It is essentially the ratio between the treatment effect and the error variance. The larger the treatment effect in relation to the error variance, the larger the F-statistic and the lower the probability that that F-statistic occurred by chance given that the null hypothesis of no-effect is true.