The purpose of this study was to investigate the frequency of countries represented in the TOP20 long-distance elite runners ranking during 1997–2020, taking into account the countries’ Human Development Index (HDI), and to verify if the Matthew effect can be observed regarding countries’ representativeness in the raking alongside the years.

The sample comprised 1852 professional runner athletes, ranked in the Senior World TOP20 half-marathon (403 female and 487 male) and marathon (480 female and 482 male) races, between the years 1997–2020. Information about the countries’ HDI was included, and categorized as “low HDI”, “medium HDI”, “high HDI”, and “very-high HDI”. Athletes were categorized according to their ranking positions (1st–3rd; 4th–10th; > 10th), and the number of athletes per country/year was summed and categorized as “total number of athletes 1997–2000”; “total number of athletes 2001–2010”; and “total number of athletes 2011–2020”. The Chi-square test and Spearman correlation were used to verify potential associations and relationships between variables.

Most of the athletes were from countries with medium HDI, followed by low HDI and very-high HDI. Chi-square test results showed significant differences among females (χ² = 15.52; P = 0.017) and males (χ² = 9.03; P = 0.014), in half-marathon and marathon, respectively. No significant association was verified between HDI and the total number of athletes, but the association was found for the number of athletes alongside the years (1997–2000 to 2001–2010: r = 0.60; P < 0.001; 2001–2010 to –2011–2020: r = 0.29; P < 0.001).

Most of the athletes were from countries with medium HDI, followed by those with low HDI and very-high HDI. The Matthew effect was observed, but a generalization of the results should not be done.

The relationships between anthropometric and training variables with running performance were previously investigated. However, it is possible that through the moderating role of anthropometric variables, the magnitude of the relationship between training and performance may be changed. The purpose of this study was to estimate the mediation role of body mass index (BMI) in the relationship between training volume and running performance among non-professional runners, taking into account sex and age category.

The sample comprised 1151 non-professional road runners (61.8% male), aged 18–72 years. Information about sex, age, body mass (kg), body height (cm), running pace, motivation for running, training volume and frequency were obtained through an online questionnaire. Taking into account athletes’ age, they were split into two age categories: “until 34 years” (adult runners) and “≥ 35 years” (master athletes). A mediation analysis was computed in Macro Process (SPSS 26), considering sex and age category. BMI was the mediating variable, while training volume/week was used as independent variable, and running pace was considered as dependent variable.

For both sexes and age categories, a significant association between training volume on running pace was observed [male adult: β = − 0.67; 95% CI (− 1.04 to − 0.53); male master: β = − 0.241; 95% CI (− 0.44 to − 0.26); female adult: −0.83; 95% CI (− 1.25 to − 0.41); female master: − 0.76; 95% CI (− 1.09 to − 0.44)], as well the association between training volume and BMI; and running BMI and running pace. Except for female adult runners, a mediation effect of BMI was observed.

The present study showed that BMI mediated the association between training volume and running pace in non-professional runners of different sexes and age categories. On the other hand, a small influence of training volume on the expression of BMI was found.