Correct, and this is a clever analogy

.
In general (not just necessarily in the context of ME/CFS) I think this is possible if something is wrong with the Complex V machinery itself (we don't know if this is the case or not in ME/CFS). But I think this would be pretty apparent and accompanied by other abnormalities, and so shouldn't confound anything - you would note it and look for how it is occurring. Plus, severe (eg: genetic) issues with Complex V are often lethal or present from birth.
We speculated what the cause of inefficient ATP synthesis in lymphoblasts could be, based on our knowledge at the time in our first ME/CFS study:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7036826/ , paragraph starting with
"What might cause such a mitochondrial Complex V inefficiency?". Sorry if this is a hassle to read, it felt like bad etiquette to paste in a text block of my own findings in a thread about somebody else's research. Anyway, the tldr is: a mutational defect seems less likely than dysregulation.
Off the top of my head... we have measured it by luciferase luminescence, and so have Lawson
et al 2016. But these were steady-state measurements of ATP levels, not sure if other techniques measuring the rate of ATP synthesis have been employed. I might be forgetful though. Vermeulen
et al 2010 also measured it but the citation for their method is throwing up a dead end for me so I can't really comment. You are right, the findings do vary, but I think these can be explained by differences in cell/tissue type or methods employed, etc. Different-shaped pieces of the same puzzle.
Oxidisable substrate provision is relevant to the IACFS talk I just gave, the results from which are unpublished. I know Cort was covering the conference but not sure if he's written up anything that I spoke about yet. I did post a little about it previously but I don't want to discuss too much detail informally before it's all finalised. That discussion alongside our new results will be in everybody's hands in the future.
I can make a couple of observations more freely based on what is already published:
-this is another study showing no changes in the rate of glycolysis in seahorse, which is in line with lymphoblasts, PBMCs, NKs (sans glycolytic reserve) so now seems pretty consistent across a few groups using seahorse with different sample types.
-Normal basal OCR and absolute rate of ATP synthesis, and elevated maximum OCR with galactose reported here, which for these parameters is the same pattern in lymphoblasts. Unchanged ROS, also the same as lymphoblasts. Glucose utilisation shows some differences in oxphos, but I need to think more about this. Could very well be an upstream defect as suggested. Glucose can also be increasingly or decreasingly utilised by the pentose phosphate pathway - this can also affect many relevant things such as redox balancing and pyruvate supply, which in turn affect mitochondrial function. Maybe another pathway to have a read about for you guys, if unfamiliar. I touched on it in the IACFS talk.
-Would be interesting to see how the cells use glutamine as well given previous metabolomic reports - through GDH and also through glutamate to aspartate (AST) - both routes assist with the provision of reducing equivalents (in different ways) to drive OXPHOS. There are other relevant amino acid degradative pathways (stay tuned) but glutamine is used the most by the cell, and has been discussed the most in the field.
-It is suggested that a metabolic inflexibility is present and lies outside of AMPK, since AMPK is proposed to behave normally. This is interesting since the chronically activated TORC1 signalling seen in lymphoblasts would likely contribute towards metabolic inflexibility. Different tissue types though, but interesting to think about.