You are not going to believe this: your weight is a better predictor of your race performance than how much or how well you train.
From Chapter One: Get Leaner, Go Faster in the new edition of Racing Weight:
The advantages of being light and lean for endurance performance are so obvious that they hardly needed to be scientifically proven, but exercise scientists have gone out and proven them anyway, and the proof is interesting. In a 1986 study Peter Bale and his colleagues at England’s Brighton Polytechnic University compared a host of anthropometric measurements in a group of 60 male runners (Bale, Bradbury, and Colley 1986). The subjects were divided into three groups of 20 based on their best 10K race times. The average weight of the men in the “average” group was 152 pounds compared to 145 pounds in the good group and 141 pounds in the elite group. Body composition measurements followed a similar pattern. Average body-fat percentages were 12.1, 10.7, and 8.0 in the average, good, and elite groups, respectively.
It bears noting that even the runners making up the average group were somewhat lighter and significantly leaner than the average nonrunner. The sport selects for naturally lighter and leaner individuals because they generally find greater initial success. The selection pressure continues within the sport right up to the top level. While most world-class runners have similar body weights (with women being lighter than men, naturally), research has shown that within the population of world-class runners, those with the lowest body-fat percentages tend to have the fastest race times.
Studies involving other types of endurance athletes have yielded similar findings. In 2011 Swiss researchers compared anthropometric variables against Ironman® swim, bike, and run split times in a group of 184 agegroup triathletes (Knechtle et al. 2011). Body weight was found to have a statistically moderate effect on total race time, while body-fat percentage had a large effect on total race time and a moderate effect (bordering on large) on swim, bike, and run splits. Both body weight and body-fat percentage were more strongly correlated with split times and total race time than are training variables such as average weekly training time.