Skip to main content
Home/Accuracy

How Accurate Is Alphr?

Every prediction we make is tracked and verified. Here's how our models have performed across 2287 resolved and held-out model-tested AFL, NRL and Super League predictions.

H2H Accuracy

81.9%AFL
61.5%NRL
68.9%Super League

All Models

73.0%AFL
58.6%NRL
65.9%Super League

Brier Score

0.1809AFL
0.2424NRL
0.2112Super League

📊Round-by-Round H2H Accuracy

Bars = round accuracy · Lines = cumulative accuracy · Dashed = 50 % baseline

25%50%75%100%ORR3R6R9R12R15R18R21R24PFPFGF
AFL
NRL
Super League
Below 50 %

💰H2H P&L, 1 Unit Flat

Simulated profit / loss betting 1 unit on every Alphr H2H tip

AFL+60.5u+28.0% ROI216u staked177/216 won
NRL-16.2u-7.6% ROI213u staked131/213 won
Super League+105.3u+31.5% ROI334u staked230/334 won
-60u-30u0u+30u+60u+90u+120uORR3R6R9R12R15R18R21R24PFPFGF
AFL
NRL
Super League

Accuracy by Model

Strike rate and Brier score per model across AFL, NRL and Super League.

Head to Head

Win/loss predictions
AFL
81.9%177/216B 0.1451
NRL
61.5%131/213B 0.2300
Super League
68.9%230/334B 0.1953

Margin Band

Victory margin range
AFL
76.9%166/216B 0.1552
NRL
60.4%128/212B 0.2325
Super League
74.0%247/334B 0.1901

Over/Under

Total score predictions
AFL
60.2%130/216B 0.2423
NRL
53.8%114/212B 0.2648
Super League
54.8%183/334B 0.2481

Edge-Filtered Picks

Predictions where our model found ≥7 % edge over market odds.

75.2%AFL

173/230

52.8%NRL

86/163

Super League

odds pending

Expected vs Actual

How closely predicted probabilities align with outcomes. A perfectly calibrated model sits on the diagonal.

50-60%

AFL61% actual · 55% exp · 282
NRL50% actual · 55% exp · 221
Super League54% actual · 55% exp · 106

60-70%

AFL64% actual · 65% exp · 78
NRL58% actual · 65% exp · 278
Super League61% actual · 65% exp · 83

70-80%

AFL77% actual · 75% exp · 70
NRL74% actual · 75% exp · 129
Super League83% actual · 75% exp · 84

80%+

AFL90% actual · 91% exp · 218
NRL56% actual · 91% exp · 9
Super League85% actual · 91% exp · 61
Expected
AFL
NRL
Super League

Brier Score by Model

Lower is better. 0 = perfect, 0.25 = coin flip.

Head to Head

0.14510.23000.1953

Margin Band

0.15520.23250.1901

Over/Under

0.24230.26480.2481

Result Composition

Breakdown of resolved predictions by outcome.

AFL73 % won · 473/648
NRL59 % won · 373/637
Super League66 % won · 660/1002

Methods

All metrics are calculated from 648 resolved AFL predictions and 637 resolved NRL predictions from the 2025 season, plus 1002 Super League held-out model-tested predictions. AFL predictions are generated by XGBoost models trained on 18+ years of match data using 49 engineered features. NRL predictions use XGBoost models trained on 18+ years of data with ~160 features including referee bias modelling. Super League uses 183 engineered features including team ELO, player lineup ELO, rolling form, venue, H2H and rest-day signals.

Super League metrics use held-out 2024-2025 official match data and the same H2H, margin and total prediction primitives as AFL and NRL. Odds ingestion is pending, so the H2H P&L chart uses a flat 1.91 display benchmark for Super League.

Accuracy is the proportion of predictions where the model's selected outcome matched reality. Brier score measures the mean squared error between predicted probability and actual outcome (0 = perfect, 0.25 = coin flip).

Edge picks are filtered to require ≥7 % edge between model probability and market-implied probability. Calibration is assessed by grouping predictions into probability buckets and comparing expected win rate to actual win rate.

All 2288 predictions are transparently tracked and verifiable on this site. Super League public settlement will be added once the odds and market-tip layer is live.

Frequently Asked Questions

Common questions about how we measure prediction accuracy.

What does "accuracy" mean here?
Accuracy is the percentage of predictions where Alphr's selected outcome (e.g., which team wins) matched the actual result.
What is a Brier score?
A Brier score measures the quality of probabilistic predictions. It's the mean squared error between the predicted probability and the actual result (1 or 0). Lower is better, 0 is perfect, 0.25 is no better than a coin flip.
What are "edge picks"?
Edge picks are predictions where our model's estimated probability is at least 7 percentage points higher than the market-implied probability (1 / odds). These are our highest-conviction selections.
How is the calibration chart read?
Each row represents a probability range. If a model says 60-70 % and the actual win rate is ~65 %, it's well-calibrated. The grey line shows the expected midpoint; the coloured bar shows actuals.
Where does this data come from?
AFL and NRL accuracy comes from recorded 2025 site tips verified against final results. Super League comes from official match-centre results and player stats, with 2024-2025 held out for model testing.