Interestingly the NEDU study wasn't the first time we'd heard that bubble models were calculating incorrect shallow stops. This is what Mark Ellyatt wrote about his experiences trying RGBM on the first generation Abyss computer in (I believe) 2003:
I had used the RGBM algorithm many times before, both shallow (60-90m) and deep (120m-260m), contrary to Bruce Wienke's website I received DCS every time using the RGBM algorithm. Satisfactory results were obtained only after manually adding significant extra shallow decompression stops, usually in the order of hours of extra time. I have used the VPM a/b algorithms since they became available, also with similar results, only the version using VPMbE (2005) with maximum conservatism provides acceptable decompression times.
Granted Mark wasn't making the same kind of dives most of us are but due to the fact that his dives were that much more extreme, we have actual evidence that RGBM and VPM weren't calculating long enough shallow stops for at least this one diver ... The interesting thing about this to me is that the NEDU study seemed to suggest the same conclusion that had been reached a few years early by people trying it in practice.
Hey Diver0001,
With regard to sufficient shallow time, I’ve seen it posted a lot that VPM "scales consistently and in the right proportions" as you go deeper and stay longer. I have another perspective on that idea. See the chart below.
Each of the lines shown on the chart represents the allowed supersaturations for each compartment upon surfacing from dives with progressively longer bottom times (all 200ft dives in this example).
For this discussion the actual dive parameters (i.e. CCR or OC, depth, set point, gases) don't matter much. This pattern is reproducible in deeper dives when the effect of the critical volume algorithm (CVA) in VPM wears off. The CVA is that part of VPM that allows it to better reflect NDLs and lighter deco requirements. The important thing to understand about the chart is that we're seeing how VPM-B adjusts in the face of increasing gas loads due to the longer bottom times and deeper dives many tech divers are familiar with.
I added the black-dashed line to highlight something. Notice that the peak supersaturation that the diver is allowed to surface with doesn't change (or changes very little) as the bottom time increases (following CVA wear-off). Even staying within VPM theory you have to see the problem with that.
Suppose you execute the 40 min dive. Based on the chart the peak SS at surfacing is in compartment 11. If you move to a bottom time of 60 min C13 has the peak SS at surfacing. C13's half-time is about 63% longer than C11's. That is, C13 will stay supersaturated far longer because it's a slower compartment. But what has VPM done? It's kept the supersatuation at which the diver is allowed to surface the same (oh, there may be minor differences, but they are effectively equal as shown on the chart). VPM methodology allows the same supersaturation upon surfacing even though it must be endured for a much longer time. This approach can only do one thing ---
increase risk. The diver, then, must know to increase VPM's conservatism or substantially pad shallow stops to adjust for this, or accept the certainty of increasing risk, because VPM is doing nothing on its own to make that adjustment. The lack of sufficient shallow time is directly related to this "feature".
So in response to the idea that VPM "scales correctly and in the right proportions" I would say the following. It's not so much that VPM doesn't scale correctly,
it's that it doesn't scale AT ALL, at least in the sense discussed above. There is no attempt in the face of mounting gas loads in slower compartments to decrease allowed supersaturations to reflect the far longer SS exposure.
In comparison, GF does reduce allowed surfacing supersaturations as slower compartments enter the picture because Buhlmann's M-values get smaller as compartment half-times get longer. I'm not saying it adjusts perfectly to keep risk the same. But it does reduce allowed supersaturations as more gas is loaded in slower compartments. So in that sense GF "scales better and in better proportions" than VPM.
For comparison, see the charts below for GF and for a Navy Mk16 set of dives. Notice the negative slope in peak allowed surfacing SS as dives become more extreme and slower compartments become more important. Again, I can't claim the adjustment is enough to keep risk the same, but at least they're playing the right game.
For completeness, this pattern occurs when dives are extreme enough to minimize the effect of the CVA and long enough so that, for any given depth, the first stop isn't changing drastically as time increases-- basically technical dives with longer decompressions.
