The distribution of losses in a confidence interval generally follow the shape of a lognormal curve. Unlike a normal curve, the mean of a lognormal curve will almost always fall to the right of the median. While a normal curve would have a mean at 50%, a lognormal curve’s mean is usually greater than 50% due to the tail to the right. The amount that the lognormal mean is greater than 50% varies based upon the standard deviation, but it generally falls between 55%-65% for most loss projections.
Articles in this section
- What is a Loss Projection?
- What is a Reserve Analysis?
- Does Loss Forecaster limit development on individual claims that have exceeded the retention/loss limit?
- Does Loss Forecaster have the capability to produce Loss Development Triangles?
- How is IBNR accounted for in Loss Forecaster?
- How do loss projections using the same data but different loss limits produce the same results?
- Is there a specific industry accepted reason that the Confidence Interval defaults to a factor between 60%-65%?
- Where in the calculation does the Excess of Threshold Data impact the calculation?
- How can I add a custom logo to my Loss Forecaster reports?
- How do I export a Loss Forecaster file to Excel?