vir non camelus est.

Sunday, September 2, 2007

Minor Round Statistics

As the Fish prepare for some September action, it is an opportune time to look at some of the stats that have emerged from Winter Season 2007:

Win:loss - 10-7-1

Average winning margin (95% CI) - 6.89 points (1.4, 12.4)

Average losing margin - 9.13 points (4.7, 13.6)

Highest winning margin - 21 points (Round 21 - Spectrum)

Highest losing margin - 17 points (Round 16 - Red Dwarfs)

A logistic regression analysis was performed to examine factors predictive of a Black Fish victory. All 17 games in which a win or loss was recorded were included in the analysis. Following univariate analysis, a multivariate model was constructed using backwards elimination. Some interesting results were obtained:

1. Despite the Fish winning 5 of their 6 games played on court 2, there was no statistical relationship between the court and the odds of winning (p=0.248, Fisher's exact).

2. The Fish played 10 matches without a player on the bench, winning only 4 of these games. In contrast, they won 6 of the 7 games in which more than 5 players were available. Therefore, the odds of a victory were 9 times higher when an interchange player was available. This relationship was not statistically significant, however (p=0.082, Fisher's exact).

3. The presence of powerhouse forward Rodgers (N) increased the odds of victory by 33%, however this also failed to reach statistical significance (p=0.581, Fisher's exact).

4. The multivariate model revealed only one statistically significant predictive factor, after accounting for all other predictor variables. And the result may come as a surprise to some (but perhaps not to all). The presence of utility Downing (K), significantly increased the odds of a Black Fish victory when all other predictors were taken into account. KD increased the odds of victory tenfold (!), 95% CI (1.05, 95.5), p=0.045.

Make of that what you will.

No comments:

Post a Comment