As the tunnelling industry is an ever-expanding market due to the switch to electric vehicles and expensive air space, the need for attaining a cost-efficient construction has never been greater. However, there is an understanding deficit on historical tunnel failures. The latter is partly due to sensitive information on insurance losses, delays not being formally released to the public and a lack of a comprehensive database. This paper examines all the significant tunnel failures from 1980 to 2019 and creates a single database including critical tunnel failure parameters as tunnel use, failure type, overburden, diameter, etc. The database shows that the introduction of The Code of Practice for Risk Management of Tunnelling Works has positively affected the number of tunnel incidents per year. More importantly, an edit distance model is created capable of entailing a total of six important tunnel parameters from which similarities and potential matches can be identified within the database of failure records. The more matches the given assessment scenario has to the database of the historical failures, the higher the likelihood of the tunnel to also experience some potential failures. Additionally, the edit distance model can give a percentage prediction of the failure type that is most likely to occur. Finally, an estimated insurance loss (cost overrun) and project delay can be given by looking at the losses and delays from the different failure types in the database.