The study of one type of error—the conviction of innocent people—has gained enormous importance, attracting increasing academic research and indeed giving birth to an activism geared towards obtaining the exoneration of innocent victims of unjust court convictions. One of the issues that has produced the greatest number of studies has been identifying factors that increase the probability that convictions of innocent people will occur. Among its results is the consensus that a group of "evidentiary practices" exists that may explain the errors. The present work sets out to describe, from the evidence available, the most problematic evidentiary practices in relation to the use of expert evidence. According to the empirical data available, this is one of the most relevant factors in the system’s production of wrong decisions. Based on a more refined diagnosis of which practices are most problematic in the use of this evidence, I hope to make it possible to gauge the system’s weaknesses. This will allow me to develop proposals and strategies for risk prevention and minimization. Diminishing and anticipating errors not only seems a realistic goal or a reasonable aspiration, but also an imperative for the system