This study aimed to identify tests that include all the criteria necessary to characterise reliably the personality of horses on a large scale: stability of measures across situations and over time, evaluation of independent dimensions, links with use and easy to organise for a large number of animals. For this purpose, two independent experiments were carried out, involving respectively 24 (experiment 1) and 15 horses (experiment 2). Existing tests were adapted to facilitate their use during breeding horse shows, while their owners held their horses in hand. They were called the “Simplified Personality Tests” (SPT). They involved three fear tests (novel surface, novel object and suddenness), two tactile sensitivity tests (Von Frey filament and hip-stifle axis stimulation) and behavioural measurements conducted during standard horse show tests (assessment of horse conformation and jumping ability). Following the first experiment, it was demonstrated that the measurements carried out during two of the fear tests were significantly correlated (correlations between “novel surface” and “suddenness”: R = 0.42, p = 0.04, N = 24), as was also the case for the two tactile sensitivity measures (correlation between “filaments” and “hip-stifle axis” R = 0.50, p = 0.01, N = 24). However, the fear and tactile sensitivity measures were not correlated, indicating that they reflected independent dimensions. To test for stability across time, we checked whether the measurements performed during these SPT correlated with equivalent measurements recorded from another series of tests (“Complete Personality Tests”) obtained one month earlier and whose stability over several years had previously been demonstrated. These correlations were all significant except for the “novel object test”. Thus, except for the latter test, all measures were stable across situations and over time and evaluated two independent dimensions: fearfulness (disposition to react to a greater or lesser extent to new or sudden stimuli) and tactile sensitivity (disposition to react to a greater or lesser extent to tactile stimulation). To determine the relationship between these measures and the horse's behaviour while competing in show jumping competitions, we observed horses ridden during training sessions or competitions one month (experiment 1) or one year (experiment 2) after the tests. Principal component analyses showed that such relationships existed: the most fearful and to a lesser extent the least sensitive horses were more difficult to ride, but performed better when competing because they knocked down fewer bars. Finally, these tests were conducted on more than 650 horses over three years, demonstrating that they are easy to organise on a large scale.