fix: purge ML validation label leakage

- Click any metric card to see historical periods where it was at a similar level
- Purple dot highlighting on chart shows matching periods
- Metric overlay line plotted on chart (dashed purple)
- Metric Context panel shows percentile, comparable days, avg forward returns,
  and historical examples from different market cycles
- New /api/metric-context endpoint for per-metric similarity analysis
- Backtest chart_data now includes per-metric raw values
- score_day() returns raw metric values alongside scores
- Fixed JS SyntaxError from broken inline onclick escaping (uses addEventListener)

Co-Authored-By: Claude Opus 4.6 <<EMAIL>>

backtesting/engine.py

@@ -146,9 +146,10 @@ _BT_ML_KEY_MAP = {
 
 
 def score_day(date, index, drawdowns, ml_weights=None):
-    """Score a single day using all available metrics. Returns (composite_score, individual_scores, n_metrics).
+    """Score a single day using all available metrics. Returns (composite_score, details, n_metrics).
 
     If ml_weights is provided, uses ML-optimized weighting instead of equal weights.
+    details includes both "score" and "raw" (the actual metric value before scoring).
     """
     scores = []
     details = {}
@@ -160,7 +161,7 @@ def score_day(date, index, drawdowns, ml_weights=None):
             s = _score_range(val, cfg["ranges"])
             if s is not None:
                 scores.append(s)
-                details[metric_key] = {"value": val, "score": s}
+                details[metric_key] = {"value": val, "score": s, "raw": val}
 
     # Ratio-based metrics (price vs reference)
     for metric_key, cfg in RATIO_SCORERS.items():
@@ -177,7 +178,7 @@ def score_day(date, index, drawdowns, ml_weights=None):
             s = _score_range(pct_above, cfg["ranges"])
             if s is not None:
                 scores.append(s)
-                details[metric_key] = {"value": pct_above, "score": s}
+                details[metric_key] = {"value": pct_above, "score": s, "raw": pct_above}
 
     # Drawdown
     dd = drawdowns.get(date)
@@ -185,7 +186,7 @@ def score_day(date, index, drawdowns, ml_weights=None):
         s = _score_range(dd, DRAWDOWN_RANGES)
         if s is not None:
             scores.append(s)
-            details["drawdown"] = {"value": dd, "score": s}
+            details["drawdown"] = {"value": dd, "score": s, "raw": dd}
 
     if not scores:
         return None, details, 0
@@ -297,12 +298,19 @@ def run_backtest(ml_mode=False):
         composite, details, n_metrics = score_day(d, index, drawdowns, ml_weights=ml_weights)
         if composite is not None and n_metrics >= 3:  # Require at least 3 metrics
             price = price_lookup.get(d)
+            # Collect raw metric values for per-metric historical exploration
+            metric_values = {}
+            for mk, info in details.items():
+                raw = info.get("raw")
+                if raw is not None:
+                    metric_values[mk] = round(raw, 6) if isinstance(raw, float) else raw
             entry = {
                 "date": d,
                 "score": composite,
                 "n_metrics": n_metrics,
                 "price": price,
                 "forward_returns": fwd_returns.get(d, {}),
+                "metric_values": metric_values,
             }
             daily_scores.append(entry)
 
@@ -462,6 +470,7 @@ def run_backtest(ml_mode=False):
 
     # --- Build time series for charting ---
     # Smart downsampling: daily for last 2 years, weekly before that
+    # Include per-metric values so the frontend can plot any metric.
     chart_data = []
     import datetime as _dt
     try:
@@ -469,14 +478,25 @@ def run_backtest(ml_mode=False):
         cutoff_date = (last_date - _dt.timedelta(days=730)).strftime("%Y-%m-%d")
     except Exception:
         cutoff_date = "2024-01-01"
+
+    # Collect all metric keys that were ever scored (for per-metric series)
+    all_metric_keys = set()
+    for d in daily_scores:
+        all_metric_keys.update(d.get("metric_values", {}).keys())
+
     for i, d in enumerate(daily_scores):
         is_recent = d["date"] >= cutoff_date
         if is_recent or i % 7 == 0 or i == len(daily_scores) - 1:
-            chart_data.append({
+            entry = {
                 "date": d["date"],
                 "score": d["score"],
                 "price": d["price"],
-            })
+            }
+            # Include per-metric values (raw metric value, not score)
+            metric_vals = d.get("metric_values", {})
+            if metric_vals:
+                entry["metrics"] = metric_vals
+            chart_data.append(entry)
 
     result = {
         "date_range": {"start": daily_scores[0]["date"], "end": daily_scores[-1]["date"]},

config/llm_settings.json

@@ -1,9 +1,9 @@
 {
-  "provider": "openrouter",
-  "model": "minimax/minimax-m2.5",
+  "provider": "ollama",
+  "model": "gemma4:12b-mlx",
   "providers": {
     "ollama": {
-      "base_url": "http://100.100.242.21:11434"
+      "base_url": "http://100.79.255.5:11434"
     },
     "lmstudio": {
       "base_url": "http://100.100.242.21:1234"

dashboard/server.py

@@ -165,7 +165,9 @@ def run_scrape(force_full=False):
         # 3. On-chain metrics — use cached values (historical data is permanent)
         onchain_keys = ["puell_multiple", "mvrv_zscore", "reserve_risk", "rhodl_ratio",
                         "nupl", "200w_sma", "lth_realized_price", "hash_ribbons",
-                        "pi_cycle_bottom", "lth_supply"]
+                        "pi_cycle_bottom", "lth_supply", "sopr", "sellside_risk",
+                        "active_address_momentum", "txcount_momentum", "nvt_price",
+                        "vdd_multiple"]
         
         has_cached_onchain = any(existing_cache.get(k, {}).get("value") is not None for k in onchain_keys)
         
@@ -176,6 +178,12 @@ def run_scrape(force_full=False):
                 from scrapers import lookintobitcoin
                 onchain = lookintobitcoin.scrape_all()
                 metrics.update(onchain)
+                try:
+                    from scrapers import checkonchain
+                    metrics.update(checkonchain.scrape_all())
+                except Exception as e:
+                    log.error("CheckOnChain scraping failed: %s\n%s", e, traceback.format_exc())
+                    _last_error = f"CheckOnChain scraping failed: {e}"
                 metrics["_onchain_timestamp"] = datetime.now(timezone.utc).isoformat()
             except Exception as e:
                 log.error("LookIntoBitcoin scraping failed: %s\n%s", e, traceback.format_exc())
@@ -341,6 +349,46 @@ def _fetch_models(provider, providers):
 
 # ── API Routes ────────────────────────────────────────────────────────────
 
+def _with_informational_onchain_metrics(scored, cache):
+    """Add non-scored on-chain data cards without changing composite scoring."""
+    if not isinstance(scored, dict):
+        return scored
+
+    enriched = dict(scored)
+    metrics = [dict(m) for m in scored.get("metrics", [])]
+    existing_keys = {m.get("key") for m in metrics}
+
+    lth_supply = cache.get("lth_supply", {})
+    lth_value = lth_supply.get("value")
+    if lth_value is not None and "lth_supply" not in existing_keys:
+        trend = lth_supply.get("trend")
+        trend_text = f" — {trend}" if trend else ""
+        metrics.append({
+            "name": "Long-Term Holder Supply",
+            "key": "lth_supply",
+            "value": lth_value,
+            "display_value": f"{lth_value:,.0f} BTC",
+            "score": None,
+            "description": "Informational on-chain metric; not included in the composite score" + trend_text,
+            "recent": lth_supply.get("recent", []),
+        })
+
+    pi_cycle = cache.get("pi_cycle_bottom", {})
+    pi_value = pi_cycle.get("value")
+    if pi_value is not None and "pi_cycle_bottom" not in existing_keys:
+        metrics.append({
+            "name": "Pi Cycle Bottom",
+            "key": "pi_cycle_bottom",
+            "value": pi_value,
+            "display_value": f"{pi_value:,.2f}" if isinstance(pi_value, (int, float)) else str(pi_value),
+            "score": None,
+            "description": "Informational on-chain cycle metric; not included in the composite score",
+            "recent": pi_cycle.get("recent", []),
+        })
+
+    enriched["metrics"] = metrics
+    return enriched
+
 @app.get("/api/data")
 def api_data(mode: str = "classic"):
     """Return current cached metrics + scores.
@@ -351,6 +399,7 @@ def api_data(mode: str = "classic"):
         scored = cache.get("_scored_ml", cache.get("_scored", {}))
     else:
         scored = cache.get("_scored", {})
+    scored = _with_informational_onchain_metrics(scored, cache)
     price_data = cache.get("price", {})
     drawdown_data = cache.get("drawdown", {})
     extras = cache.get("price_extras", {})
@@ -503,8 +552,17 @@ DASHBOARD_HTML = """<!DOCTYPE html>
 .meta-row{display:flex;gap:16px;flex-wrap:wrap;margin-top:8px;font-size:.8rem;color:var(--text-dim)}
 .meta-row span{display:flex;align-items:center;gap:4px}
 .metrics-grid{display:grid;grid-template-columns:repeat(auto-fill,minmax(300px,1fr));gap:12px;margin-bottom:20px}
-.metric-card{background:var(--card);border-radius:10px;padding:14px;border:1px solid var(--border);transition:border-color .15s}
+.metric-card{background:var(--card);border-radius:10px;padding:14px;border:1px solid var(--border);transition:border-color .15s;cursor:pointer}
 .metric-card:hover{border-color:var(--text-dim)}
+.metric-card.selected{border-color:#a78bfa;box-shadow:0 0 0 1px #a78bfa,0 0 12px rgba(167,139,250,0.15)}
+.metric-click-hint{font-size:.6rem;margin-left:4px;opacity:0;transition:opacity .15s}
+.metric-card:hover .metric-click-hint{opacity:.5}
+.metric-card.selected .metric-click-hint{opacity:1}
+.mc-examples-title{font-size:.75rem;color:#94a3b8;text-transform:uppercase;letter-spacing:.06em;margin-bottom:6px}
+.mc-example{font-size:.8rem;font-family:var(--mono);padding:4px 0;border-bottom:1px solid rgba(255,255,255,0.03)}
+.mc-ex-date{color:#e2e8f0}
+.mc-ex-cycle{color:#a78bfa;font-size:.7rem}
+.mc-ex-price{color:#94a3b8}
 .metric-header{display:flex;justify-content:space-between;align-items:flex-start;margin-bottom:8px}
 .metric-name{font-size:.85rem;font-weight:600}
 .metric-score{display:flex;align-items:center;gap:6px}
@@ -588,6 +646,20 @@ DASHBOARD_HTML = """<!DOCTYPE html>
   <a href="/backtest" style="font-size:.8rem;color:#22d3ee;text-decoration:none;margin-top:8px;display:inline-block">View full backtest &rarr;</a>
 </div>
 
+<!-- Metric Context Panel (shown when a metric is selected) -->
+<div class="card" id="metricContextPanel" style="margin-bottom:20px;display:none;border-color:#a78bfa">
+  <div style="display:flex;justify-content:space-between;align-items:center;margin-bottom:8px">
+    <h2 style="color:#a78bfa" id="mcTitle">Metric Context</h2>
+    <button onclick="if(selectedMetric) selectMetric(selectedMetric, '')" style="background:none;border:1px solid #a78bfa;color:#a78bfa;padding:4px 10px;border-radius:4px;cursor:pointer;font-family:var(--mono);font-size:.75rem">✕ Clear</button>
+  </div>
+  <div id="mcCurrent" style="font-size:1rem;font-family:var(--mono);margin-bottom:4px"></div>
+  <div id="mcPercentile" style="font-size:.8rem;color:#94a3b8;font-family:var(--mono);margin-bottom:4px"></div>
+  <div id="mcComparable" style="font-size:.8rem;color:#94a3b8;font-family:var(--mono);margin-bottom:8px"></div>
+  <div id="mcReturns" style="font-size:.9rem;font-family:var(--mono);line-height:1.6;margin-bottom:8px"></div>
+  <div id="mcExamples" style="display:none"></div>
+  <a href="/backtest" style="font-size:.8rem;color:#a78bfa;text-decoration:none;margin-top:8px;display:inline-block">View full backtest &rarr;</a>
+</div>
+
 <!-- Metrics Grid -->
 <h2>On-Chain Metrics</h2>
 <div class="metrics-grid" id="metricsGrid">
@@ -690,6 +762,8 @@ function drawSparkline(canvasId, data, color) {
   ctx.stroke();
 }
 
+let selectedMetric = null;
+
 function renderMetrics(metrics) {
   const grid = document.getElementById('metricsGrid');
   if (!metrics || !metrics.length) {
@@ -703,10 +777,11 @@ function renderMetrics(metrics) {
     const color = m.score != null ? scoreColor(m.score, 10) : '#64748b';
     const fillPct = m.score != null ? (m.score / 10 * 100) : 0;
     const hasSparkline = m.recent && m.recent.length > 2;
+    const isSelected = selectedMetric === m.key ? ' selected' : '';
 
-    html += '<div class="metric-card">';
+    html += '<div class="metric-card' + isSelected + '" data-key="' + m.key + '" data-name="' + m.name.replace('"', '&quot;') + '">';
     html += '<div class="metric-header">';
-    html += '<div class="metric-name">' + m.name + '</div>';
+    html += '<div class="metric-name">' + m.name + '<span class="metric-click-hint">👆</span></div>';
     html += '<div class="metric-score">';
     html += '<div class="metric-score-bar"><div class="metric-score-fill" style="width:' + fillPct + '%;background:' + color + '"></div></div>';
     html += '<div class="metric-score-num" style="color:' + color + '">' + score + '</div>';
@@ -733,11 +808,112 @@ function renderMetrics(metrics) {
       }
     });
   });
+
+  // Attach click handlers to metric cards
+  document.querySelectorAll('.metric-card').forEach(card => {
+    card.addEventListener('click', function() {
+      const key = this.getAttribute('data-key');
+      const name = this.getAttribute('data-name');
+      if (key) selectMetric(key, name);
+    });
+  });
+}
+
+// Metric selection + context panel
+function selectMetric(metricKey, metricName) {
+  if (selectedMetric === metricKey) {
+    // Deselect if clicking the same one
+    selectedMetric = null;
+    window._highlightMetric = null;
+    document.getElementById('metricContextPanel').style.display = 'none';
+    const panel = document.getElementById('histContext');
+    if (panel) panel.style.display = 'block';
+    applyChartRange(currentRange); // Re-render chart without highlight
+  } else {
+    selectedMetric = metricKey;
+    loadMetricContext(metricKey, metricName);
+  }
+  poll(); // Re-render metric cards with highlight
+}
+
+async function loadMetricContext(metricKey, metricName) {
+  try {
+    const r = await fetch('/api/metric-context?metric=' + encodeURIComponent(metricKey) + '&mode=' + currentMode);
+    const ctx = await r.json();
+    if (ctx.error) {
+      showToast(ctx.error, 'error');
+      return;
+    }
+
+    // Show metric context panel, hide composite context
+    const panel = document.getElementById('histContext');
+    if (panel) panel.style.display = 'none';
+
+    const mcp = document.getElementById('metricContextPanel');
+    mcp.style.display = 'block';
+
+    document.getElementById('mcTitle').textContent = metricName;
+    document.getElementById('mcCurrent').textContent = 'Current: ' + (ctx.current_raw != null ? ctx.current_raw : 'N/A');
+    document.getElementById('mcPercentile').textContent = 'Metric value in top ' + (100 - ctx.percentile).toFixed(1) + '% historically';
+    document.getElementById('mcComparable').textContent = ctx.comparable_days + ' comparable days found';
+
+    const fmtR = (v) => v == null ? '--' : (v >= 0 ? '+' : '') + v.toFixed(1) + '%';
+    const cR = v => v != null && v >= 0 ? '#22c55e' : '#ef4444';
+    const periods = [['30d', ctx.avg_30d_return], ['90d', ctx.avg_90d_return], ['180d', ctx.avg_180d_return], ['1yr', ctx.avg_1yr_return]];
+    let retHtml = '';
+    for (const [label, val] of periods) {
+      if (val != null) retHtml += '<strong style="color:' + cR(val) + '">' + label + ': ' + fmtR(val) + '</strong> · ';
+    }
+    document.getElementById('mcReturns').innerHTML = retHtml ? 'Avg returns when ' + metricName + ' was similar: ' + retHtml : 'No forward return data available';
+
+    // Examples
+    const exEl = document.getElementById('mcExamples');
+    if (ctx.examples && ctx.examples.length) {
+      let exHtml = '<div class="mc-examples-title">Historical examples:</div>';
+      ctx.examples.forEach(ex => {
+        const fwd30 = ex.forward_returns['30d'];
+        const fwd365 = ex.forward_returns['365d'];
+        exHtml += '<div class="mc-example">';
+        exHtml += '<span class="mc-ex-date">' + ex.date + '</span> ';
+        exHtml += '<span class="mc-ex-cycle">' + ex.cycle + '</span> ';
+        exHtml += '<span class="mc-ex-price">$' + (ex.price ? ex.price.toLocaleString() : 'N/A') + '</span>';
+        if (fwd30 != null) exHtml += ' <span style="color:' + cR(fwd30) + '">30d: ' + fmtR(fwd30) + '</span>';
+        if (fwd365 != null) exHtml += ' <span style="color:' + cR(fwd365) + '">1yr: ' + fmtR(fwd365) + '</span>';
+        exHtml += '</div>';
+      });
+      exEl.innerHTML = exHtml;
+      exEl.style.display = 'block';
+    } else {
+      exEl.innerHTML = '';
+      exEl.style.display = 'none';
+    }
+
+    // Highlight matching periods on the chart
+    highlightMetricPeriods(metricKey, ctx.current_raw, ctx.margin);
+  } catch(e) {
+    console.error('Metric context load failed:', e);
+  }
+}
+
+function highlightMetricPeriods(metricKey, currentRaw, margin) {
+  if (!fullDailyScores || !currentRaw || margin == null) return;
+
+  // Build an array of {date, rawValue} for the selected metric
+  const metricSeries = fullDailyScores
+    .filter(d => d.metrics && d.metrics[metricKey] != null)
+    .map(d => ({ date: d.date, value: d.metrics[metricKey], isSimilar: Math.abs(d.metrics[metricKey] - currentRaw) <= margin }));
+
+  // Store for use in chart rendering
+  window._highlightMetric = { key: metricKey, series: metricSeries, currentRaw, margin };
+
+  // Re-render chart with highlight
+  applyChartRange(currentRange);
 }
 
 let histChart = null;
 let fullDailyScores = null;
 let currentRange = 0; // 0 = ALL
+let currentMode = 'classic';
 
 function renderHistory(history) {
   // Legacy: still called by loadData but we'll use backtest data instead
@@ -784,7 +960,61 @@ function renderHistoryFromData(history) {
     });
   }
 
-  // Accumulation zone backgrounds
+  // If a metric is selected, add its overlay + highlight similar periods
+  const highlight = window._highlightMetric;
+  let metricColor = '#a78bfa';
+  if (highlight && highlight.series && highlight.series.length) {
+    // Build a sparse array aligned to current chart labels
+    const metricByDate = {};
+    highlight.series.forEach(s => { metricByDate[s.date] = s; });
+    const metricData = labels.map(l => {
+      const entry = metricByDate[l];
+      return entry ? entry.value : null;
+    });
+    const hasMetricData = metricData.some(v => v != null);
+
+    if (hasMetricData) {
+      datasets.push({
+        label: 'Selected Metric',
+        data: metricData,
+        borderColor: metricColor,
+        borderWidth: 1.5,
+        borderDash: [2, 2],
+        fill: false,
+        tension: 0.2,
+        pointRadius: 0,
+        yAxisID: 'y2',
+      });
+    }
+
+    // Highlight similar periods with point dots on the score line
+    const similarIndices = [];
+    labels.forEach((l, i) => {
+      const entry = metricByDate[l];
+      if (entry && entry.isSimilar) similarIndices.push(i);
+    });
+
+    if (similarIndices.length) {
+      const highlightData = labels.map((l, i) =>
+        similarIndices.includes(i) ? scores[i] : null
+      );
+      datasets.push({
+        label: 'Similar Periods',
+        data: highlightData,
+        borderColor: 'rgba(167,139,250,0)',
+        backgroundColor: '#a78bfa',
+        pointRadius: 3,
+        pointHoverRadius: 5,
+        showLine: false,
+        yAxisID: 'y',
+      });
+    }
+  }
+
+  // Determine y2 scale for the metric overlay
+  const hasMetricDataset = datasets.some(d => d.yAxisID === 'y2');
+
+  // Accumulation zone backgrounds + metric highlight bands
   const zonePlugin = {
     id: 'zones',
     beforeDraw(chart) {
@@ -812,9 +1042,55 @@ function renderHistoryFromData(history) {
         ctx.stroke();
         ctx.setLineDash([]);
       });
+
+      // Draw vertical highlight bands for similar periods
+      if (highlight && highlight.series) {
+        const metricByDate = {};
+        highlight.series.forEach(s => { metricByDate[s.date] = s; });
+        const xScale = chart.scales.x;
+        labels.forEach((l, i) => {
+          const entry = metricByDate[l];
+          if (entry && entry.isSimilar) {
+            const x = xScale.getPixelForValue(i);
+            ctx.fillStyle = 'rgba(167,139,250,0.08)';
+            ctx.fillRect(x - 3, top, 6, bottom - top);
+          }
+        });
+      }
     }
   };
 
+  const scales = {
+    x: {
+      ticks: { color: '#64748b', maxTicksLimit: 12, font: { family: 'monospace', size: 10 } },
+      grid: { color: 'rgba(255,255,255,0.03)' }
+    },
+    y: {
+      min: 0, max: 100,
+      ticks: { color: '#22d3ee', font: { family: 'monospace', size: 10 } },
+      grid: { color: 'rgba(255,255,255,0.03)' },
+      title: { display: true, text: 'Score', color: '#22d3ee', font: { family: 'monospace', size: 11 } }
+    },
+    y1: {
+      position: 'right',
+      ticks: {
+        color: '#f7931a',
+        font: { family: 'monospace', size: 10 },
+        callback: v => '$' + (v >= 1000 ? (v/1000).toFixed(0) + 'k' : v)
+      },
+      grid: { drawOnChartArea: false },
+      title: { display: true, text: 'BTC Price', color: '#f7931a', font: { family: 'monospace', size: 11 } }
+    },
+  };
+
+  if (hasMetricDataset) {
+    scales['y2'] = {
+      position: 'right',
+      display: false,
+      grid: { drawOnChartArea: false },
+    };
+  }
+
   histChart = new Chart(ctx, {
     type: 'line',
     plugins: [zonePlugin],
@@ -834,6 +1110,8 @@ function renderHistoryFromData(history) {
           callbacks: {
             label: function(ctx) {
               if (ctx.dataset.yAxisID === 'y1') return 'BTC: $' + ctx.raw.toLocaleString();
+              if (ctx.dataset.yAxisID === 'y2') return 'Metric: ' + (ctx.raw != null ? ctx.raw.toFixed(4) : 'N/A');
+              if (ctx.dataset.label === 'Similar Periods') return '★ Similar period (Score: ' + ctx.raw.toFixed(1) + ')';
               const s = ctx.raw;
               let zone = s >= 80 ? 'Extreme Accum' : s >= 65 ? 'Strong Accum' : s >= 50 ? 'Moderate' : s >= 35 ? 'Neutral' : 'Caution';
               return 'Score: ' + s.toFixed(1) + ' (' + zone + ')';
@@ -841,28 +1119,7 @@ function renderHistoryFromData(history) {
           }
         }
       },
-      scales: {
-        x: {
-          ticks: { color: '#64748b', maxTicksLimit: 12, font: { family: 'monospace', size: 10 } },
-          grid: { color: 'rgba(255,255,255,0.03)' }
-        },
-        y: {
-          min: 0, max: 100,
-          ticks: { color: '#22d3ee', font: { family: 'monospace', size: 10 } },
-          grid: { color: 'rgba(255,255,255,0.03)' },
-          title: { display: true, text: 'Score', color: '#22d3ee', font: { family: 'monospace', size: 11 } }
-        },
-        y1: {
-          position: 'right',
-          ticks: {
-            color: '#f7931a',
-            font: { family: 'monospace', size: 10 },
-            callback: v => '$' + (v >= 1000 ? (v/1000).toFixed(0) + 'k' : v)
-          },
-          grid: { drawOnChartArea: false },
-          title: { display: true, text: 'BTC Price', color: '#f7931a', font: { family: 'monospace', size: 11 } }
-        }
-      }
+      scales,
     }
   });
 }
@@ -992,8 +1249,6 @@ async function doRefresh(full) {
   setTimeout(() => { btn.disabled = false; btn.textContent = origText; }, delay);
 }
 
-let currentMode = 'classic';
-
 function setMode(mode) {
   currentMode = mode;
   document.querySelectorAll('.mode-btn').forEach(b => {
@@ -1286,6 +1541,245 @@ def api_backtest_status():
     return status
 
 
+@app.get("/api/metric-context")
+def api_metric_context(metric: str, margin: float = 0.0, mode: str = "classic"):
+    """Find historical periods where a specific metric was at a similar level.
+
+    Returns forward returns for those periods, analogous to the composite-score
+    current_context but filtered to a single metric's historical similarity.
+
+    margin: absolute tolerance for "similar" (auto-computed from metric scale if 0).
+    """
+    try:
+        from backtesting.engine import run_backtest, HISTORY_PATH, _build_daily_index, _get_all_dates, _last_known_value, METRIC_SCORERS, RATIO_SCORERS, DRAWDOWN_RANGES, _score_range
+        import os as _os
+
+        if not _os.path.exists(HISTORY_PATH):
+            return JSONResponse({"error": "No historical data. Run history collector first."}, status_code=404)
+
+        with open(HISTORY_PATH) as f:
+            history = json.load(f)
+
+        index = _build_daily_index(history)
+        all_dates = _get_all_dates(index)
+
+        # Get current metric value from cache
+        cache = {}
+        if _os.path.exists(CACHE_PATH):
+            with open(CACHE_PATH) as f:
+                cache = json.load(f)
+
+        current_raw = _get_current_metric_raw(metric, cache)
+        if current_raw is None:
+            return JSONResponse({"error": f"No current value for metric '{metric}'"}, status_code=404)
+
+        # Auto-compute margin from metric scale
+        if margin <= 0:
+            margin = _auto_metric_margin(metric, current_raw)
+
+        # Build price lookup
+        price_lookup = {}
+        for pk in ["btc_price_coingecko", "btc_price", "btc_price_sma", "btc_price_lth"]:
+            if pk in index:
+                for d, v in index[pk].items():
+                    if d not in price_lookup:
+                        price_lookup[d] = v
+
+        # Find historical days where this metric was similar
+        comparable = []
+        for d in all_dates:
+            raw_val = _get_historical_metric_raw(metric, index, d)
+            if raw_val is not None and abs(raw_val - current_raw) <= margin:
+                price = price_lookup.get(d)
+                fwd = _compute_day_forward_returns(price_lookup, d)
+                if fwd:
+                    comparable.append({
+                        "date": d,
+                        "raw_value": round(raw_val, 6) if isinstance(raw_val, float) else raw_val,
+                        "price": price,
+                        "forward_returns": fwd,
+                    })
+
+        # Compute average returns across comparable periods
+        avg_returns = {}
+        for period in ["30d", "90d", "180d", "365d"]:
+            vals = [c["forward_returns"][period] for c in comparable if period in c["forward_returns"]]
+            if vals:
+                avg_returns[period] = round(sum(vals) / len(vals), 2)
+
+        # Pick best examples (one per market cycle)
+        cycle_bins = [
+            ("pre-2016", "2010-01-01", "2015-12-31"),
+            ("2016-17 Bull", "2016-01-01", "2017-12-31"),
+            ("2018-19 Bear", "2018-01-01", "2019-12-31"),
+            ("2020-21 Bull", "2020-01-01", "2021-12-31"),
+            ("2022-23 Bear", "2022-01-01", "2023-12-31"),
+            ("2024+", "2024-01-01", "2099-12-31"),
+        ]
+        examples = []
+        used_cycles = set()
+        sorted_comp = sorted(comparable, key=lambda c: abs(c["raw_value"] - current_raw))
+        for c in sorted_comp:
+            for label, start, end in cycle_bins:
+                if start <= c["date"] <= end and label not in used_cycles:
+                    used_cycles.add(label)
+                    examples.append({
+                        "date": c["date"],
+                        "raw_value": c["raw_value"],
+                        "price": c["price"],
+                        "forward_returns": c["forward_returns"],
+                        "cycle": label,
+                    })
+                    break
+            if len(examples) >= 6:
+                break
+        examples.sort(key=lambda e: e["date"])
+
+        # Percentile: what % of all days had this metric at or below current value
+        all_raw_vals = []
+        for d in all_dates:
+            rv = _get_historical_metric_raw(metric, index, d)
+            if rv is not None:
+                all_raw_vals.append(rv)
+        all_raw_vals.sort()
+        below = len([v for v in all_raw_vals if v <= current_raw])
+        percentile = round(below / len(all_raw_vals) * 100, 1) if all_raw_vals else 50.0
+
+        return {
+            "metric": metric,
+            "current_raw": current_raw,
+            "margin": margin,
+            "comparable_days": len(comparable),
+            "percentile": percentile,
+            "avg_30d_return": avg_returns.get("30d"),
+            "avg_90d_return": avg_returns.get("90d"),
+            "avg_180d_return": avg_returns.get("180d"),
+            "avg_1yr_return": avg_returns.get("365d"),
+            "examples": examples,
+        }
+    except Exception as e:
+        log.error("Metric context error: %s", traceback.format_exc())
+        return JSONResponse({"error": str(e)}, status_code=500)
+
+
+def _get_current_metric_raw(metric, cache):
+    """Get the current raw value for a metric from the cache."""
+    # Direct cache keys
+    direct_keys = {
+        "fear_greed": ("fear_greed", "value"),
+        "puell_multiple": ("puell_multiple", "value"),
+        "mvrv_zscore": ("mvrv_zscore", "value"),
+        "reserve_risk": ("reserve_risk", "value"),
+        "rhodl_ratio": ("rhodl_ratio", "value"),
+        "nupl": ("nupl", "value"),
+        "drawdown": ("drawdown", "value"),
+        "hash_ribbons": ("hash_ribbons", "value"),
+        "sopr": ("sopr", "value"),
+        "sellside_risk": ("sellside_risk", "value"),
+        "active_address_momentum": ("active_address_momentum", "value"),
+        "txcount_momentum": ("txcount_momentum", "value"),
+        "nvt_price": ("nvt_price", "value"),
+        "vdd_multiple": ("vdd_multiple", "value"),
+        "lth_supply": ("lth_supply", "value"),
+    }
+    # Ratio-based metrics: compute from price vs reference
+    ratio_metrics = {
+        "price_vs_200w_sma": ("price", "200w_sma"),
+        "lth_realized_price": ("price", "lth_realized_price"),
+    }
+
+    if metric in direct_keys:
+        k, sub = direct_keys[metric]
+        val = cache.get(k, {})
+        if isinstance(val, dict):
+            return val.get(sub)
+        return val
+    elif metric in ratio_metrics:
+        price_key, ref_key = ratio_metrics[metric]
+        price_val = cache.get(price_key, {}).get("price") or cache.get(price_key, {}).get("value")
+        ref_val = cache.get(ref_key, {}).get("value")
+        if price_val and ref_val and ref_val > 0:
+            return ((price_val - ref_val) / ref_val) * 100
+    return None
+
+
+def _get_historical_metric_raw(metric, index, date):
+    """Get the raw value for a metric on a specific historical date."""
+    from backtesting.engine import _last_known_value
+    direct_keys = {
+        "fear_greed": "fear_greed",
+        "puell_multiple": "puell_multiple",
+        "mvrv_zscore": "mvrv_zscore",
+        "reserve_risk": "reserve_risk",
+        "rhodl_ratio": "rhodl_ratio",
+        "nupl": "nupl",
+        "drawdown": "drawdown",
+        "hash_ribbons": "hash_ribbons",
+        "sopr": "sopr",
+        "sellside_risk": "sellside_risk",
+        "active_address_momentum": "active_address_momentum",
+        "txcount_momentum": "txcount_momentum",
+        "nvt_price": "nvt_price",
+        "vdd_multiple": "vdd_multiple",
+        "lth_supply": "lth_supply",
+    }
+    if metric in direct_keys:
+        return _last_known_value(index.get(direct_keys[metric], {}), date)
+    # Ratio-based
+    if metric == "price_vs_200w_sma":
+        price_val = _last_known_value(index.get("btc_price", {}), date)
+        ref_val = _last_known_value(index.get("200w_sma", {}), date)
+        if price_val and ref_val and ref_val > 0:
+            return ((price_val - ref_val) / ref_val) * 100
+    if metric == "lth_realized_price":
+        price_val = _last_known_value(index.get("btc_price", {}), date)
+        ref_val = _last_known_value(index.get("lth_realized_price", {}), date)
+        if price_val and ref_val and ref_val > 0:
+            return ((price_val - ref_val) / ref_val) * 100
+    return None
+
+
+def _auto_metric_margin(metric, current_val):
+    """Compute a reasonable similarity margin based on metric type and scale."""
+    margins = {
+        "fear_greed": 5.0,
+        "puell_multiple": 0.15,
+        "mvrv_zscore": 0.5,
+        "reserve_risk": 0.002,
+        "rhodl_ratio": 300,
+        "nupl": 0.1,
+        "drawdown": 8.0,
+        "sopr": 0.02,
+        "sellside_risk": 0.001,
+        "active_address_momentum": 0.05,
+        "txcount_momentum": 0.05,
+        "nvt_price": 5000,
+        "vdd_multiple": 0.15,
+        "price_vs_200w_sma": 10.0,
+        "lth_realized_price": 10.0,
+    }
+    if metric in margins:
+        return margins[metric]
+    # Fallback: 15% of current value
+    return abs(current_val) * 0.15 if current_val != 0 else 1.0
+
+
+def _compute_day_forward_returns(price_lookup, date):
+    """Compute forward returns for a single date."""
+    from datetime import datetime as _dt, timedelta as _td
+    p0 = price_lookup.get(date)
+    if p0 is None or p0 <= 0:
+        return {}
+    r = {}
+    dt = _dt.strptime(date, "%Y-%m-%d")
+    for days in [30, 90, 180, 365]:
+        future = (dt + _td(days=days)).strftime("%Y-%m-%d")
+        pf = price_lookup.get(future)
+        if pf is not None:
+            r[f"{days}d"] = round(((pf - p0) / p0) * 100, 2)
+    return r
+
+
 # ── Backtest HTML Page ─────────────────────────────────────────────────
 
 BACKTEST_HTML = """<!DOCTYPE html>

data/score_history.jsonl

@@ -110,3 +110,40 @@
 {"timestamp": "2026-03-21T22:54:22.144542+00:00", "composite_score": 70.0, "scored_count": 9, "metrics": {"fear_greed": {"score": 10, "value": 12}, "puell_multiple": {"score": 8, "value": 0.6602699608966011}, "mvrv_zscore": {"score": 8, "value": 0.5211180167687892}, "drawdown": {"score": null, "value": null}, "price_vs_200w_sma": {"score": 7, "value": 58895.78086828114}, "reserve_risk": {"score": 10, "value": 0.0012985709697654493}, "rhodl_ratio": {"score": 4, "value": 1230.6243545314708}, "nupl": {"score": 8, "value": 0.22243290955405431}, "lth_realized_price": {"score": 5, "value": 43346.58756410873}, "hash_ribbons": {"score": 3, "value": null}}}
 {"timestamp": "2026-03-21T22:55:08.385540+00:00", "composite_score": 71.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 10, "value": 12}, "puell_multiple": {"score": 8, "value": 0.6602699608966011}, "mvrv_zscore": {"score": 8, "value": 0.5211180167687892}, "drawdown": {"score": 8, "value": 44.26554568527919}, "price_vs_200w_sma": {"score": 7, "value": 58895.78086828114}, "reserve_risk": {"score": 10, "value": 0.0012985709697654493}, "rhodl_ratio": {"score": 4, "value": 1230.6243545314708}, "nupl": {"score": 8, "value": 0.22243290955405431}, "lth_realized_price": {"score": 5, "value": 43346.58756410873}, "hash_ribbons": {"score": 3, "value": null}}}
 {"timestamp": "2026-03-21T22:55:33.933753+00:00", "composite_score": 71.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 10, "value": 12}, "puell_multiple": {"score": 8, "value": 0.6602699608966011}, "mvrv_zscore": {"score": 8, "value": 0.5211180167687892}, "drawdown": {"score": 8, "value": 44.26316624365482}, "price_vs_200w_sma": {"score": 7, "value": 58895.78086828114}, "reserve_risk": {"score": 10, "value": 0.0012985709697654493}, "rhodl_ratio": {"score": 4, "value": 1230.6243545314708}, "nupl": {"score": 8, "value": 0.22243290955405431}, "lth_realized_price": {"score": 5, "value": 43346.58756410873}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T18:10:22.517545+00:00", "composite_score": 63.3, "scored_count": 3, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": null, "value": null}, "mvrv_zscore": {"score": null, "value": null}, "drawdown": {"score": 8, "value": 52.07407994923858}, "price_vs_200w_sma": {"score": null, "value": null}, "reserve_risk": {"score": null, "value": null}, "rhodl_ratio": {"score": null, "value": null}, "nupl": {"score": null, "value": null}, "lth_realized_price": {"score": null, "value": null}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T18:25:30.189079+00:00", "composite_score": 63.3, "scored_count": 3, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": null, "value": null}, "mvrv_zscore": {"score": null, "value": null}, "drawdown": {"score": 8, "value": 52.098667512690355}, "price_vs_200w_sma": {"score": null, "value": null}, "reserve_risk": {"score": null, "value": null}, "rhodl_ratio": {"score": null, "value": null}, "nupl": {"score": null, "value": null}, "lth_realized_price": {"score": null, "value": null}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T18:26:00.556392+00:00", "composite_score": 63.3, "scored_count": 3, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": null, "value": null}, "mvrv_zscore": {"score": null, "value": null}, "drawdown": {"score": 8, "value": 52.098667512690355}, "price_vs_200w_sma": {"score": null, "value": null}, "reserve_risk": {"score": null, "value": null}, "rhodl_ratio": {"score": null, "value": null}, "nupl": {"score": null, "value": null}, "lth_realized_price": {"score": null, "value": null}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T18:40:30.805472+00:00", "composite_score": 63.3, "scored_count": 3, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": null, "value": null}, "mvrv_zscore": {"score": null, "value": null}, "drawdown": {"score": 8, "value": 52.04473350253808}, "price_vs_200w_sma": {"score": null, "value": null}, "reserve_risk": {"score": null, "value": null}, "rhodl_ratio": {"score": null, "value": null}, "nupl": {"score": null, "value": null}, "lth_realized_price": {"score": null, "value": null}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T18:55:31.544772+00:00", "composite_score": 63.3, "scored_count": 3, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": null, "value": null}, "mvrv_zscore": {"score": null, "value": null}, "drawdown": {"score": 8, "value": 52.066148477157356}, "price_vs_200w_sma": {"score": null, "value": null}, "reserve_risk": {"score": null, "value": null}, "rhodl_ratio": {"score": null, "value": null}, "nupl": {"score": null, "value": null}, "lth_realized_price": {"score": null, "value": null}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T18:57:42.639217+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.04790609137056}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T19:12:44.307793+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.02014593908629}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T19:27:45.019177+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.22874365482234}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T19:42:45.679921+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.35009517766498}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T19:57:46.474768+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.34850888324873}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T20:12:47.228456+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.2604695431472}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T20:27:47.904578+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.27633248730964}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T20:42:48.631663+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.26919416243655}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T20:57:49.365132+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.264435279187815}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T21:12:50.093194+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.202569796954315}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T21:27:50.802146+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.16132614213198}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T21:42:51.460898+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.24857233502538}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T21:57:52.223283+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.318369289340104}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T22:12:52.898551+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.37071700507614}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T22:27:54.711182+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.429409898477154}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-27T22:42:55.510257+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 15}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.31043781725888}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-28T20:31:14.232026+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.70066624365482}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-28T20:35:38.591732+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.73080583756345}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-28T20:42:21.081121+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.721288071065985}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-28T20:57:21.749415+00:00", "composite_score": 74.0, "scored_count": 10, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.75301395939086}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}}}
+{"timestamp": "2026-06-28T21:09:38.418891+00:00", "composite_score": 68.1, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.73159898477158}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 4, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T21:11:19.117377+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.723667512690355}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T21:26:19.843356+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.723667512690355}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T21:39:03.697500+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.68797588832488}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T21:42:37.143098+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.71890862944163}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T21:57:37.803751+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.96002538071066}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T22:12:38.557670+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.72049492385786}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T22:27:39.293792+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.94733502538072}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T22:42:40.175733+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 53.05678934010152}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T22:57:40.874889+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 53.22255710659899}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T23:12:41.665346+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 53.02744289340101}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}
+{"timestamp": "2026-06-28T23:27:42.349311+00:00", "composite_score": 69.4, "scored_count": 16, "metrics": {"fear_greed": {"score": 8, "value": 18}, "puell_multiple": {"score": 5, "value": 0.7044739567707577}, "mvrv_zscore": {"score": 8, "value": 0.22409759021503936}, "drawdown": {"score": 8, "value": 52.95923223350254}, "price_vs_200w_sma": {"score": 10, "value": 62284.65298428873}, "reserve_risk": {"score": 10, "value": 0.0010258831016337609}, "rhodl_ratio": {"score": 7, "value": 882.5868942025234}, "nupl": {"score": 8, "value": 0.11309045542914359}, "lth_realized_price": {"score": 7, "value": 49767.33015910989}, "hash_ribbons": {"score": 3, "value": null}, "sopr": {"score": 8, "value": 0.990007085019496}, "sellside_risk": {"score": 10, "value": 0.000734882488382521}, "active_address_momentum": {"score": 4, "value": -0.09386733861382784}, "txcount_momentum": {"score": 6, "value": 0.04454611494295241}, "nvt_price": {"score": 5, "value": 54673.815447216126}, "vdd_multiple": {"score": 4, "value": -0.030495759573562122}}}

ml/optimizer.py

@@ -43,6 +43,9 @@ START_DATE = "2018-02-01"
 TRAIN_CUTOFF_DAYS = 365
 # Target: forward 365d return > 30% = "good time to buy"
 GOOD_BUY_THRESHOLD = 30.0
+# Validation embargo/purge horizon: labels use 365-day forward returns.
+LABEL_HORIZON_DAYS = 365
+VALIDATION_SPLITS = 5
 
 # The 8 core metrics we score
 METRIC_KEYS = [
@@ -95,6 +98,112 @@ def score_range(value, ranges):
     return 0
 
 
+SCORE_KEYS = [
+    "puell_multiple", "mvrv_zscore", "reserve_risk", "rhodl_ratio",
+    "nupl", "fear_greed", "drawdown", "pct_above_200w_sma", "pct_above_lth_rp",
+]
+
+SCORE_FEATURES = [f"score_{k}" for k in SCORE_KEYS]
+RAW_FEATURES = [
+    "raw_puell_multiple", "raw_mvrv_zscore", "raw_reserve_risk",
+    "raw_rhodl_ratio", "raw_nupl", "raw_fear_greed",
+    "raw_pct_above_200w_sma", "raw_pct_above_lth_rp", "raw_drawdown",
+]
+DELTA_FEATURES = [
+    "delta_30d_mvrv_zscore", "delta_30d_nupl",
+    "delta_30d_puell_multiple", "delta_30d_reserve_risk",
+]
+INTERACTION_FEATURES = ["mvrv_x_nupl", "puell_x_reserve"]
+CYCLE_FEATURES = ["days_since_ath"]
+FEATURE_COLS = SCORE_FEATURES + RAW_FEATURES + DELTA_FEATURES + INTERACTION_FEATURES + CYCLE_FEATURES
+
+BRACKETS = [
+    (0, 20, "Extreme Caution"),
+    (21, 40, "Caution"),
+    (41, 55, "Neutral"),
+    (56, 70, "Moderate Opportunity"),
+    (71, 85, "Strong Accumulation"),
+    (86, 100, "Extreme Accumulation"),
+]
+
+
+def _row_date(row):
+    return datetime.strptime(row["date"], "%Y-%m-%d")
+
+
+def purged_time_series_splits(rows, n_splits=VALIDATION_SPLITS,
+                              label_horizon_days=LABEL_HORIZON_DAYS,
+                              embargo_days=0):
+    """Yield expanding-window splits with overlapping forward-label windows removed.
+
+    A row dated T with a 365-day forward-return label consumes information up to
+    T+365. For validation beginning at V, any training row whose label window
+    reaches V is removed. This keeps validation metrics out-of-sample for the
+    forward-return label, not just for features.
+    """
+    base_splitter = TimeSeriesSplit(n_splits=n_splits)
+    row_dates = [_row_date(r) for r in rows]
+    horizon = timedelta(days=label_horizon_days)
+    embargo = timedelta(days=embargo_days)
+
+    for train_idx, val_idx in base_splitter.split(np.arange(len(rows))):
+        val_start = row_dates[val_idx[0]]
+        val_end = row_dates[val_idx[-1]]
+        purged_train = []
+        for idx in train_idx:
+            label_end = row_dates[idx] + horizon
+            before_validation_label_window = label_end <= val_start - embargo
+            after_validation_embargo = row_dates[idx] > val_end + embargo
+            if before_validation_label_window or after_validation_embargo:
+                purged_train.append(idx)
+        if purged_train:
+            yield np.array(purged_train, dtype=int), np.array(val_idx, dtype=int)
+
+
+def _build_model():
+    return GradientBoostingClassifier(
+        n_estimators=300,
+        learning_rate=0.05,
+        max_depth=4,
+        subsample=0.8,
+        min_samples_leaf=20,
+        random_state=42,
+    )
+
+
+def derive_metric_weights(feature_cols, importances):
+    """Aggregate feature importances back to transparent score metric weights."""
+    metric_names = list(SCORE_KEYS)
+    feature_to_metric = {}
+    for m in metric_names:
+        feature_to_metric[f"score_{m}"] = m
+        feature_to_metric[f"raw_{m}"] = m
+    feature_to_metric["delta_30d_mvrv_zscore"] = "mvrv_zscore"
+    feature_to_metric["delta_30d_nupl"] = "nupl"
+    feature_to_metric["delta_30d_puell_multiple"] = "puell_multiple"
+    feature_to_metric["delta_30d_reserve_risk"] = "reserve_risk"
+
+    metric_importances = {m: 0.0 for m in metric_names}
+    for name, imp in zip(feature_cols, importances):
+        if name in feature_to_metric:
+            metric_importances[feature_to_metric[name]] += float(imp)
+        elif name == "mvrv_x_nupl":
+            metric_importances["mvrv_zscore"] += float(imp) / 2
+            metric_importances["nupl"] += float(imp) / 2
+        elif name == "puell_x_reserve":
+            metric_importances["puell_multiple"] += float(imp) / 2
+            metric_importances["reserve_risk"] += float(imp) / 2
+        elif name == "days_since_ath":
+            metric_importances["drawdown"] += float(imp)
+
+    total_imp = sum(metric_importances.values())
+    if total_imp > 0:
+        weights = {k: round(v / total_imp, 4) for k, v in metric_importances.items()}
+    else:
+        weights = {k: round(1 / len(metric_importances), 4) for k in metric_importances}
+    return dict(sorted(weights.items(), key=lambda x: x[1], reverse=True))
+
+
 def build_dataset(index, thresholds):
     """Build aligned training dataset: metric scores + forward returns."""
     # Get all dates from 2018-02-01 onward
@@ -257,39 +366,33 @@ def train_model(rows):
     log.info("Target distribution: %d positive (%.1f%%), %d negative",
              positive, positive / len(labeled) * 100, len(labeled) - positive)
 
-    # Feature columns: scores + raw values + deltas + interactions + cycle position
-    score_features = [
-        "score_puell_multiple", "score_mvrv_zscore", "score_reserve_risk",
-        "score_rhodl_ratio", "score_nupl", "score_fear_greed",
-        "score_drawdown", "score_pct_above_200w_sma", "score_pct_above_lth_rp",
-    ]
-    raw_features = [
-        "raw_puell_multiple", "raw_mvrv_zscore", "raw_reserve_risk",
-        "raw_rhodl_ratio", "raw_nupl", "raw_fear_greed",
-        "raw_pct_above_200w_sma", "raw_pct_above_lth_rp", "raw_drawdown",
-    ]
-    delta_features = [
-        "delta_30d_mvrv_zscore", "delta_30d_nupl",
-        "delta_30d_puell_multiple", "delta_30d_reserve_risk",
-    ]
-    interaction_features = ["mvrv_x_nupl", "puell_x_reserve"]
-    cycle_features = ["days_since_ath"]
-
-    feature_cols = score_features + raw_features + delta_features + interaction_features + cycle_features
+    feature_cols = FEATURE_COLS
 
     X = np.array([[r[f] for f in feature_cols] for r in labeled])
     y = np.array([r["target"] for r in labeled])
 
     log.info("Feature matrix: %d samples x %d features", X.shape[0], X.shape[1])
 
-    # Time-series cross-validation (expanding window, 5 splits)
-    tscv = TimeSeriesSplit(n_splits=5)
+    # Purged time-series cross-validation. Standard TimeSeriesSplit is not
+    # enough here because each label consumes the next 365 days of returns.
     cv_scores = []
     cv_f1 = []
     cv_precision = []
     cv_recall = []
+    fold_results = []
 
-    for fold, (train_idx, val_idx) in enumerate(tscv.split(X)):
+    splits = list(purged_time_series_splits(
+        labeled,
+        n_splits=VALIDATION_SPLITS,
+        label_horizon_days=LABEL_HORIZON_DAYS,
+        embargo_days=0,
+    ))
+    if not splits:
+        log.error("No viable purged validation splits. Need more history for %dd label horizon.",
+                  LABEL_HORIZON_DAYS)
+        return None
+
+    for fold, (train_idx, val_idx) in enumerate(splits):
         X_train, X_val = X[train_idx], X[val_idx]
         y_train, y_val = y[train_idx], y[val_idx]
 
@@ -297,14 +400,7 @@ def train_model(rows):
         X_train_s = scaler.fit_transform(X_train)
         X_val_s = scaler.transform(X_val)
 
-        model = GradientBoostingClassifier(
-            n_estimators=300,
-            learning_rate=0.05,
-            max_depth=4,
-            subsample=0.8,
-            min_samples_leaf=20,
-            random_state=42,
-        )
+        model = _build_model()
         model.fit(X_train_s, y_train)
 
         y_pred = model.predict(X_val_s)
@@ -320,27 +416,40 @@ def train_model(rows):
         cv_precision.append(prec)
         cv_recall.append(rec)
 
-        train_dates = f"{labeled[train_idx[0]]['date']} to {labeled[train_idx[-1]]['date']}"
-        val_dates = f"{labeled[val_idx[0]]['date']} to {labeled[val_idx[-1]]['date']}"
+        fold_weights = derive_metric_weights(feature_cols, model.feature_importances_)
+        fold_results.append({
+            "fold": fold + 1,
+            "train_idx": train_idx.tolist(),
+            "val_idx": val_idx.tolist(),
+            "weights": fold_weights,
+            "metrics": {
+                "auc": round(float(auc), 4),
+                "f1": round(float(f1), 4),
+                "precision": round(float(prec), 4),
+                "recall": round(float(rec), 4),
+            },
+            "date_ranges": {
+                "train": f"{labeled[train_idx[0]]['date']} to {labeled[train_idx[-1]]['date']}",
+                "validation": f"{labeled[val_idx[0]]['date']} to {labeled[val_idx[-1]]['date']}",
+            },
+            "n_train": len(train_idx),
+            "n_validation": len(val_idx),
+        })
+
+        train_dates = fold_results[-1]["date_ranges"]["train"]
+        val_dates = fold_results[-1]["date_ranges"]["validation"]
         log.info("Fold %d: Train %s | Val %s | AUC=%.3f F1=%.3f P=%.3f R=%.3f",
                  fold + 1, train_dates, val_dates, auc, f1, prec, rec)
 
-    log.info("CV Mean AUC: %.3f (+/- %.3f)", np.mean(cv_scores), np.std(cv_scores))
-    log.info("CV Mean F1:  %.3f (+/- %.3f)", np.mean(cv_f1), np.std(cv_f1))
+    log.info("Purged CV Mean AUC: %.3f (+/- %.3f)", np.mean(cv_scores), np.std(cv_scores))
+    log.info("Purged CV Mean F1:  %.3f (+/- %.3f)", np.mean(cv_f1), np.std(cv_f1))
 
     # Train final model on all labeled data
     log.info("Training final model on all %d labeled samples...", len(labeled))
     scaler = StandardScaler()
     X_scaled = scaler.fit_transform(X)
 
-    final_model = GradientBoostingClassifier(
-        n_estimators=300,
-        learning_rate=0.05,
-        max_depth=4,
-        subsample=0.8,
-        min_samples_leaf=20,
-        random_state=42,
-    )
+    final_model = _build_model()
     final_model.fit(X_scaled, y)
 
     # Feature importances
@@ -357,48 +466,7 @@ def train_model(rows):
         bar = "#" * int(imp * 200)
         log.info("  %-30s %.4f %s", name, imp, bar)
 
-    # Extract optimal weights by aggregating importance per metric
-    # Map each feature back to its parent metric
-    metric_names = [
-        "puell_multiple", "mvrv_zscore", "reserve_risk", "rhodl_ratio",
-        "nupl", "fear_greed", "drawdown", "pct_above_200w_sma", "pct_above_lth_rp",
-    ]
-    feature_to_metric = {}
-    for m in metric_names:
-        feature_to_metric[f"score_{m}"] = m
-        feature_to_metric[f"raw_{m}"] = m
-    # Delta features map to their base metric
-    feature_to_metric["delta_30d_mvrv_zscore"] = "mvrv_zscore"
-    feature_to_metric["delta_30d_nupl"] = "nupl"
-    feature_to_metric["delta_30d_puell_multiple"] = "puell_multiple"
-    feature_to_metric["delta_30d_reserve_risk"] = "reserve_risk"
-    # Interaction terms split evenly between constituent metrics
-    # mvrv_x_nupl -> mvrv_zscore + nupl
-    # puell_x_reserve -> puell_multiple + reserve_risk
-
-    metric_importances = {m: 0.0 for m in metric_names}
-    for name, imp in feat_imp:
-        if name in feature_to_metric:
-            metric_importances[feature_to_metric[name]] += imp
-        elif name == "mvrv_x_nupl":
-            metric_importances["mvrv_zscore"] += imp / 2
-            metric_importances["nupl"] += imp / 2
-        elif name == "puell_x_reserve":
-            metric_importances["puell_multiple"] += imp / 2
-            metric_importances["reserve_risk"] += imp / 2
-        # days_since_ath maps to drawdown conceptually
-        elif name == "days_since_ath":
-            metric_importances["drawdown"] += imp
-
-    # Normalize weights to sum to 1
-    total_imp = sum(metric_importances.values())
-    if total_imp > 0:
-        weights = {k: round(v / total_imp, 4) for k, v in metric_importances.items()}
-    else:
-        weights = {k: round(1 / len(metric_importances), 4) for k in metric_importances}
-
-    # Sort by weight descending
-    weights = dict(sorted(weights.items(), key=lambda x: x[1], reverse=True))
+    weights = derive_metric_weights(feature_cols, importances)
 
     log.info("\nOptimal Metric Weights:")
     log.info("-" * 50)
@@ -413,6 +481,7 @@ def train_model(rows):
     log.info("COMPARISON BACKTEST: ML-Weighted vs Equal-Weight")
     log.info("=" * 60)
     comparison = run_comparison(rows, weights)
+    out_of_sample_comparison = run_out_of_sample_comparison(labeled, fold_results)
 
     # Build output
     result = {
@@ -424,6 +493,9 @@ def train_model(rows):
             "mean_f1": round(float(np.mean(cv_f1)), 4),
             "mean_precision": round(float(np.mean(cv_precision)), 4),
             "mean_recall": round(float(np.mean(cv_recall)), 4),
+            "validation_method": "purged_expanding_window",
+            "label_horizon_days": LABEL_HORIZON_DAYS,
+            "folds": fold_results,
         },
         "training_info": {
             "n_samples": len(labeled),
@@ -435,66 +507,47 @@ def train_model(rows):
             "model": "GradientBoostingClassifier",
         },
         "comparison": comparison,
+        "out_of_sample_comparison": out_of_sample_comparison,
         "trained_at": datetime.now(tz=__import__('datetime').timezone.utc).isoformat(),
     }
 
     return result
 
 
-def run_comparison(rows, ml_weights):
-    """Compare ML-weighted scoring vs equal-weight scoring across score brackets."""
-    # Metrics used in scoring (maps to score_* columns)
-    score_keys = [
-        "puell_multiple", "mvrv_zscore", "reserve_risk", "rhodl_ratio",
-        "nupl", "fear_greed", "drawdown", "pct_above_200w_sma", "pct_above_lth_rp",
-    ]
-    n_metrics = len(score_keys)
-    equal_weight = 1.0 / n_metrics
+def _composite_score(row, mode, ml_weights=None):
+    scores = [row[f"score_{k}"] for k in SCORE_KEYS]
+    if mode == "equal_weight" or not ml_weights:
+        return sum(scores) / len(SCORE_KEYS) * 10
+    equal_weight = 1.0 / len(SCORE_KEYS)
+    weighted_sum = sum(row[f"score_{k}"] * ml_weights.get(k, equal_weight) for k in SCORE_KEYS)
+    return weighted_sum * 10
 
-    brackets = [
-        (0, 20, "Extreme Caution"),
-        (21, 40, "Caution"),
-        (41, 55, "Neutral"),
-        (56, 70, "Moderate Opportunity"),
-        (71, 85, "Strong Accumulation"),
-        (86, 100, "Extreme Accumulation"),
-    ]
 
-    # Only use rows with forward returns
-    scored_rows = [r for r in rows if "fwd_365d" in r]
-
-    results = {"equal_weight": [], "ml_weighted": []}
-
-    for mode in ["equal_weight", "ml_weighted"]:
-        for r in scored_rows:
-            scores = [r[f"score_{k}"] for k in score_keys]
-            if mode == "equal_weight":
-                composite = sum(scores) / n_metrics * 10
-            else:
-                weighted_sum = sum(r[f"score_{k}"] * ml_weights.get(k, equal_weight) for k in score_keys)
-                composite = weighted_sum * 10
-            r[f"composite_{mode}"] = composite
-
-        for low, high, label in brackets:
-            days_in = [r for r in scored_rows if low <= r[f"composite_{mode}"] <= high]
-            if not days_in:
-                results[mode].append({
-                    "range": f"{low}-{high}", "label": label,
-                    "days": 0, "avg_365d": None,
-                })
-                continue
-            returns_365 = [r["fwd_365d"] for r in days_in]
-            win_rate = len([r for r in returns_365 if r > 0]) / len(returns_365) * 100
-            results[mode].append({
-                "range": f"{low}-{high}",
-                "label": label,
-                "days": len(days_in),
-                "avg_365d": round(sum(returns_365) / len(returns_365), 2),
-                "median_365d": round(sorted(returns_365)[len(returns_365) // 2], 2),
-                "win_rate_365d": round(win_rate, 1),
+def _summarize_brackets(scored_rows, score_key):
+    results = []
+    for low, high, label in BRACKETS:
+        days_in = [r for r in scored_rows if low <= r[score_key] <= high]
+        if not days_in:
+            results.append({
+                "range": f"{low}-{high}", "label": label,
+                "days": 0, "avg_365d": None,
             })
+            continue
+        returns_365 = [r["fwd_365d"] for r in days_in]
+        returns_sorted = sorted(returns_365)
+        win_rate = len([r for r in returns_365 if r > 0]) / len(returns_365) * 100
+        results.append({
+            "range": f"{low}-{high}",
+            "label": label,
+            "days": len(days_in),
+            "avg_365d": round(sum(returns_365) / len(returns_365), 2),
+            "median_365d": round(returns_sorted[len(returns_sorted) // 2], 2),
+            "win_rate_365d": round(win_rate, 1),
+        })
+    return results
 
-    # Print comparison
+
+def _log_comparison_table(results):
     log.info("\n%-18s | %-8s %-8s %-8s | %-8s %-8s %-8s",
              "Bracket", "EQ Avg", "EQ Med", "EQ Win%", "ML Avg", "ML Med", "ML Win%")
     log.info("-" * 80)
@@ -508,6 +561,47 @@ def run_comparison(rows, ml_weights):
         log.info("%-18s | %-8s %-8s %-8s | %-8s %-8s %-8s",
                  eq["label"], eq_avg, eq_med, eq_win, ml_avg, ml_med, ml_win)
 
+
+def run_comparison(rows, ml_weights):
+    """Compare final ML-weighted scoring vs equal-weight scoring across all labeled rows.
+
+    This is retained for backwards compatibility with existing output. It is an
+    in-sample/full-history comparison; prefer out_of_sample_comparison for model
+    selection decisions.
+    """
+    scored_rows = [dict(r) for r in rows if "fwd_365d" in r]
+    for r in scored_rows:
+        r["composite_equal_weight"] = _composite_score(r, "equal_weight")
+        r["composite_ml_weighted"] = _composite_score(r, "ml_weighted", ml_weights)
+
+    results = {
+        "equal_weight": _summarize_brackets(scored_rows, "composite_equal_weight"),
+        "ml_weighted": _summarize_brackets(scored_rows, "composite_ml_weighted"),
+    }
+    _log_comparison_table(results)
+    return results
+
+
+def run_out_of_sample_comparison(rows, fold_results):
+    """Compare fold-specific ML weights on validation rows only."""
+    validation_rows = []
+    for fold in fold_results:
+        weights = fold.get("weights", {})
+        for idx in fold.get("val_idx", []):
+            if idx >= len(rows) or "fwd_365d" not in rows[idx]:
+                continue
+            r = dict(rows[idx])
+            r["fold"] = fold.get("fold")
+            r["composite_equal_weight"] = _composite_score(r, "equal_weight")
+            r["composite_ml_weighted"] = _composite_score(r, "ml_weighted", weights)
+            validation_rows.append(r)
+
+    results = {
+        "folds": len(fold_results),
+        "validation_days": len(validation_rows),
+        "equal_weight": _summarize_brackets(validation_rows, "composite_equal_weight"),
+        "ml_weighted": _summarize_brackets(validation_rows, "composite_ml_weighted"),
+    }
     return results
 
 

scoring/engine.py

@@ -259,6 +259,65 @@ def score_hash_ribbons(data, thresholds=None):
     return 3, "Normal mining activity"
 
 
+def score_sopr(value, thresholds=None):
+    if value is None:
+        return None, "No data"
+    if value < 0.98:
+        return 10, "Deep loss realization — capitulation, strong accumulation"
+    if value < 1.0:
+        return 8, "Below breakeven — capitulation, good accumulation"
+    if value < 1.02:
+        return 5, "Near breakeven — neutral"
+    if value < 1.05:
+        return 2, "Moderate profit taking"
+    return 0, "Elevated profit taking — caution"
+
+
+def score_sellside_risk(value, thresholds=None):
+    if value is None:
+        return None, "No data"
+    if value < 0.001:
+        return 10, "Very low sell-side risk — strong accumulation"
+    if value < 0.002:
+        return 8, "Low sell-side risk — good accumulation"
+    if value < 0.005:
+        return 5, "Moderate sell-side risk"
+    if value < 0.01:
+        return 2, "Elevated sell-side risk"
+    return 0, "High sell-side risk"
+
+
+def score_momentum_pct(value):
+    if value is None:
+        return None, "No data"
+    pct = value * 100
+    if pct >= 20:
+        return 10, f"Strong positive momentum (+{pct:.0f}%)"
+    if pct >= 0:
+        return 6, f"Mild positive momentum (+{pct:.0f}%)"
+    if pct >= -10:
+        return 4, f"Slightly negative momentum ({pct:.0f}%)"
+    if pct >= -25:
+        return 2, f"Weak momentum ({pct:.0f}%)"
+    return 1, f"Strong negative momentum ({pct:.0f}%)"
+
+
+def score_nvt_price(nvt_price, spot_price):
+    if nvt_price is None or spot_price is None or spot_price <= 0:
+        return None, "No data"
+    premium = (nvt_price - spot_price) / spot_price * 100
+    if premium < -25:
+        return 10, f"NVT price {abs(premium):.0f}% below spot — deep value"
+    if premium < -10:
+        return 8, f"NVT price {abs(premium):.0f}% below spot — undervalued"
+    if premium < 10:
+        relation = "below" if premium < 0 else "above"
+        return 5, f"NVT price {abs(premium):.0f}% {relation} spot — fair value"
+    if premium < 30:
+        return 2, f"NVT price {premium:.0f}% above spot — extended"
+    return 0, f"NVT price {premium:.0f}% above spot — overheated"
+
+
 def score_all(metrics):
     """Score all metrics and return individual + composite scores."""
     thresholds = load_thresholds()
@@ -399,6 +458,84 @@ def score_all(metrics):
         "recent": [],
     })
 
+    # SOPR
+    sopr = metrics.get("sopr", {})
+    sopr_score, sopr_desc = score_sopr(sopr.get("value"), thresholds)
+    results.append({
+        "name": "SOPR",
+        "key": "sopr",
+        "value": sopr.get("value"),
+        "display_value": f"{sopr.get('value', 'N/A'):.4f}" if sopr.get("value") is not None else "N/A",
+        "score": sopr_score,
+        "description": sopr_desc,
+        "recent": sopr.get("recent", []),
+    })
+
+    # Sell-side Risk Ratio
+    ssr = metrics.get("sellside_risk", {})
+    ssr_score, ssr_desc = score_sellside_risk(ssr.get("value"), thresholds)
+    results.append({
+        "name": "Sell-side Risk Ratio",
+        "key": "sellside_risk",
+        "value": ssr.get("value"),
+        "display_value": f"{ssr.get('value', 'N/A'):.6f}" if ssr.get("value") is not None else "N/A",
+        "score": ssr_score,
+        "description": ssr_desc,
+        "recent": ssr.get("recent", []),
+    })
+
+    # Active Address Momentum
+    aam = metrics.get("active_address_momentum", {})
+    aam_score, aam_desc = score_momentum_pct(aam.get("value"))
+    results.append({
+        "name": "Active Address Momentum",
+        "key": "active_address_momentum",
+        "value": aam.get("value"),
+        "display_value": f"{aam.get('value') * 100:.1f}%" if aam.get("value") is not None else "N/A",
+        "score": aam_score,
+        "description": aam_desc,
+        "recent": aam.get("recent", []),
+    })
+
+    # Transaction Count Momentum
+    txm = metrics.get("txcount_momentum", {})
+    txm_score, txm_desc = score_momentum_pct(txm.get("value"))
+    results.append({
+        "name": "Transaction Count Momentum",
+        "key": "txcount_momentum",
+        "value": txm.get("value"),
+        "display_value": f"{txm.get('value') * 100:.1f}%" if txm.get("value") is not None else "N/A",
+        "score": txm_score,
+        "description": txm_desc,
+        "recent": txm.get("recent", []),
+    })
+
+    # NVT Price
+    nvt = metrics.get("nvt_price", {})
+    nvt_score, nvt_desc = score_nvt_price(nvt.get("value"), current_price)
+    results.append({
+        "name": "NVT Price",
+        "key": "nvt_price",
+        "value": nvt.get("value"),
+        "display_value": f"${nvt.get('value'):,.0f}" if nvt.get("value") is not None else "N/A",
+        "score": nvt_score,
+        "description": nvt_desc,
+        "recent": nvt.get("recent", []),
+    })
+
+    # VDD Multiple
+    vdd = metrics.get("vdd_multiple", {})
+    vdd_score, vdd_desc = score_momentum_pct(vdd.get("value"))
+    results.append({
+        "name": "VDD Multiple",
+        "key": "vdd_multiple",
+        "value": vdd.get("value"),
+        "display_value": f"{vdd.get('value') * 100:.1f}%" if vdd.get("value") is not None else "N/A",
+        "score": vdd_score,
+        "description": vdd_desc,
+        "recent": vdd.get("recent", []),
+    })
+
     # Compute composite
     valid_scores = [r["score"] for r in results if r["score"] is not None]
     if valid_scores:
@@ -481,31 +618,34 @@ def score_all_ml(metrics):
 
     results = classic["metrics"]
 
-    # Compute ML-weighted composite
-    weighted_sum = 0.0
-    weight_total = 0.0
-
+    # Compute raw ML weights first, then normalize across only the currently
+    # scored metrics. This keeps the dashboard's displayed per-metric weights and
+    # contribution points consistent with the normalized composite score even
+    # when optional metrics are missing or hash ribbons receives its fallback.
+    weighted_metrics = []
     for m in results:
         if m["score"] is None:
             continue
         ml_key = _ML_KEY_MAP.get(m["key"])
         if ml_key is None:
             # Hash ribbons or unknown metric — use small default weight
-            w = 0.01
+            raw_weight = 0.01
         else:
-            w = ml_weights.get(ml_key, 0.0)
+            raw_weight = ml_weights.get(ml_key, 0.0)
+        weighted_metrics.append((m, raw_weight))
 
-        m["ml_weight"] = round(w, 4)
-        m["ml_contribution"] = round(m["score"] * w * 10, 2)
-        weighted_sum += m["score"] * w
-        weight_total += w
-
-    # Normalize if weights don't sum to 1 (e.g., missing metrics)
+    weight_total = sum(raw_weight for _, raw_weight in weighted_metrics)
     if weight_total > 0:
-        composite = weighted_sum / weight_total * 10
+        composite = sum(m["score"] * raw_weight for m, raw_weight in weighted_metrics) / weight_total * 10
     else:
         composite = 0
 
+    for m, raw_weight in weighted_metrics:
+        effective_weight = raw_weight / weight_total if weight_total > 0 else 0.0
+        m["ml_raw_weight"] = round(raw_weight, 4)
+        m["ml_weight"] = round(effective_weight, 4)
+        m["ml_contribution"] = round(m["score"] * effective_weight * 10, 2)
+
     # Assessment text (same thresholds as classic)
     if composite >= 80:
         assessment = "EXTREME ACCUMULATION ZONE"
@@ -528,4 +668,5 @@ def score_all_ml(metrics):
         "total_count": classic["total_count"],
         "ml_mode": True,
         "classic_score": classic["composite_score"],
+        "ml_weight_total": round(weight_total, 4),
     }

scrapers/checkonchain.py

@@ -0,0 +1,170 @@
+"""Scraper for static CheckOnChain Plotly chart HTML pages."""
+
+from __future__ import annotations
+
+import array
+import base64
+import json
+import logging
+import re
+from html import unescape
+
+import requests
+
+log = logging.getLogger(__name__)
+
+CHARTS = {
+    "sopr": {
+        "url": "https://charts.checkonchain.com/btconchain/realised/sopr/sopr_light.html",
+        "traces": ["SOPR"],
+    },
+    "sellside_risk": {
+        "url": "https://charts.checkonchain.com/btconchain/realised/sellsideriskratio_all/sellsideriskratio_all_light.html",
+        "traces": ["Sell-side Risk Ratio", "Sellside Risk Ratio", "SSR"],
+    },
+    "active_address_momentum": {
+        "url": "https://charts.checkonchain.com/btconchain/adoption/actaddress_momentum/actaddress_momentum_light.html",
+        "traces": ["30DMA", "30 Day", "Active Address"],
+    },
+    "txcount_momentum": {
+        "url": "https://charts.checkonchain.com/btconchain/adoption/txcount_momentum/txcount_momentum_light.html",
+        "traces": ["30DMA", "30 Day", "Transaction"],
+    },
+    "nvt_price": {
+        "url": "https://charts.checkonchain.com/btconchain/pricing/pricing_nvtprice/pricing_nvtprice_light.html",
+        "traces": ["NVT Price", "NVT"],
+    },
+    "vdd_multiple": {
+        "url": "https://charts.checkonchain.com/btconchain/lifespan/vddmultiple/vddmultiple_light.html",
+        "traces": ["VDD Multiple", "Value Days Destroyed"],
+    },
+}
+
+
+def _extract_plotly_traces(html_text: str):
+    """Extract first Plotly.newPlot trace array from a static Plotly HTML page."""
+    marker = "Plotly.newPlot("
+    start = html_text.find(marker)
+    if start < 0:
+        return []
+    first_array = html_text.find("[", start)
+    if first_array < 0:
+        return []
+
+    depth = 0
+    in_string = False
+    escape = False
+    quote = ""
+    for idx in range(first_array, len(html_text)):
+        ch = html_text[idx]
+        if in_string:
+            if escape:
+                escape = False
+            elif ch == "\\":
+                escape = True
+            elif ch == quote:
+                in_string = False
+            continue
+        if ch in {'"', "'"}:
+            in_string = True
+            quote = ch
+        elif ch == "[":
+            depth += 1
+        elif ch == "]":
+            depth -= 1
+            if depth == 0:
+                raw = html_text[first_array:idx + 1]
+                return json.loads(raw)
+    return []
+
+
+def scrape_chart(url: str, timeout=30):
+    resp = requests.get(url, headers={"User-Agent": "Mozilla/5.0"}, timeout=timeout)
+    resp.raise_for_status()
+    return _extract_plotly_traces(unescape(resp.text))
+
+
+def _find_trace(traces, names):
+    names = [n.lower() for n in names if n]
+    # Prefer non-price traces with the requested terms.
+    for trace in traces:
+        trace_name = str(trace.get("name", "")).lower()
+        if "price" in trace_name and not any("price" in n for n in names):
+            continue
+        if any(n in trace_name for n in names):
+            return trace
+    # Fallback: first numeric non-price trace.
+    for trace in traces:
+        trace_name = str(trace.get("name", "")).lower()
+        if "price" in trace_name:
+            continue
+        y = trace.get("y") or []
+        if any(v is not None for v in y[-30:]):
+            return trace
+    return None
+
+
+def _decode_plotly_array(values):
+    """Decode Plotly typed-array JSON ({dtype, bdata}) or return plain values."""
+    if not isinstance(values, dict) or "bdata" not in values:
+        return values or []
+
+    dtype = values.get("dtype")
+    typecodes = {
+        "f8": "d", "float64": "d",
+        "f4": "f", "float32": "f",
+        "i8": "q", "int64": "q",
+        "i4": "i", "int32": "i",
+        "u8": "Q", "uint64": "Q",
+        "u4": "I", "uint32": "I",
+    }
+    typecode = typecodes.get(dtype)
+    if not typecode:
+        return []
+    decoded = base64.b64decode(values["bdata"])
+    arr = array.array(typecode)
+    arr.frombytes(decoded)
+    if values.get("byteorder") == "big":
+        arr.byteswap()
+    return arr.tolist()
+
+
+def _numeric_values(trace):
+    values = []
+    for value in _decode_plotly_array((trace or {}).get("y", [])):
+        if value is None:
+            continue
+        try:
+            values.append(float(value))
+        except (TypeError, ValueError):
+            pass
+    return values
+
+
+def _latest(values):
+    return values[-1] if values else None
+
+
+def _momentum(values, window=30):
+    if len(values) <= window or values[-window] == 0:
+        return None
+    return (values[-1] - values[-window]) / values[-window]
+
+
+def scrape_all():
+    results = {}
+    for key, cfg in CHARTS.items():
+        log.info("Scraping CheckOnChain %s ...", key)
+        try:
+            traces = scrape_chart(cfg["url"])
+            trace = _find_trace(traces, cfg.get("traces", []))
+            values = _numeric_values(trace)
+            value = _latest(values)
+            if key in {"active_address_momentum", "txcount_momentum", "vdd_multiple"}:
+                # The card value is momentum, while the sparkline shows the raw metric.
+                value = _momentum(values)
+            results[key] = {"value": value, "recent": values[-30:]}
+        except Exception as exc:
+            log.error("CheckOnChain scrape failed for %s: %s", key, exc)
+            results[key] = {"value": None, "error": str(exc)}
+    return results

tests/test_ml_optimizer_validation.py

@@ -0,0 +1,74 @@
+from datetime import datetime, timedelta
+
+from ml import optimizer
+
+
+def _row(date, returns=10.0, score_200w=10, score_drawdown=0):
+    row = {
+        "date": date,
+        "price": 100.0,
+        "fwd_365d": returns,
+        "score_puell_multiple": 0,
+        "score_mvrv_zscore": 0,
+        "score_reserve_risk": 0,
+        "score_rhodl_ratio": 0,
+        "score_nupl": 0,
+        "score_fear_greed": 0,
+        "score_drawdown": score_drawdown,
+        "score_pct_above_200w_sma": score_200w,
+        "score_pct_above_lth_rp": 0,
+    }
+    return row
+
+
+def test_purged_time_series_splits_remove_overlapping_forward_label_windows():
+    start = datetime(2020, 1, 1)
+    rows = [_row((start + timedelta(days=i)).strftime("%Y-%m-%d")) for i in range(900)]
+
+    splits = list(
+        optimizer.purged_time_series_splits(
+            rows,
+            n_splits=3,
+            label_horizon_days=365,
+            embargo_days=0,
+        )
+    )
+
+    assert splits, "expected at least one viable split"
+    for train_idx, val_idx in splits:
+        val_start = datetime.strptime(rows[val_idx[0]]["date"], "%Y-%m-%d")
+        latest_allowed_train_date = val_start - timedelta(days=365)
+        assert len(train_idx) > 0, "purging should keep non-overlapping expanding-window training rows"
+        for idx in train_idx:
+            train_date = datetime.strptime(rows[idx]["date"], "%Y-%m-%d")
+            assert train_date <= latest_allowed_train_date
+
+
+def test_run_out_of_sample_comparison_scores_only_validation_rows_with_fold_weights():
+    rows = [
+        _row("2020-01-01", returns=-10, score_200w=0, score_drawdown=10),
+        _row("2020-01-02", returns=-5, score_200w=0, score_drawdown=10),
+        _row("2020-01-03", returns=100, score_200w=10, score_drawdown=10),
+        _row("2020-01-04", returns=120, score_200w=10, score_drawdown=10),
+    ]
+    fold_results = [
+        {
+            "fold": 1,
+            "val_idx": [2, 3],
+            "weights": {"pct_above_200w_sma": 1.0, "drawdown": 0.0},
+        }
+    ]
+
+    comparison = optimizer.run_out_of_sample_comparison(rows, fold_results)
+
+    assert comparison["validation_days"] == 2
+    assert comparison["folds"] == 1
+    assert sum(bucket["days"] for bucket in comparison["ml_weighted"]) == 2
+    assert sum(bucket["days"] for bucket in comparison["equal_weight"]) == 2
+
+    extreme_ml = next(bucket for bucket in comparison["ml_weighted"] if bucket["label"] == "Extreme Accumulation")
+    assert extreme_ml["days"] == 2
+    assert extreme_ml["avg_365d"] == 110.0
+
+    caution_equal = next(bucket for bucket in comparison["equal_weight"] if bucket["label"] == "Caution")
+    assert caution_equal["days"] == 2

tests/test_scoring_engine_ml.py

@@ -0,0 +1,63 @@
+import math
+
+from scoring import engine
+
+
+def _complete_metrics():
+    return {
+        "fear_greed": {"value": 10, "classification": "Extreme Fear"},
+        "puell_multiple": {"value": 0.3},
+        "mvrv_zscore": {"value": -0.1},
+        "drawdown": {"value": 60.0, "ath": 250.0},
+        "price": {"price": 100.0},
+        "200w_sma": {"value": 120.0},
+        "reserve_risk": {"value": 0.001},
+        "rhodl_ratio": {"value": 50.0},
+        "nupl": {"value": -0.1},
+        "lth_realized_price": {"value": 120.0},
+        "hash_ribbons": {"buy_signal": False},
+    }
+
+
+def test_score_all_ml_normalizes_displayed_weights_and_contributions(monkeypatch):
+    monkeypatch.setattr(
+        engine,
+        "load_ml_weights",
+        lambda: {
+            "fear_greed": 0.40,
+            "puell_multiple": 0.20,
+            "mvrv_zscore": 0.15,
+            "drawdown": 0.10,
+            "pct_above_200w_sma": 0.05,
+            "reserve_risk": 0.04,
+            "rhodl_ratio": 0.03,
+            "nupl": 0.02,
+            "pct_above_lth_rp": 0.01,
+        },
+    )
+
+    scored = engine.score_all_ml(_complete_metrics())
+
+    assert scored["ml_mode"] is True
+    valid_metrics = [m for m in scored["metrics"] if m["score"] is not None]
+    assert scored["ml_weight_total"] == 1.01  # trained weights + small hash-ribbons fallback
+    assert math.isclose(sum(m["ml_weight"] for m in valid_metrics), 1.0, abs_tol=0.001)
+    assert math.isclose(
+        sum(m["ml_contribution"] for m in valid_metrics),
+        scored["composite_score"],
+        abs_tol=0.05,
+    )
+
+    hash_ribbons = next(m for m in valid_metrics if m["key"] == "hash_ribbons")
+    assert hash_ribbons["ml_raw_weight"] == 0.01
+    assert hash_ribbons["ml_weight"] == round(0.01 / 1.01, 4)
+
+
+def test_score_all_ml_preserves_classic_fallback_when_weights_missing(monkeypatch):
+    monkeypatch.setattr(engine, "load_ml_weights", lambda: {})
+
+    scored = engine.score_all_ml(_complete_metrics())
+
+    assert scored["ml_mode"] is False
+    assert scored["ml_error"] == "ML weights not found — run ml/optimizer.py"
+    assert "classic_score" not in scored