tradeblocks-mcp 4.3.0 → 4.5.0

package/README.md

@@ -549,6 +549,63 @@ Field-based `suggest_filters` retains all-trade `currentStats` and reports per-f
 Suggestions compare against those covered populations; `filter_curve.baseline`
 reports the same covered and uncovered counts and P&L.
 
+New in 4.4.0: [filter statistics](../../releases/v4.4.0.md) for asking whether a
+chosen filter survives a full threshold search, with explicit assumptions and work caps.
+
+`find_predictive_fields.method` selects `pearson` (default), `spearman`, or `kendall`
+(tau-b with ties). Explicit requests name their method; omission preserves the Pearson
+response. Ranking uses the chosen coefficient on the same causal entry-time values,
+including `maxMissedSessions` refusals.
+Kendall admits up to 75 million estimated pairs across fields, using each field's
+valid field/target coverage after strategy/date filtering. It refuses by name
+(`kendall_pair_work_limit`) before quadratic calculations. A 1,100-trade block with
+100 fully covered fields passes (60,445,000 pairs); a 20,000-trade field is refused.
+Pearson and Spearman retain their existing behavior.
+
+`filter_curve` accepts either the existing `field` or up to 10 distinct `fields`.
+Multi-field calls, an explicit `objective` (`totalPl` or `winRate`), or opt-in
+`permutation: { iterations: 499, block: "month", seed: 42 }` add the best threshold
+and per-year kept/removed stability. Permutation shuffles whole calendar-month outcome
+blocks and reruns the complete threshold sweep, not just the selected threshold.
+It reports add-one raw p-values and Holm-adjusted p-values across requested fields.
+Single-field calls without these options retain their old outputs.
+At least 30 covered trades and 8 observed months are required; refusals are named.
+Iterations are capped at 499. Only permutation requests limit threshold and percentile
+arrays to 100 entries, with a named `permutation_candidate_limit` refusal.
+Without permutation, larger arrays retain their previous schema acceptance.
+Sort-aware work admission caps permutation at 30 million units per field and
+90 million per request, refusing the whole request before any sweeps. At 1,100 trades,
+55 months, 499 iterations and default thresholds, the largest admitted family is
+3 fully covered fields (87,751,152 estimated units); 4 fields exceed the cap.
+Field, objective and mode choices across separate requests are outside the Holm family.
+Monthly blocks retain within-month dependence, not cross-month dependence; block
+exchangeability is an assumption. See [the full contract](../../docs/mcp-tools.md#filter-curve-significance-and-stability).
+
+In 4.5.0, `unit: "day"` aggregates entry-date P&L before scoring and month
+permutation; win rate counts profitable days. A field varying within a day is
+refused as `day_field_varies`, with the varying dates. Cheap request/work admission
+precedes causal field resolution and counts the chosen unit conservatively before
+coverage. Omitted `unit` retains trade-mode outputs.
+Day mode also offers `sharpe` (daily dollar P&L, rf=0, sample std, √252),
+`mar` (calendar-day CAGR / drawdown fraction of observed starting capital), and
+`maxDrawdown` (dollar drawdown reduction; smaller is better). Filtered dates are
+flat. Trade units refuse these objectives; MAR additionally refuses missing capital.
+`objectiveMetadata` documents definitions and undefined cases. This dollar-P&L
+Sharpe intentionally differs from the capital-based portfolio return Sharpe.
+`groupBy: "regime.label"` adds chosen-threshold `perGroup` rows in either unit,
+with the same covered baseline/kept/removed shape as `perYear`. Numeric groups use
+`bins` or shared regime-analysis quartiles; categorical groups use values.
+Missing values form an explicit uncovered group, small groups stay flagged,
+and within-day varying groups refuse as `day_group_varies`. Group values honor
+the same causal availability and optional freshness limit.
+
+Missing groups use `group: null`, `groupValue: null`, `uncoveredGroup: true`, distinct
+from a real `"uncovered"` category. Varchar market fields such as `Trend_Direction`
+are now categorical for filter suggestions and regime analysis. Explicit-field
+tools resolve only consumed fields without changing their timing or values. See the
+[4.5.0 release notes](../../releases/v4.5.0.md) for the timings, MAR capital/window
+policy and the under-90-day per-year/per-group `marUnstable` guard.
+
 `suggest_filters` defaults to `objective: "totalPl"`: only positive P&L improvements,
 ranked by P&L improvement with win-rate gain as tie-breaker. `objective: "winRate"`
 admits/ranks by win-rate gain with P&L tie-breaks. `minImprovementPct` (default 3

package/dist/{chunk-TLX7IBCM.js → chunk-6IUDE55E.js}

@@ -6,7 +6,7 @@ import {
   getNetPl,
   getRiskFreeRate,
   resolveTreasuryRateByKey
-} from "./chunk-ENXBPJ7Y.js";
+} from "./chunk-UOGG5TZQ.js";
 import {
   PlBasis
 } from "./chunk-BH6SQOWF.js";
@@ -3790,7 +3790,7 @@ async function openReadWriteConnection(dbPath, threads, memoryLimit) {
   await createMarketParquetViews(connection, dataRoot);
   await ensureMarketDataTables(connection);
   try {
-    const blocksDir = (await import("./sync-EW33QZPD.js")).getBlocksDir(dataRoot);
+    const blocksDir = (await import("./sync-R7BQBTN4.js")).getBlocksDir(dataRoot);
     await migrateMetadataToJson(connection, dataRoot, blocksDir);
   } catch (err) {
     console.warn(
@@ -20622,7 +20622,7 @@ function calculateCorrelationMatrix(trades, options = {}) {
       }
       let correlation;
       if (method === "pearson") {
-        correlation = pearsonCorrelation2(returns1, returns2);
+        correlation = pearsonCorrelation(returns1, returns2);
       } else if (method === "spearman") {
         correlation = spearmanCorrelation(returns1, returns2);
       } else {
@@ -20635,7 +20635,7 @@ function calculateCorrelationMatrix(trades, options = {}) {
   }
   return { strategies, correlationData, sampleSizes };
 }
-function pearsonCorrelation2(x, y) {
+function pearsonCorrelation(x, y) {
   if (x.length !== y.length || x.length === 0) return 0;
   const meanX = mean(x);
   const meanY = mean(y);
@@ -20657,7 +20657,7 @@ function spearmanCorrelation(x, y) {
   if (x.length !== y.length || x.length === 0) return 0;
   const rankX = getRanks(x);
   const rankY = getRanks(y);
-  return pearsonCorrelation2(rankX, rankY);
+  return pearsonCorrelation(rankX, rankY);
 }
 function kendallCorrelation(x, y) {
   if (x.length !== y.length || x.length === 0) return 0;
@@ -23270,4 +23270,4 @@ decimal.js/decimal.mjs:
    *  MIT Licence
    *)
 */
-//# sourceMappingURL=chunk-TLX7IBCM.js.map
+//# sourceMappingURL=chunk-6IUDE55E.js.map