"Lognormally distributed share returns": If share returns are lognormally distributed it means that the logarithm of [1 + the share return] has a normal probability distribution. Normal distributions have infinitely long â€˜tailsâ€™ both upside and downside - so implying unlimited downside potential when used for modelling share returns. But the theoretically worst outcome for a share investor is to lose the whole of their investment - in other words a negative return of -100%. It is not theoretically possible to suffer a return of worse than -100%. Lognormal distributions - unlike normal distributions - also have a limited downside, so they do not suffer from this theoretical shortcoming.