I've been speculating similarly about testosterone: What if peak levels of LH are important, independent of average levels? This would explain why such relatively high doses of hCG are needed to get the effects we want, which in turn increase the risk of side effects. Very crudely, at least, the numbers may reconcile. It's claimed that on a per-IU basis, hCG is six to eight times more potent than LH. Dr. Saya's case study on hCG suggests a response area-under-the-curve (AUC) of ~6 mIU/mL*days for a 500 IU injection. One study showed normal intratesticular testosterone would be achieved with injections of 300 IU every other day. This would translate to average serum levels of 1.8 mIU/mL. Applying a factor of six or so is supposed to give an LH equivalent—10.8 mIU/mL. And this figure corresponds pretty well with peak LH values reported in this work. It's pretty tenuous evidence, and much more would be needed to say if the idea really has merit.