{ "band": 0.15, "band_sweep": { "0.05": { "hits": 45, "n": 78, "neutral_predicted": 5 }, "0.1": { "hits": 43, "n": 78, "neutral_predicted": 13 }, "0.15": { "hits": 43, "n": 78, "neutral_predicted": 16 }, "0.2": { "hits": 39, "n": 78, "neutral_predicted": 22 }, "0.3": { "hits": 37, "n": 78, "neutral_predicted": 31 }, "0.5": { "hits": 23, "n": 78, "neutral_predicted": 56 } }, "blocks": { "DeepSeek live (xau through band \u00b10.15)": { "accuracy": 1.0, "confusion": [ [ 32, 0, 0 ], [ 0, 8, 0 ], [ 0, 0, 38 ] ], "directional_accuracy": 1.0, "directional_hits": 70, "directional_n": 70, "hits": 78, "macro_f1": 1.0, "n": 78, "sign_agreement": 1.0, "sign_hits": 70, "sign_n": 70 }, "keyword fallback (band \u00b10.15)": { "accuracy": 0.3076923076923077, "confusion": [ [ 6, 15, 11 ], [ 0, 8, 0 ], [ 12, 16, 10 ] ], "directional_accuracy": 0.22857142857142856, "directional_hits": 16, "directional_n": 70, "hits": 24, "macro_f1": 0.30646952758745044, "n": 78, "sign_agreement": 0.41025641025641024, "sign_hits": 16, "sign_n": 39 }, "model (native argmax label)": { "accuracy": 0.6025641025641025, "confusion": [ [ 20, 1, 11 ], [ 0, 5, 3 ], [ 10, 6, 22 ] ], "directional_accuracy": 0.6, "directional_hits": 42, "directional_n": 70, "hits": 47, "macro_f1": 0.579918628305725, "n": 78, "sign_agreement": 0.6666666666666666, "sign_hits": 42, "sign_n": 63 }, "model (score through band \u00b10.15)": { "accuracy": 0.5512820512820513, "confusion": [ [ 19, 4, 9 ], [ 0, 3, 5 ], [ 8, 9, 21 ] ], "directional_accuracy": 0.5714285714285714, "directional_hits": 40, "directional_n": 70, "hits": 43, "macro_f1": 0.48980342078786476, "n": 78, "sign_agreement": 0.7017543859649122, "sign_hits": 40, "sign_n": 57 } }, "deepseek": { "batch_ms_max": null, "batch_ms_mean": null, "cache_hits": 78, "calls": 0, "failed_calls": 0, "headlines_sent": 0, "model": "deepseek-v4-flash", "per_headline_ms": null, "prompt_sha1": "d32e04c6a87d019f326f111b39c99b6dd8ce794d" }, "deepseek_scored": 78, "generated_at": "20260923_0037_xau_all_20260922", "mcnemar": { "model_vs_deepseek": { "deepseek_only_right": 31, "model_only_right": 0, "p": 9.313225746154785e-10 }, "model_vs_keyword": { "keyword_only_right": 8, "model_only_right": 31, "p": 0.0002940768754342571 } }, "model": { "label_distribution": { "bearish": 909, "bullish": 818, "neutral": 893 }, "model_path": "../models/candidates/xau_all_20260922.joblib", "model_type": "domain_augmented_tfidf_logistic_regression", "neutral_threshold": 0.05, "sklearn_version": "1.9.1", "trained_at": "2026-09-23T00:35:47.838884+00:00", "training_examples": 2620 }, "model_latency_ms": { "max": 4.03, "mean": 0.9802564102564102, "median": 0.935 }, "n": 78, "per_category": { "distractors": { "deepseek": 8, "deepseek_n": 8, "keyword": 8, "model": 5, "n": 8 }, "dovish_fed": { "deepseek": 8, "deepseek_n": 8, "keyword": 7, "model": 4, "n": 8 }, "gold_direct": { "deepseek": 8, "deepseek_n": 8, "keyword": 4, "model": 7, "n": 8 }, "hawkish_fed": { "deepseek": 8, "deepseek_n": 8, "keyword": 2, "model": 8, "n": 8 }, "inverted_metrics": { "deepseek": 6, "deepseek_n": 6, "keyword": 0, "model": 3, "n": 6 }, "negation_traps": { "deepseek": 8, "deepseek_n": 8, "keyword": 2, "model": 1, "n": 8 }, "real_yields": { "deepseek": 6, "deepseek_n": 6, "keyword": 1, "model": 3, "n": 6 }, "risk_on": { "deepseek": 4, "deepseek_n": 4, "keyword": 0, "model": 0, "n": 4 }, "safe_haven": { "deepseek": 8, "deepseek_n": 8, "keyword": 0, "model": 7, "n": 8 }, "strong_usd": { "deepseek": 7, "deepseek_n": 7, "keyword": 0, "model": 5, "n": 7 }, "weak_usd": { "deepseek": 7, "deepseek_n": 7, "keyword": 0, "model": 4, "n": 7 } }, "spearman_model_vs_deepseek": 0.4233409741735717 }