Air drag typically increases approximately like $v^2$. The velocity of the pendulum scales like $L^{1/2}$. Therefore the force probably scales like $L$, and the work per oscillation like $L^2$. The energy in the pendulum is proportional to $L$, so the fraction of the energy dissipated per cycle should scale like $L^2/L=L$. The decrease in angle per cycle should therefore also increase in proportion to $L$.
↧