Current arbitrary guidelines for between-limb performances are often thought to ideally be less than 10% of one another. For example, if an athlete jumps 40cm in a single-leg countermovement jump, but only 32cm using their left leg, this would indicate an inter-limb asymmetry of 80% (32 / 40 * 100 = 80). So, the question here would be: should we attempt to reduce this asymmetry?
Some studies have shown that kicking accuracy and resultant ball speed is negatively affected by larger asymmetries, suggesting that reducing the asymmetry ‘may’ improve kicking performance. However, in my opinion, attempting to measure a complex skill such as kicking a ball and relating it to limb strength is always going to be problematic. Athletes are often better at kicking with their dominant leg due to several neurophysiological adaptations/skills which are specific to that movement (e.g. intra-musculature coordination and the magnitude and direction of force production). So to me, a score on a balance, or strength test, will have little validity with kicking a ball.
With regards to return to play parameters (e.g. ACL rehabilitation), I believe the context is very different, and I would most certainly be looking for a base level of symmetry in things such as balance, strength and power. And although there is minimal research supporting it, I believe symmetry of less than 10% is an acceptable range, in most cases.