fix(anomaly): exclude non-directional kinds from grading and backtest

Review follow-up (HIGH): the three detectors fed the same evaluator/backtest, but
SuspensionFreeze is non-directional (favourite unchanged) — grading it as "favourite
won" polluted the hit-rate with the base favourite-win rate, and its high frozen-ness
score always cleared the backtest threshold.

- Add AnomalyKind.IsDirectional() (flip + steam = true, freeze = false).
- AnomalyOutcomeEvaluator returns Unresolved for non-directional kinds (favourites
  still surfaced for display) so they don't distort calibration.
- RunBacktestUseCase skips non-directional anomalies when building candidates.
- Tests for the classification, the evaluator path, and the backtest skip.

tests/Marathon.Domain.Tests/AnomalyDetection/AnomalyOutcomeEvaluatorTests.cs

@@ -73,6 +73,27 @@ public sealed class AnomalyOutcomeEvaluatorTests
         verdict.ActualWinner.Should().Be(Side.Side1);
     }
 
+    [Fact]
+    public void Should_ReportUnresolved_When_KindIsNonDirectional()
+    {
+        // SuspensionFreeze is informational (the favourite did not change) — it must NOT be
+        // graded hit/miss even when a result exists, or its base favourite-win rate would
+        // pollute the calibration. Favourites are still surfaced for display.
+        var anomaly = new Anomaly(
+            Id: Guid.NewGuid(),
+            EventId: DefaultEventId,
+            DetectedAt: new DateTimeOffset(2026, 5, 10, 18, 5, 0, MoscowOffset),
+            Kind: AnomalyKind.SuspensionFreeze,
+            Score: 0.9m,
+            EvidenceJson: ThreeWayFlipJson);
+        var result = MakeResult(Side.Side2, s1: 0, s2: 2);
+
+        var verdict = AnomalyOutcomeEvaluator.Evaluate(anomaly, new SportCode(6), result);
+
+        verdict.Outcome.Should().Be(AnomalyOutcomeKind.Unresolved);
+        verdict.PostFlipFavourite.Should().Be(Side.Side2);
+    }
+
     [Fact]
     public void Should_ReportMiss_When_DrawOccurred_AndPostFlipFavouriteIsNotDraw()
     {