Table 4 Details of selected Farrington variants, including parameter settings

A1

Original Farrington method,

Base settings,

No seasonality,

No population offset

Window half size (w) = 3; Years (b) = 5; Seasonality periods = 1; Trend conditionally included (at 0.05 alpha threshold); Past weeks to exclude = w

A2

Original Farrington method,

Base settings,

Seasonality with four periods

Window half size (w) = 3; Years (b) = 5;

Seasonality periods = 4; Trend conditionally included (at 0.05 alpha threshold); Past weeks to exclude = w

B1

Improved Farrington method,

Base settings,

No seasonality

Window half size (w) = 3; Years (b) = 5;

Seasonality periods = 1; Trend included; Past weeks to exclude = 26

B2

Improved Farrington method,

Base settings,

Seasonality with four periods

Window half size (w) = 3; Years (b) = 5; Seasonality periods = 4; Trend included; Past weeks to exclude = 26

C1

Improved with population offset;

Highest true positive in highest caught set

Window half size (w) = 5; Years (b) = 4; Seasonality periods = 4; Trend included; Past weeks to exclude = 26

C2

Improved with population offset;

Highest caught in highest true positive set

Window half size (w) = 3; Years (b) = maximum; Seasonality periods = 8; Trend excluded; Past weeks to exclude = w

C3

Improved with population offset;

Balanced trade-off option

Window half size (w) = 3; Years (b) = 5; Seasonality periods = 8; Trend excluded; Past weeks to exclude = w

C4

Improved with population offset;

Selected balanced trade-off

Window half size (w) = 5; Years (b) = 3; Seasonality periods = 4; Trend excluded; Past weeks to exclude = w