Fix shift value for confidence interval

[StephenNneji]

I missed fixing shift value on confidence interval

[DrPaulSharp]

I see that this change is to match up the shift used for the confidence intervals with the one used for the data elsewhere in the plotting module. However, I'm not sure that the "+1" is right here. For the first four values of i, with a shift_value of 100, we have div = 1, 100, 1000, 10000 - i.e., the factor 10 is missing. Looking at a plot of the current behaviour, for the second contrast the current behaviour gives the shown confidence region, whist this change gives the data. What do you think?

[DrPaulSharp]

I can confirm the same bug is in the matlab - it should be i-1 there. I've a branch ready to go if you agree with me here.

[DrPaulSharp]

All good!