Abstract: Shadow mapping is a technique, widely used in computer graphics, for efficiently calculating shadows in real-time applications. However, shadows created with this technique can not be filtered as regular textures, causing severe aliasing. To solve this problem, several techniques based on statistical analysis, including variance shadow maps and exponential shadow maps, have been introduced by other authors in the past. However these techniques introduce different kinds of "light leaking" artifacts, which are clearly visible in moderately complex scenes. In this paper we propose a new statistical filtering method which approximates the cumulative distribution function (CDF) of depths by a Gaussian CDF instead of bounding it with Chebyshev Inequality. This approximation significantly reduces "light bleeding" artifacts, keeping the same performance and spatial cost as the original variance shadow maps.