Use Grigorten formula for ssd_ecd()

[joethorley]

From email from David Fox

I personally think we should be using the Gringorten formula (replace both 0.375 and 0.25 with 0.44 in the formula above) as it has better tail behaviour for the type of distributions we are dealing with.

ssd_ecd(1:10)

[1] 0.06097561 0.15853659 0.25609756 0.35365854 0.45121951 0.54878049 0.64634146 0.74390244 0.84146341

[10] 0.93902439

((1:10) - 0.375) / (10.25)

[1] 0.06097561 0.15853659 0.25609756 0.35365854 0.45121951 0.54878049 0.64634146 0.74390244 0.84146341

[10] 0.93902439

Gringorten

((1:10) - 0.44) / (10.44)

[1] 0.05363985 0.14942529 0.24521073 0.34099617 0.43678161 0.53256705 0.62835249 0.72413793 0.81992337

[10] 0.91570881