Thanks for this example! I'm not convinced that we can label the two solutions good or bad: While the sizes of the red pie chart may be biased, the proportions still feel more correct than the blue version. With a dataset ``` const data = [5,5,5,1,1,1,1,1]; ``` The five smaller pieces take up 30% of the pie and not just 25%. With this specific dataset, I almost feel that the red pie chart is returns a more precise picture than the blue one. My take-away is using padAngle with care.