how can 'they' make a wrong calculation for call drop rate

Oh man...!!! its truly foolish man that always make a wrong calculation, isn't he/she don't learning math or calculus when they get in school or university??the simple question for 'them', there are a lot of data percentage, can average that percentage to get new percentage statistic..?oh its really bad idea..
in my opinion, its cannot be a new percentage is we average a lot of data percentage, cause there are too many dividend if we calculate one by one, let we take one simple case, if we have a normal release about 1000 a day and have a radio drop about 5, it say we have 0.5% radio drop, and we have 2000 normal release and 5 radio drop in second day, that we can say we have 0.25% radio drop in second day, if we just calculate directly with just average the percentage how many percentage we get?
average(0.5%+0.25%), its equal 0.375% -> is it valid or not? i think not, the second way is we summary our normal release and radio drop first and take the average for this all summary let we calculate...
so we have 3000 normal release and have 10 radio drop, so average for this two day is 0.33%
so what kind of calculation that can near the true statistic..?
yah its can be a proof concept for 'they' that call himself a master...