Discrete Mathematics is an inevitable part of any undergraduate computer science degree programme. However, computer science students typically find this to be at best a necessary evil with which they struggle to engage. Twenty years ago, we started to address this issue seriously in our university, and we have instituted a number of innovations throughout the years which have had a positive effect on engagement and, thus, attainment. At the turn of the century, a mere 2% of our first-year students attained a 1st-class mark (a mark over 70%) in the discrete mathematics course whilst over half of the class were awarded a failing grade (a mark under 40%). Despite the course syllabus and assessment remaining as difficult as ever (if not more challenging), and despite maintaining the same entrance requirements to the programme whilst more than tripling the class size, for the past two years, two-thirds of the class attained a first-class mark whilst less than 2% of the class failed. In this paper, we describe and motivate the innovations which we introduced, and provide a detailed analysis of how and why attainment levels varied over two decades as a direct result of these innovations.