Battles are milestones of human history. For the Middle Ages, much more so than for any subsequent period, we rely heavily on the accounts of contemporary historians, as we rarely have many archival records. However, as medieval historians were mainly clerics, they were often misinformed. Even a chronicler fond of chivalric prowess such as Froissart never went to a battlefield himself, and this was the case with many chroniclers. Thus, my aim is to answer a simple question: what types of sources did medieval chroniclers use in order to tell what happened in battles? I shall also offer a classification of these sources and an assessment of their accuracy and significance. First, I will consider the case of testimonies provided by warriors directly involved in combat. The role of the heralds of arms, so often quoted by the chroniclers, will also be studied. Then, I will turn to the use of archival documents, such as letters from kings and generals, and also other types of sources. Finally, I will consider the case of warrior-chroniclers, such as Villehardouin. Their use of personal memories does not necessarily mean that their narratives are more accurate. Like many others, they used many commonplaces and often narrated what should have happened in battles rather than what actually happened.