Continuous postgame learning

daily_20260807 · dates 20260807 · generated 2026-08-08T02:50:05.532591+00:00

6-2

Winner picks · 75.0%

Comparable starters · 0 identity mismatches

0/0

Starter ER within ±1

0/0

Starter workload within ±1 inning

Where the score pipeline gained or lost quality

Direct comparison against official final runs. Positive deltas mean the stage was worse than historical_ml; this is score error, not a synthetic one-game ERA.

Stage	Games	Team-run MAE	Total-run MAE	Margin MAE	Winner accuracy	Δ team MAE vs base	Δ margin MAE vs base
historical_ml	8	2.2882	2.8205	3.5636	50.0%	0.0	0.0
variance_applied	8	2.4375	3.125	3.625	50.0%	0.1493	0.0614
enhanced_prediction	8	2.25	3.5	2.75	75.0%	-0.0382	-0.8136
embedded_savant_run_environment	8	2.25	3.5	2.75	75.0%	-0.0382	-0.8136
fundamental_edge_target_review	8	2.25	3.5	2.75	75.0%	-0.0382	-0.8136
capped_target	8	2.25	3.5	2.75	75.0%	-0.0382	-0.8136
simulation_result	8	2.5625	3.625	3.375	75.0%	0.2743	-0.1886
published_final	8	2.5625	3.625	3.375	75.0%	0.2743	-0.1886

Largest measured regression

{
  "stage": "simulation_result",
  "games": 8,
  "team_run_mae": 2.5625,
  "total_run_mae": 3.625,
  "margin_mae": 3.375,
  "winner_hits": 6,
  "winner_accuracy": 0.75,
  "team_run_mae_delta_vs_historical_ml": 0.2743,
  "margin_mae_delta_vs_historical_ml": -0.1886
}

Automatic questions

Which recorded stage contributed most to losses?

partial_attribution sample=2

Largest recorded pipeline-stage movement toward the final pick. This is deterministic trace attribution, not causal SHAP attribution.

Stage	Total Toward Pick Run Delta	Games
simulation_result	4.0	2
enhanced_prediction	1.0	2
embedded_savant_run_environment	0.0	2
fundamental_edge_target_review	0.0	2
capped_target	0.0	2
variance_applied	-0.6369	2

Complete structured answer

{
  "status": "partial_attribution",
  "sample_size": 2,
  "top_stage": "simulation_result",
  "ranking": [
    {
      "stage": "simulation_result",
      "total_toward_pick_run_delta": 4.0,
      "games": 2
    },
    {
      "stage": "enhanced_prediction",
      "total_toward_pick_run_delta": 1.0,
      "games": 2
    },
    {
      "stage": "embedded_savant_run_environment",
      "total_toward_pick_run_delta": 0.0,
      "games": 2
    },
    {
      "stage": "fundamental_edge_target_review",
      "total_toward_pick_run_delta": 0.0,
      "games": 2
    },
    {
      "stage": "capped_target",
      "total_toward_pick_run_delta": 0.0,
      "games": 2
    },
    {
      "stage": "variance_applied",
      "total_toward_pick_run_delta": -0.6369,
      "games": 2
    }
  ],
  "interpretation": "Largest recorded pipeline-stage movement toward the final pick. This is deterministic trace attribution, not causal SHAP attribution."
}

Which recorded stage contributed most to wins?

partial_attribution sample=6

Largest recorded pipeline-stage movement toward the final pick. This is deterministic trace attribution, not causal SHAP attribution.

Stage	Total Toward Pick Run Delta	Games
enhanced_prediction	8.0	6
simulation_result	3.0	6
embedded_savant_run_environment	0.0	6
fundamental_edge_target_review	0.0	6
capped_target	0.0	6
variance_applied	-2.5746	6

Complete structured answer

{
  "status": "partial_attribution",
  "sample_size": 6,
  "top_stage": "enhanced_prediction",
  "ranking": [
    {
      "stage": "enhanced_prediction",
      "total_toward_pick_run_delta": 8.0,
      "games": 6
    },
    {
      "stage": "simulation_result",
      "total_toward_pick_run_delta": 3.0,
      "games": 6
    },
    {
      "stage": "embedded_savant_run_environment",
      "total_toward_pick_run_delta": 0.0,
      "games": 6
    },
    {
      "stage": "fundamental_edge_target_review",
      "total_toward_pick_run_delta": 0.0,
      "games": 6
    },
    {
      "stage": "capped_target",
      "total_toward_pick_run_delta": 0.0,
      "games": 6
    },
    {
      "stage": "variance_applied",
      "total_toward_pick_run_delta": -2.5746,
      "games": 6
    }
  ],
  "interpretation": "Largest recorded pipeline-stage movement toward the final pick. This is deterministic trace attribution, not causal SHAP attribution."
}

Are Overs performing worse than Unders?

not_measurable sample=0

The current artifacts contain predicted totals but no total line, Over/Under selection, price, or settled total bet.

Complete structured answer

{
  "status": "not_measurable",
  "sample_size": 0,
  "reason": "The current artifacts contain predicted totals but no total line, Over/Under selection, price, or settled total bet.",
  "missing_fields": [
    "total_line",
    "total_selection",
    "odds",
    "settled_result"
  ]
}

Are hitter props outperforming pitcher props?

not_measurable sample=0

Player stat projections are not stored as priced, selected, and settled prop predictions.

Complete structured answer

{
  "status": "not_measurable",
  "sample_size": 0,
  "reason": "Player stat projections are not stored as priced, selected, and settled prop predictions.",
  "missing_fields": [
    "prop_family",
    "prop_line",
    "selection",
    "odds",
    "settled_result"
  ]
}

Which sportsbooks produce better opportunities?

not_measurable sample=0

No settled prediction was linked to a sportsbook snapshot.

Complete structured answer

{
  "status": "not_measurable",
  "sample_size": 0,
  "reason": "No settled prediction was linked to a sportsbook snapshot.",
  "missing_fields": [
    "sportsbook",
    "selection_odds",
    "settled_profit"
  ],
  "books": []
}

Which confidence ranges underperform?

insufficient_sample sample=8

Range	Sample Size	Hits	Misses	Accuracy
[0.000, 0.525)	2	2	0	1.0
[0.525, 0.550)	2	2	0	1.0
[0.550, 0.600)	1	1	0	1.0
[0.600, 1.010)	3	1	2	0.3333

Complete structured answer

{
  "status": "insufficient_sample",
  "sample_size": 8,
  "minimum_recommended_sample": 20,
  "worst_observed_range": {
    "range": "[0.600, 1.010)",
    "sample_size": 3,
    "hits": 1,
    "misses": 2,
    "accuracy": 0.3333
  },
  "ranges": [
    {
      "range": "[0.000, 0.525)",
      "sample_size": 2,
      "hits": 2,
      "misses": 0,
      "accuracy": 1.0
    },
    {
      "range": "[0.525, 0.550)",
      "sample_size": 2,
      "hits": 2,
      "misses": 0,
      "accuracy": 1.0
    },
    {
      "range": "[0.550, 0.600)",
      "sample_size": 1,
      "hits": 1,
      "misses": 0,
      "accuracy": 1.0
    },
    {
      "range": "[0.600, 1.010)",
      "sample_size": 3,
      "hits": 1,
      "misses": 2,
      "accuracy": 0.3333
    }
  ]
}

Which monitored features are drifting?

insufficient_sample sample=15

Three slates are descriptive only and cannot establish statistical feature drift.

Date	Sample Size	Mean Confidence	Mean Predicted Total	Accuracy
20260807	15	0.589	7.4	0.75

Complete structured answer

{
  "status": "insufficient_sample",
  "sample_size": 15,
  "minimum_baseline_sample": 30,
  "minimum_recent_sample": 20,
  "reason": "Three slates are descriptive only and cannot establish statistical feature drift.",
  "daily_monitors": [
    {
      "date": "20260807",
      "sample_size": 15,
      "mean_confidence": 0.589,
      "mean_predicted_total": 7.4,
      "accuracy": 0.75
    }
  ],
  "features_monitored": [
    "confidence",
    "predicted_total",
    "winner_accuracy"
  ]
}

Which model version introduced errors?

insufficient_sample sample=8

Version error claims require at least 30 settled predictions per version.

Pipeline Version	Model Schema Version	Policy Version	Sample Size	Hits	Misses	Accuracy
enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season	historical_game_temporal_v7	model_winner_simulation_score_v1	8	6	2	0.75

Complete structured answer

{
  "status": "insufficient_sample",
  "sample_size": 8,
  "minimum_sample_per_version": 30,
  "versions": [
    {
      "pipeline_version": "enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season",
      "model_schema_version": "historical_game_temporal_v7",
      "policy_version": "model_winner_simulation_score_v1",
      "sample_size": 8,
      "hits": 6,
      "misses": 2,
      "accuracy": 0.75
    }
  ],
  "reason": "Version error claims require at least 30 settled predictions per version."
}

Every settled game

Date	Game	Pick	Confidence	Pred A-H	Actual A-H	Result	Largest recorded movement	Pipeline
20260807	20260807_ATL@NYY	ATL	64.0%	4.0-3.0	-	pending_actual	enhanced_prediction (2.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_BAL@TEX	TEX	51.1%	3.0-5.0	1-2	hit	simulation_result (1.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_CHC@KC	CHC	64.0%	5.0-2.0	6-4	hit	enhanced_prediction (3.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_CIN@WAS	WAS	53.0%	2.0-4.0	3-5	hit	simulation_result (1.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_CLE@CHW	CLE	51.7%	5.0-4.0	8-2	hit	variance_applied (-0.3491)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_COL@STL	STL	66.0%	3.0-4.0	-	pending_actual	embedded_savant_run_environment (1.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_DET@SF	DET	66.0%	5.0-2.0	-	pending_actual	enhanced_prediction (3.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_HOU@SD	SD	55.7%	4.0-5.0	-	pending_actual	variance_applied (-0.1116)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_LAA@MIA	MIA	66.0%	1.0-5.0	4-3	miss	simulation_result (2.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_LAD@ARI	LAD	64.0%	5.0-4.0	-	pending_actual	enhanced_prediction (-2.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_MIN@MIL	MIL	60.0%	2.0-5.0	8-6	miss	simulation_result (2.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_NYM@PIT	NYM	52.8%	3.0-2.0	6-4	hit	enhanced_prediction (2.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_OAK@BOS	BOS	56.0%	3.0-6.0	1-13	hit	enhanced_prediction (3.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_TB@SEA	TB	53.0%	4.0-3.0	-	pending_actual	enhanced_prediction (1.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season
20260807	20260807_TOR@PHI	PHI	60.0%	2.0-6.0	-	pending_actual	simulation_result (2.0)	enhanced_game_prediction_v4_continuous_learning__player_form_v3_5_10_season

Starting-pitcher component quality

H	{ "mae": null, "bias_predicted_minus_actual": null }
R	{ "mae": null, "bias_predicted_minus_actual": null }
ER	{ "mae": null, "bias_predicted_minus_actual": null }
BB	{ "mae": null, "bias_predicted_minus_actual": null }
SO	{ "mae": null, "bias_predicted_minus_actual": null }
HR	{ "mae": null, "bias_predicted_minus_actual": null }
IP_outs	{ "mae_outs": null, "mae_innings": null, "bias_outs": null }

IP error is measured in outs. No single-outing ERA is used; ER and workload are compared directly.

Date	Game	Side	Pred starter	Actual starter	Match	Pred IP	Actual IP	Δ outs	Pred ER	Actual ER	Δ ER	H P/A	BB P/A	SO P/A	All context