Talk:Kled/@comment-24951773-20160927091604/@comment-28977071-20161028211542

This is not just a matter of subjective preference, but also of objective optimality.

Regarding the damage carries, they fall into two basic categories, burst and DPS. If you optimized the burst champion, then you could really seek only the pure damage to max the carry potential. But in case of DPS carry you have to optimize the product of DPS and the time that you were allowed to apply that DPS. That time is dependant on your sustain, hence defense effectively (albeit indirectly) increases your damage output.

Now if you had 0 of both offensive and defensive stat and to invest into them, you'd wish to optimize something like OxD(G-x), where O and D would represent the offense and defense exchange ratios when trading  for the stat, respectively. As they are cca. constant, they can be omitted from the product during the optimization and you are left with simple x(G-x), where x'd be the invested into defense. This can be solved in various ways, but put simply it is a maximization of rectangle area x(G-x) with constant circumference of 2G, so naturally the optimal solution is a square. Hence you'd theoretically invest half of the into offense and the other half into defense to maximize the DPS carry's output.

Practically, however, you are starting with some stats already and the existence of human reflexes makes the burst slightly increased in value, so the scales are naturally shifted a bit more in favor of investing gold into offense if one wished to reach optimum.