Table 1.
Estimated relative effects of introducing auto-play.
| Relative effect (95% CI) and Bayesian posterior tail probability | ||||
|---|---|---|---|---|
| Weekly seasonality included | Monthly seasonality included | |||
| Standard prior | Robust prior | Standard prior | Robust prior | |
| Betted amount | +6.9% (0.62—14%), pBayes=.0149 | +6.9% (-0.88—16%), pBayes=.0424† | +8.7% (2.2—16%), pBayes=.0044 | +9.1% (0.92—18%), pBayes=.0146 |
| Net losses | -8.9% (-30—22%), pBayes=.2199 | -8% (-33—33%), pBayes=.2627 | -8.6% (-29—23%), pBayes=.2211 | -7.9% (-33—34%), pBayes=.261 |
| Spins | +3.2% (0.15—6.4%), pBayes=.0202 | +1.9% (-2.1—6.2%), pBayes=.1762 | +3.4% (0.34—6.6%), pBayes=.01547 | +2.2% (-1.8—6.5%), pBayes=.14795 |
†Note that probability estimates and credibility intervals may not always agree when probabilities are calculated from randomly drawn, asymmetric distributions. Standard priors were package-default 0.01, while robust were 0.02.