Most soccer, rugby union, rugby league, American football, Australian football and Gaelic football competitions over the world are played on natural grass over seasons that commence in the early autumn (fall) and extend through winter. Injury surveillance in these competitions has usually reported high rates of injury to the lower limb and an increased incidence of injuries early in the season. This 'early-season' bias has not usually been reported in summer football competitions, or in sports played indoors, such as basketball. Although easily compared rates have not often been published there has also been a reported trend towards a greater injury incidence in football played in warmer and/or drier conditions. Injury incidence in American football played on artificial turf has often been reported to be higher than in games played on natural grass. This review concludes that the most plausible explanation for all of these reported findings involves variations in playing surface characteristics. Shoe-surface traction for the average player is the specific relevant variable that is most likely to correlate with injury incidence in a given game of football. Shoe-surface traction will usually have a positive correlation with ground hardness, dryness, grass cover and root density, length of cleats on player boots and relative speed of the game. It is possible that measures to reduce shoe-surface traction, such as, ground watering and softening, play during the winter months, use of natural grasses such as perennial ryegrass (Lolium perenne L.) and player use of boots with shorter cleats, would all reduce the risk of football injuries. The most pronounced protective effect is likely to be on injuries to the lower limb of a noncontact nature, including anterior cruciate ligament injuries. Intervention studies should be performed, both using randomised and historical controls.