NHL Prediction Model Performance & Calibration

See how our NHL prediction model performs: accuracy, Brier scores, calibration, and error metrics, with prospective records separated from historical diagnostics.

To understand each metric, read Understanding Performance Metrics . To apply this in practice, view today's NHL predictions, our NHL playoff odds, and in-game win probability charts.

Prospective Record — Saved Before Puck Drop

Regular-season raw-model forecasts recorded before scheduled puck drop. Each game uses the latest saved forecast before both its archived and current start time. Results include overtime and shootouts. Historical date-only records are excluded.

Model 09a777065076

2026-10-09 to 2026-10-09 · 4 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	75.0%	0.2695	0.7360
50/50 probabilities	—	0.2500	0.6931
Always pick home	50.0%	—	—

Accuracy: 75.0%; approximate 95% interval 30.1%–95.4%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 1.81 goals (4 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
46%–46%	1	45.9%	0.0%	0.0%–79.3%
52%–56%	2	53.8%	100.0%	34.2%–100.0%
66%–66%	1	66.4%	0.0%	0.0%–79.3%

Model 12a77d125e2d

2026-10-08 to 2026-10-08 · 10 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	60.0%	0.1926	0.5668
50/50 probabilities	—	0.2500	0.6931
Always pick home	60.0%	—	—

Accuracy: 60.0%; approximate 95% interval 31.3%–83.2%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 1.66 goals (10 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
38%–38%	1	38.4%	0.0%	0.0%–79.3%
49%–49%	1	49.4%	100.0%	20.7%–100.0%
54%–56%	3	54.9%	33.3%	6.1%–79.2%
62%–65%	2	63.6%	50.0%	9.5%–90.5%
73%–77%	2	74.6%	100.0%	34.2%–100.0%
82%–82%	1	82.2%	100.0%	20.7%–100.0%

Model 2ee1c783ab2a

2026-09-30 to 2026-10-04 · 34 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	61.8%	0.2430	0.6807
50/50 probabilities	—	0.2500	0.6931
Always pick home	58.8%	—	—

Accuracy: 61.8%; approximate 95% interval 45.0%–76.1%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 1.92 goals (34 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
22%–30%	3	25.2%	33.3%	6.1%–79.2%
36%–36%	2	36.1%	100.0%	34.2%–100.0%
43%–50%	12	46.1%	41.7%	19.3%–68.0%
51%–59%	8	55.2%	75.0%	40.9%–92.9%
63%–70%	6	65.9%	66.7%	30.0%–90.3%
72%–76%	3	74.2%	66.7%	20.8%–93.9%

Model 3c307b4cc1b9

2026-10-12 to 2026-10-12 · 2 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	0.0%	0.5141	1.2624
50/50 probabilities	—	0.2500	0.6931
Always pick home	50.0%	—	—

Accuracy: 0.0%; approximate 95% interval 0.0%–65.8%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 0.61 goals (2 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
29%–29%	1	28.6%	100.0%	20.7%–100.0%
72%–72%	1	72.0%	0.0%	0.0%–79.3%

Model 9a870d96081b

2026-10-05 to 2026-10-05 · 4 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	50.0%	0.2620	0.7172
50/50 probabilities	—	0.2500	0.6931
Always pick home	50.0%	—	—

Accuracy: 50.0%; approximate 95% interval 15.0%–85.0%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 1.13 goals (4 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
61%–61%	1	60.8%	0.0%	0.0%–79.3%
74%–76%	2	74.6%	50.0%	9.5%–90.5%
80%–80%	1	80.3%	100.0%	20.7%–100.0%

Model a998a268e982

2026-10-13 to 2026-10-13 · 16 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	68.8%	0.2094	0.6103
50/50 probabilities	—	0.2500	0.6931
Always pick home	43.8%	—	—

Accuracy: 68.8%; approximate 95% interval 44.4%–85.8%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 1.64 goals (16 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
27%–27%	1	27.3%	100.0%	20.7%–100.0%
35%–40%	4	37.5%	25.0%	4.6%–69.9%
41%–47%	3	45.1%	0.0%	0.0%–56.1%
51%–58%	4	54.4%	25.0%	4.6%–69.9%
62%–67%	2	64.5%	100.0%	34.2%–100.0%
72%–73%	2	72.2%	100.0%	34.2%–100.0%

Model b0997d329740

2026-10-14 to 2026-10-14 · 2 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	50.0%	0.3111	0.8179
50/50 probabilities	—	0.2500	0.6931
Always pick home	50.0%	—	—

Accuracy: 50.0%; approximate 95% interval 9.5%–90.5%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 1.96 goals (2 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
51%–51%	1	50.6%	100.0%	20.7%–100.0%
62%–62%	1	61.5%	0.0%	0.0%–79.3%

Model b9cc1fbec83e

2026-10-06 to 2026-10-06 · 9 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	66.7%	0.2391	0.6709
50/50 probabilities	—	0.2500	0.6931
Always pick home	55.6%	—	—

Accuracy: 66.7%; approximate 95% interval 35.4%–87.9%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 2.00 goals (9 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
40%–49%	3	44.1%	33.3%	6.1%–79.2%
56%–59%	2	57.8%	50.0%	9.5%–90.5%
61%–67%	3	63.4%	66.7%	20.8%–93.9%
71%–71%	1	70.6%	100.0%	20.7%–100.0%

Model d5a2e9de4f19

2026-10-10 to 2026-10-10 · 14 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	64.3%	0.2098	0.6090
50/50 probabilities	—	0.2500	0.6931
Always pick home	50.0%	—	—

Accuracy: 64.3%; approximate 95% interval 38.8%–83.7%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 1.83 goals (14 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
19%–19%	1	19.0%	0.0%	0.0%–79.3%
31%–31%	1	31.2%	0.0%	0.0%–79.3%
43%–50%	2	46.2%	50.0%	9.5%–90.5%
58%–58%	1	58.1%	0.0%	0.0%–79.3%
61%–70%	6	65.1%	66.7%	30.0%–90.3%
76%–79%	2	77.5%	50.0%	9.5%–90.5%
82%–82%	1	81.7%	100.0%	20.7%–100.0%

Model def0cfad25b6

2026-10-11 to 2026-10-12 · 4 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	75.0%	0.1685	0.5109
50/50 probabilities	—	0.2500	0.6931
Always pick home	50.0%	—	—

Accuracy: 75.0%; approximate 95% interval 30.1%–95.4%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 1.78 goals (4 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
18%–18%	1	18.3%	0.0%	0.0%–79.3%
53%–58%	2	55.9%	50.0%	9.5%–90.5%
72%–72%	1	71.7%	100.0%	20.7%–100.0%

Model e4f04ea9af43

2026-10-07 to 2026-10-07 · 3 games

Same-game probability comparison
Forecast	Accuracy	Brier	Log loss
Model	33.3%	0.2653	0.7173
50/50 probabilities	—	0.2500	0.6931
Always pick home	66.7%	—	—

Accuracy: 33.3%; approximate 95% interval 6.1%–79.2%. These Wilson intervals assume independent games; shared teams and dates can add uncertainty.

Median total-goals MAE: 1.33 goals (3 games). Lower Brier, log loss, and MAE are better.

Probability calibration

Home-probability range	Games	Mean forecast	Home win rate	Approx. 95% interval
25%–25%	1	25.3%	0.0%	0.0%–79.3%
36%–36%	1	36.2%	100.0%	20.7%–100.0%
43%–43%	1	43.0%	100.0%	20.7%–100.0%

Versions are scored separately. Small samples can vary substantially. These metrics measure how closely game forecasts match observed outcomes.

Historical Diagnostics — Not the Prospective Record

Historical game summaries can include training games. Each table identifies its date window; the summary exports do not identify a model version or certify forecasts were frozen before puck drop. Treat these as historical diagnostics. Cross-validation and xG holdout results below are separate evaluations. Read the evaluation guide.

How to Read These Metrics

Accuracy: The percentage of games where the predicted winner (team with >50% win probability) actually won. Simple but incomplete — it ignores how confident the model was.
Brier Score: Measures the mean squared error of probability predictions (0–1 scale, lower is better). Assigning 50% to every game yields 0.25. Compare scores on the same games; a low Brier score alone does not establish calibration.
Log Loss: A logarithmic scoring rule that heavily penalises confident wrong predictions. Assigning 90% to a team that loses costs far more than assigning 55%. This keeps the model honest about uncertainty.
Calibration: Shows whether stated probabilities match real outcomes. In a well-calibrated model, games given a 70% win probability should be won about 70% of the time. The calibration tables below group predictions into decile bins so you can verify this directly.
MAE / RMSE (Total Goals Error): Both measure how far predicted totals are from actual totals, in goal units. The game-level windows report MAE (mean absolute error) because the point estimate is the median of the simulated total, and the median minimizes MAE. The cross-validation folds report RMSE because they score the mean (expected goals), where RMSE is the consistent metric.

Read the metrics together: Accuracy measures correct winners; calibration compares probabilities with observed rates. A model can match the overall win rate yet fail to distinguish easier games from harder ones. Check sample sizes and uncertainty before interpreting differences between models or groups. Learn more in our analytics guide and methodology.

Game Predictions (Multi-Window)

Window	Start	End	Games	Accuracy	Brier	Log Loss	Avg Winner Prob	MAE Total
last 30	2026-09-14	2026-10-14	107	65.4%	0.2227	0.6352	55.0%	1.850
season to date	2026-10-01	2026-10-14	99	65.7%	0.2215	0.6325	55.3%	1.732
multi season	2021-10-12	2026-10-14	6667	59.9%	0.2323	0.6562	53.1%	1.862

Totals (Over 5.5)

Window	Games	Accuracy	Brier	Log Loss	Avg Outcome Prob
last 30	107	53.3%	0.2565	0.7068	50.5%
season to date	99	53.5%	0.2537	0.7008	50.7%
multi season	6667	57.1%	0.2468	0.6872	51.7%

Playoff Game Performance

Start	End	Games	Accuracy	Brier	Log Loss
2022-05-02	2026-06-14	433	57.0%	0.2394	0.6714

Current Matchups

TOR vs BUF model odds OTT vs NYI model odds TBL vs SEA model odds COL vs PIT model odds ANA vs BOS model odds WSH vs NJD model odds

Daily Performance

Show:

Date	Games	Accuracy	Brier	Log Loss

Prediction Recap Highlights

Definition: High confidence means the model assigned the predicted team a win probability well above 50%. Correct predictions are the highest-confidence correct calls, misses are the highest-confidence incorrect calls, and surprise results show the largest absolute gap between win probability and the actual outcome.

Show:

Highest-Confidence Correct Predictions

Date	Matchup	Win Prob	Outcome

Biggest Misses (High Confidence)

Date	Matchup	Win Prob	Outcome

Surprise Results

Date	Matchup	Surprise	Outcome

Calibration (Win Prob Deciles) — Last 30

Bin	Count	Mean Pred	Observed
1	1	19.3%	0.0%
2	4	24.2%	25.0%
3	10	35.2%	40.0%
4	24	45.3%	33.3%
5	31	55.4%	58.1%
6	18	64.4%	61.1%
7	15	73.7%	73.3%
8	4	82.6%	100.0%

Calibration (Win Prob Deciles) — Season To Date

Bin	Count	Mean Pred	Observed
1	1	19.3%	0.0%
2	4	24.2%	25.0%
3	10	35.2%	40.0%
4	21	45.3%	33.3%
5	27	55.4%	59.3%
6	17	64.6%	58.8%
7	15	73.7%	73.3%
8	4	82.6%	100.0%

Calibration (Win Prob Deciles) — Multi Season

Bin	Count	Mean Pred	Observed
1	3	18.9%	33.3%
2	74	26.6%	17.6%
3	352	36.4%	29.5%
4	1799	46.0%	43.9%
5	2461	54.8%	52.7%
6	1403	64.3%	66.5%
7	514	73.8%	75.7%
8	61	82.5%	91.8%

Calibration (Over 5.5) — Last 30

Bin	Count	Mean Pred	Observed
4	7	47.9%	57.1%
5	51	55.5%	58.8%
6	46	63.2%	47.8%
7	3	73.1%	66.7%

Calibration (Over 5.5) — Season To Date

Bin	Count	Mean Pred	Observed
4	7	47.9%	57.1%
5	49	55.4%	57.1%
6	40	63.1%	50.0%
7	3	73.1%	66.7%

Calibration (Over 5.5) — Multi Season

Bin	Count	Mean Pred	Observed
4	44	48.4%	54.5%
5	3057	57.0%	54.9%
6	3269	63.4%	58.8%
7	295	71.8%	62.0%
8	2	85.5%	0.0%

Team Calibration (Home, Top 15 by Volume)

Team	Count	Mean Pred	Observed	Bias
CAR	249	65.1%	70.3%	-5.2%
EDM	246	59.3%	61.4%	-2.0%
DAL	244	55.7%	61.5%	-5.7%
FLA	243	60.4%	63.8%	-3.4%
VGK	241	57.0%	60.2%	-3.2%
COL	238	60.7%	65.5%	-4.9%
TBL	231	56.5%	64.1%	-7.5%
NYR	231	51.1%	57.1%	-6.0%
TOR	229	57.3%	59.0%	-1.6%
BOS	226	55.0%	62.8%	-7.8%
WPG	221	56.2%	60.6%	-4.5%
MIN	221	55.7%	58.4%	-2.7%
LAK	220	58.8%	54.1%	+4.7%
WSH	220	54.4%	53.6%	+0.8%
MTL	219	50.2%	43.8%	+6.3%

Team Calibration (Pred vs Observed)

Mean Pred Observed

Starter Calibration (Home)

Window	Starter Status	Games	Accuracy	Brier	Log Loss
last 30	Starter	107	65.4%	0.2227	0.6352
season to date	Starter	99	65.7%	0.2215	0.6325
multi season	Unknown	20	40.0%	0.2654	0.7278
multi season	Starter	6647	60.0%	0.2322	0.6560

Cross-Validation (Expanding Window)

Summary: 3 folds | Brier: 0.2545 | Log Loss: 0.7027 | RMSE Total: 2.410

Show fold details

Fold	Train N	Val N	Brier	Log Loss	RMSE
Fold 1	703	2,089	0.2559	0.7056	2.466
Fold 2	1,396	1,396	0.2548	0.7033	2.390
Fold 3	2,094	698	0.2529	0.6992	2.374

In-Game Checkpoints — Last 30

Checkpoint	Games	Accuracy	Brier	Log Loss
end_p1	39	76.9%	0.1484	0.4549
end_p2	39	82.1%	0.1067	0.3340
ot_start	9	33.3%	0.2193	0.6145
p3_10	39	87.2%	0.0704	0.2350
p3_5	39	87.2%	0.0684	0.2233
pregame	39	66.7%	0.2199	0.6295

In-Game Checkpoints — Season To Date

Checkpoint	Games	Accuracy	Brier	Log Loss
end_p1	31	77.4%	0.1416	0.4377
end_p2	31	83.9%	0.0992	0.3152
ot_start	7	42.9%	0.2094	0.5898
p3_10	31	87.1%	0.0704	0.2354
p3_5	31	87.1%	0.0681	0.2229
pregame	31	64.5%	0.2152	0.6192

In-Game Calibration — Pregame (Last 30 Days)

Bin	Count	Mean Pred	Observed
2	1	23.9%	0.0%
3	2	34.0%	50.0%
4	13	45.6%	38.5%
5	12	54.9%	66.7%
6	5	65.2%	60.0%
7	5	72.5%	80.0%
8	1	81.3%	100.0%

In-Game Calibration — End P2 (Last 30 Days)

Bin	Count	Mean Pred	Observed
0	4	1.3%	0.0%
1	4	15.6%	0.0%
2	2	25.5%	0.0%
3	2	32.9%	50.0%
4	4	48.5%	50.0%
5	4	54.4%	50.0%
6	2	65.5%	50.0%
7	6	74.0%	83.3%
8	2	87.3%	100.0%
9	9	96.5%	100.0%

In-Game Calibration — P3 10 (Last 30 Days)

Bin	Count	Mean Pred	Observed
0	9	2.0%	0.0%
1	4	15.7%	0.0%
2	1	21.6%	0.0%
3	2	39.0%	50.0%
4	4	48.6%	75.0%
6	1	69.9%	100.0%
7	1	79.9%	100.0%
8	3	86.8%	100.0%
9	14	97.2%	92.9%

xG Holdout — Contextual

Train: 2021-10-12 – 2025-06-06 | Test: 2025-06-09 – 2026-10-10

Games (test): 1481 | Shots (test): 127463 | ROC AUC: 0.777 | Log Loss: 0.2217 | Brier: 0.0600

xG Splits — Contextual Strength State

Split	Shots	Goal Rate	AUC	Log Loss	Brier
Even	101408	6.2%	0.771	0.2029	0.0540
PP	22118	10.4%	0.717	0.2892	0.0797
PK	2806	7.2%	0.806	0.2169	0.0608
EmptyNet	1131	51.1%	0.741	0.6039	0.2100

xG Splits — Contextual Shot Type

Split	Shots	Goal Rate	AUC	Log Loss	Brier
wrist	54723	7.0%	0.805	0.2048	0.0552
snap	32298	8.4%	0.769	0.2479	0.0694
slap	14907	4.9%	0.709	0.1820	0.0451
tip-in	12043	6.7%	0.653	0.2387	0.0615
backhand	9409	9.2%	0.811	0.2444	0.0685
deflected	2061	11.9%	0.687	0.3319	0.0964
wrap-around	824	4.1%	0.730	0.1613	0.0378
bat	638	8.6%	0.802	0.2537	0.0678
poke	428	11.4%	0.745	0.3132	0.0898
between-legs	84	11.9%	0.785	0.3051	0.0869
nan	38	86.8%	0.591	0.4402	0.1179
cradle	10	10.0%	1.000	0.2103	0.0561

xG Holdout — Neutral

Train: 2021-10-12 – 2025-06-06 | Test: 2025-06-09 – 2026-10-10

Games (test): 1481 | Shots (test): 127463 | ROC AUC: 0.773 | Log Loss: 0.2258 | Brier: 0.0614

xG Splits — Neutral Strength State

Split	Shots	Goal Rate	AUC	Log Loss	Brier
Even	101408	6.2%	0.770	0.2035	0.0541
PP	22118	10.4%	0.698	0.3087	0.0874
PK	2806	7.2%	0.804	0.2178	0.0610
EmptyNet	1131	51.1%	0.721	0.6164	0.2154

xG Splits — Neutral Shot Type

Split	Shots	Goal Rate	AUC	Log Loss	Brier
wrist	54723	7.0%	0.800	0.2097	0.0570
snap	32298	8.4%	0.767	0.2512	0.0708
slap	14907	4.9%	0.711	0.1819	0.0451
tip-in	12043	6.7%	0.654	0.2384	0.0614
backhand	9409	9.2%	0.797	0.2563	0.0730
deflected	2061	11.9%	0.686	0.3335	0.0968
wrap-around	824	4.1%	0.689	0.1639	0.0381
bat	638	8.6%	0.814	0.2529	0.0680
poke	428	11.4%	0.736	0.3155	0.0913
between-legs	84	11.9%	0.797	0.3086	0.0888
nan	38	86.8%	0.539	1.2494	0.1176
cradle	10	10.0%	1.000	0.2097	0.0569

Monthly Performance Trends

Track how model performance varies month-to-month across the season.

Month	Games	Accuracy	Brier	Log Loss
2021-10	131	54.2%	0.2476	0.6895
2021-11	209	54.5%	0.2446	0.6818
2021-12	153	60.1%	0.2377	0.6679
2022-01	202	59.9%	0.2322	0.6565
2022-02	151	52.3%	0.2483	0.6899
2022-03	234	59.8%	0.2367	0.6655
2022-04	231	61.0%	0.2342	0.6608
2022-05	1	0.0%	0.2891	—
2022-10	146	51.4%	0.2529	0.6988
2022-11	216	59.7%	0.2394	0.6721
2022-12	222	60.4%	0.2333	0.6582
2023-01	217	55.8%	0.2406	0.6738
2023-02	164	49.4%	0.2541	0.7023
2023-03	236	56.8%	0.2394	0.6714
2023-04	111	63.1%	0.2214	0.6339
2023-10	140	56.4%	0.2351	0.6601
2023-11	213	54.0%	0.2396	0.6709
2023-12	219	58.4%	0.2365	0.6661
2024-01	208	53.8%	0.2363	0.6633
2024-02	172	62.8%	0.2302	0.6529
2024-03	228	64.0%	0.2270	0.6458
2024-04	132	53.8%	0.2484	0.6897
2024-10	166	69.9%	0.2129	0.6148
2024-11	220	65.5%	0.2204	0.6305
2024-12	214	70.1%	0.2078	0.6045
2025-01	224	59.8%	0.2344	0.6611
2025-02	122	59.0%	0.2216	0.6328
2025-03	234	65.8%	0.2155	0.6202
2025-04	132	66.7%	0.2242	0.6387
2025-10	180	70.6%	0.2036	0.5955
2025-11	225	64.9%	0.2195	0.6292
2025-12	226	61.1%	0.2258	0.6430
2026-01	240	61.3%	0.2310	0.6540
2026-02	74	64.9%	0.2296	0.6538
2026-03	242	49.6%	0.2523	0.6982
2026-04	125	64.0%	0.2226	0.6357
2026-09	8	62.5%	0.2381	0.6690
2026-10	99	65.7%	0.2215	0.6325

Playoff Model Performance

Game-level and series-level accuracy across playoff rounds.

Playoff Games

Round	Games	Accuracy	Brier	Log Loss
All Rounds	82	56.1%	0.2455	0.6841
Round 1	45	57.8%	0.2456	0.6843
Round 2	22	59.1%	0.2417	0.6761
Round 3	9	44.4%	0.2568	0.7069
Round 4	6	50.0%	0.2426	0.6781

Playoff Series

Round	Series	Accuracy	Brier	Log Loss
All Rounds	15	66.7%	0.2298	0.6520
Round 1	8	62.5%	0.2352	0.6636
Round 2	4	75.0%	0.2320	0.6567
Round 3	2	50.0%	0.2213	0.6313
Round 4	1	100.0%	0.1950	0.5827

Playoff Calibration (Pred vs Observed)

Mean Pred Observed