The aim of the research is to determine the technical efficiency of legendary hockey players in the National Hockey League (NHL), to create a ranking of these players and to reveal the best NHL players of all time. The research uses statistical data on 379 players from the 1944/45 season to the 2023/24 season. The methodology is based on multi-criteria analysis, specifically the concept of data envelopment analysis (DEA). The DEA provides an objective measure of the overall playing profile of hockey players and can help supplement the information in rankings provided by sports journalists. Andersen and Petersen's model is used to evaluate the data collected, providing super efficiency scores by aggregating NHL statistics related to various aspects of the game to produce a final ranking of hockey legends. The concept of data envelopment analysis works with multiple variables and allows for greater objectivity to be incorporated into the rankings. The number of games played is chosen as one of the model's input variables. The output variables include: number of goals scored, number of assists, plus/minus, inverse of penalty minutes, points per game, number of shots, number of individual awards, and number of Stanley Cups won. The research named Wayne Gretzky, Butch Goring and Serge Savard as the best players in the NHL historically in terms of technical efficiency. Among other things, the ranking of legendary hockey players revealed that players with a high number of games played or points scored are not necessarily technically efficient.