There is a growing interest to provide more individualized estimates of treatment effect. An intuitive approach to facilitate personalized decision making is to predict the patient's risk with and without intervention. The mere difference between these two absolute outcome predictions then provides the absolute intervention effect ("treatment benefit") for that specific individual. Unfortunately, estimation of absolute intervention effects is not straightforward, even in randomized clinical trials. For this reason, advanced statistical methods are often necessary to determine the outcome risks of individual patients, and to evaluate individualized predictions of absolute treatment benefit. In this talk, I will introduce the main principles of absolute intervention effect prediction, and outline key challenges. Subsequently, I will illustrate statistical methods that may help to improve the estimation of absolute treatment effects, and their validation.