// MODEL BACKTEST · 730D

Prediction model accuracy

Our Elo strength model powers the win probabilities on every team head-to-head. Here's how it actually performed — each match graded against the prediction the model would have made from only the games before it, over the last 730 days.

56.5%
favorite won
35/62
correct predictions
0.2497
Brier score (lower is better)
Calibration

When the model says it's X% confident, how often is it actually right? A well-calibrated model tracks the diagonal.

Confidence Actual win rate Predictions
50–60% 56.5% 62
Recent predictions
Date Model favorite Confidence Result
Aug 2 Spirit over MOUZ 55.6%
Aug 1 MOUZ over Astralis 50.1%
Aug 1 Spirit over FaZe 51.5%
Jul 31 FaZe over The MongolZ 54.8%
Jul 30 MOUZ over 3DMAX 51.5%
Jul 26 Spirit over 100 Thieves 54.5%
Jul 26 The MongolZ over Wildcard 52.2%
Jul 26 3DMAX over FUT 51.1%
Jul 26 Ninjas in Pyjamas over paiN 52.1%
Jul 25 FOKUS over MOUZ 50.2%
Jul 24 HEROIC over The MongolZ 59.7%
Jul 23 FUT over SINNERS 52.5%
Jul 23 Aurora over FOKUS 56.3%
Jul 22 FaZe over EYEBALLERS 56.5%
Jul 18 K27 over HEROIC 51.3%
Jul 18 K27 over Ninjas in Pyjamas 55.3%
Jul 17 K27 over 3DMAX 52.4%
Jul 17 Phantom over K27 56.8%
Jul 17 Gentle Mates over 3DMAX 53.3%
Jul 16 Gentle Mates over Ninjas in Pyjamas 55.4%
Jul 15 K27 over Ninjas in Pyjamas 54.4%
Jul 12 9z over PARIVISION 52.6%
Jul 11 FaZe over PARIVISION 55.8%
Jul 11 Alliance over 9z 52.6%
Jul 10 FaZe over BetBoom 51%