Experiment 3 Example Figures

Tip

Click the figure to view in fullscreen.

Warning

Please note that the figures presented are based on only 1,000 games, which is a much smaller sample size than the 200,000 games used in the paper. As a result, the figures may not be as precise as those in the paper.

Figure 7 in the paper

$\tau$ prediction accuracy (ACC) along inter and intra episodes of BSI-PT and the BPR variants against the pitcher with $\omega_{op}$ under different $\epsilon$ values

$\epsilon=0$

BSI-PT's ACC for $\epsilon=0$

Bayes-OKR's ACC for $\epsilon=0$

DBPR+'s ACC for $\epsilon=0$

BPR+'s ACC for $\epsilon=0$

$\epsilon=0.2$

BSI-PT's ACC for $\epsilon=0.2$

Bayes-OKR's ACC for $\epsilon=0.2$

DBPR+'s ACC for $\epsilon=0.2$

BPR+'s ACC for $\epsilon=0.2$

$\epsilon=0.5$

BSI-PT's ACC for $\epsilon=0.5$

Bayes-OKR's ACC for $\epsilon=0.5$

DBPR+'s ACC for $\epsilon=0.5$

BPR+'s ACC for $\epsilon=0.5$

$\epsilon=1$

BSI-PT's ACC for $\epsilon=1$

Bayes-OKR's ACC for $\epsilon=1$

DBPR+'s ACC for $\epsilon=1$

BPR+'s ACC for $\epsilon=1$

Figure 8 in the paper

Winning percentage (WP) along inter and intra episodes of BSI-PT and the BPR variants against the pitcher with $\omega_{op}$ under different $\epsilon$ values

Winning percentage (WP) for $\epsilon=0$

Winning percentage (WP) for $\epsilon=0.2$

Winning percentage (WP) for $\epsilon=0.5$

Winning percentage (WP) for $\epsilon=1$