I expect that's getting to be more like two standard deviations. However, I'd give an allowance of an extra 10-20% to compensate for a lack of diurnal variation. That is, if the average healthy young guy makes 6.5 mg of testosterone daily, then that average free testosterone of 14-15 ng/dL refers to the morning peak. The evening trough could easily be 20-40% lower. If we assume peak testosterone is the important parameter then a guy on TRT using relatively frequent injections of a longer testosterone ester would need to use a little more testosterone to reach the peak value, because it's also close to his average value, given that he has little fluctuation in serum testosterone. In the case of the exactly average guy it means his free testosterone on TRT is 14-15 ng/dL all the time; it doesn't decline as the day goes on. Maintaining the level all day then requires a bit more testosterone than the 6.5 mg per day.

Having argued that free testosterone is proportional to production rate or dose rate, I agree with this logic—subject to the adjustment for lack of serum variation when needed.