Yeah - here's a real world example that gives a similar way to look at how bad luck effects low datasets in short time frames.