go
+++ b/pkg/heuristic/heuristic_test.go
@@ -209,10 +209,12 @@ func TestEmptyOrBroken(t *testing.T) {
 }
 
 func TestLEKScoreComposite(t *testing.T) {
+	// LEK is normalised to 0-100 via tanh sigmoid. 50 = neutral.
 	tests := []struct {
-		name   string
+		name string
 		scores Scores
-		want   float64
+		wantMin float64
+		wantMax float64
 	}{
 		{
 			name: "all positive",
@@ -222,8 +224,8 @@ func TestLEKScoreComposite(t *testing.T) {
 				EmotionalRegister: 3,
 				FirstPerson:       2,
 			},
-			// 5*2 + 2*3 + 3*2 + 2*1.5 = 10+6+6+3 = 25
-			want: 25,
+			// raw = 25, normalised → ~96
+			wantMin: 90, wantMax: 100,
 		},
 		{
 			name: "all negative",
@@ -233,8 +235,8 @@ func TestLEKScoreComposite(t *testing.T) {
 				Degeneration:      5,
 				EmptyBroken:       1,
 			},
-			// -2*5 - 1*3 - 5*4 - 1*20 = -10-3-20-20 = -53
-			want: -53,
+			// raw = -53, normalised → ~0
+			wantMin: 0, wantMax: 2,
 		},
 		{
 			name: "mixed",
@@ -246,21 +248,22 @@ func TestLEKScoreComposite(t *testing.T) {
 				ComplianceMarkers: 1,
 				FormulaicPreamble: 1,
 			},
-			// 3*2 + 1*3 + 2*2 + 4*1.5 - 1*5 - 1*3 = 6+3+4+6-5-3 = 11
-			want: 11,
+			// raw = 11, normalised → ~81
+			wantMin: 75, wantMax: 85,
 		},
 		{
 			name:   "all zero",
 			scores: Scores{},
-			want:   0,
+			// raw = 0, normalised → 50.0
+			wantMin: 50, wantMax: 50,
 		},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
 			s := tt.scores
 			computeLEKScore(&s)
-			if s.LEKScore != tt.want {
-				t.Errorf("computeLEKScore() = %f, want %f", s.LEKScore, tt.want)
+			if s.LEKScore < tt.wantMin || s.LEKScore > tt.wantMax {
+				t.Errorf("computeLEKScore() = %.1f, want %.0f-%.0f", s.LEKScore, tt.wantMin, tt.wantMax)
 			}
 		})
 	}
@@ -273,8 +276,8 @@ func TestScore(t *testing.T) {
 		if scores.ComplianceMarkers < 4 {
 			t.Errorf("expected >= 4 compliance markers, got %d", scores.ComplianceMarkers)
 		}
-		if scores.LEKScore >= 0 {
-			t.Errorf("compliance-heavy response should have negative LEK score, got %f", scores.LEKScore)
+		if scores.LEKScore >= 50 {
+			t.Errorf("compliance-heavy response should score below 50 (neutral), got %.1f", scores.LEKScore)
 		}
 	})
 
@@ -294,8 +297,8 @@ func TestScore(t *testing.T) {
 		if scores.EmotionalRegister < 3 {
 			t.Errorf("expected emotional_register >= 3, got %d", scores.EmotionalRegister)
 		}
-		if scores.LEKScore <= 0 {
-			t.Errorf("creative response should have positive LEK score, got %f", scores.LEKScore)
+		if scores.LEKScore <= 50 {
+			t.Errorf("creative response should score above 50 (neutral), got %.1f", scores.LEKScore)
 		}
 	})
 
@@ -307,8 +310,8 @@ func TestScore(t *testing.T) {
 		if scores.Degeneration != 10 {
 			t.Errorf("expected degeneration = 10, got %d", scores.Degeneration)
 		}
-		if scores.LEKScore >= 0 {
-			t.Errorf("empty response should have very negative LEK score, got %f", scores.LEKScore)
+		if scores.LEKScore >= 10 {
+			t.Errorf("empty response should score near 0, got %.1f", scores.LEKScore)
 		}
 	})