About French Comex/MT92 tables:
- These are mostly statistical. COMEX (JP Imbert and B Gardette) used a databank of 64,000 (sixty four thousand, recorded by the COMEX) commercial dives (using air only) to improve the MT74 tables. Whatever theoretical model they used was a "black box" useful mostly to handle the data (the authors state that "they doubted of the predictive value of analytical models").
About deep air diving and DCS risk:
- French Navy itself reckons that for dives deeper than 40m/130ft there is an incidence of 1 DCS/3,000 dives, while for dives shallower the rate is 1/30,000 (using French Navy MN90 tables, and air only, also for deco). Their "deep stops" protocols were an attempt to improve this.
- Data of the COMEX about DCS cases show a similar increasement of 10 times of the DCS occurence when the severity of the dive (correlated with the depth, as they define severity as depth * square root of bottom time) increases from "mild" (eg 20' at 39m/130 ft) to "medium" and finally to "severe" (eg 20' at 60m/200ft).
So for sure diving deeper than 45 m/150ft with air increases the DCS risk.
I'm not advocating for air being better than trimix for the 40m/60m range, far from it, but Helium comes at a price and has also its drawbacks.
BTW deeper means more risk, all else being equal.