We introduce a new scientometric index, inspired by the Lobby index from complex networks literature, that we call K-index. The K-index grows with the impact of the citing papers and can be thought of as a measure of scientific creativity and innovation. We show that the K-index can be easily computed from the Web of Science platform and presents several advantages over other bibliometric indexes. The K-index is robust to self-citations, is not limited by the total number of papers published by a researcher and is able to distinguish in a consistent way researchers that have the same h index but different scientific impacts: Einstein and Hirsch, for example. The K-index successfully detects a known case of inflated numbers for papers, citations and h index due to scientific career fraud. Finally, we show that, in a sample of twenty-nine physics Nobel laureates and thirty highly cited non-Nobel-laureate physicists, the K-index correlates better to the achievement of scientific prizes than the number of papers, citations, citations per paper, citing articles and the h index. Clustering researchers in a K versus h plot reveals interesting patterns that can be interpreted in terms of innovation and recognition.