Artículo Materias > Educación física y el deporte Universidad Europea del Atlántico > Investigación > Artículos y libros Cerrado Inglés The aim of this study was to compare pacing and packing behaviours between sex and performance level at elite Berlin marathon races. Official electronic split and finishing times from 279 (149 male and 130 female) marathon performances, including 5 male world records, were obtained from 11 Berlin marathon races held from 2008 to 2018, and from two previous world records and the second world all-time fastest performance also achieved at that same Berlin course. Male performances displaying an even pacing behaviour were significantly faster than those adopting a positive behaviour (p < 0.001; d = 0.75). Male world records were characterized by even profiles with fast endspurts, being especially remarkable at world all-time two fastest performances which were assisted by the use of a new shoe technology. Female marathon runners decreased their speed less than men during the second half marathon and especially from the 35th km onwards (p < 0.001; 0.51 ≤ d ≤ 0.55). The latest race stages were usually run individually in both sexes. Significant pace differences between performance groups at every race segment were found in women (p < 0.01; 1.0 ≤ d ≤ 2.0), who also covered an important part of the race alone. Prior to participation in meet marathon races such as Berlin marathon, elite runners should select the group that they will join during the race according to their current performance level as a preassigned pace set by a pacemaker will be adopted. Therefore, they could follow an even rather than positive pacing behaviour which will allow them to achieve a more optimal performance. metadata Muñoz-Pérez, Iker; Lago-Fuentes, Carlos; Mecías-Calvo, Marcos y Casado, Arturo mail SIN ESPECIFICAR, carlos.lago@uneatlantico.es, marcos.mecias@uneatlantico.es, SIN ESPECIFICAR (2022) Pacing and packing behaviour in elite and world record performances at Berlin marathon. European Journal of Sport Science. pp. 1-8. ISSN 1746-1391