This document is a thesis by Peter Schindler from Imperial College London about improving tennis ranking systems and conducting in-play probability analysis of tennis matches. It uses two main datasets: 1) match result data from 2000-2015 containing winners, scores, rankings, and other information, and 2) point-by-point data from 2014-2015 matches scraped from websites. The thesis develops an alternative Elo rating system to rank players, compares it to the ATP rankings, and analyzes the 2014 Wimbledon final point-by-point to track match probability changes and identify important points.