UTD Ratio deco discussion

[huwporter]

It is probably worth also explicitly stating (since people do love to put two and two together and come up with 100) that average depth in isolation is far from a complete way of measuring the stress of a profile.

I mean you could do 60 minutes at 100m, then ascend directly to 1m and stay there for 24 hours and have a nice shallow average depth, but I hope no-one would actually recommend that as a sensible profile...

 

[Kevrumbo]

Except that the bubble models claim exactly the opposite - i.e. that their bubble/deep stop profiles are safer despite having a deeper average depth.

This seems to me to be so obvious that I am really completely baffled that it is an issue that is open for debate. Following a deep stops philosophy without adding an additional significant amount of time on shallow stops will certainly create a deeper average depth, and if you think that is important, then, well, ....

Well it seems obvious in hindsight. . .

Bruce Wienke of the RGBM Model had a seemingly compelling argument in this monograph on deepstops:
http://tecvault.t101.ro/Deep_Stops-BW.pdf

The first four paragraphs in the above linked article are particularly damning and now "cringeworthy" (the claim that deepstops can result in the overall shortening of decompression times -no wonder we haven't heard from BRW in a while. . .) all in the context of the unexpected results of the NEDU Deepstops Study.

 

[elmo]

The trend away from deep stop profiles, whether they are calculated by UTD Ratio Deco, GFs, VPM, RGBM, BVM, etc., is certainly understandable following the results of the NEDU study. However, most of us are still diving something like Buhlmann GF profiles where GF-low is substantially lower (typically half to two-thirds) than GF-high. Something in the ballpark of 40/75.
I used Subsurface to match as closely as possible the shallow and deep stop profiles from the NEDU study with GF and VPM-B deco models, and the "fit" is not what many people would expect.
The shallow stops profile fits a Buhlmann GF 100/44 plan very closely. Yes, the numbers are the right way around: GF-low is more than twice the value of GF-high! I know nobody that plans dives with anything like that.
The deep stops profile gets an ok match (bubble model fans may disagree) to either Buhlmann GF 20/65 (at sea level), or VPM-B +4 (most conservative setting in Subsurface and GUE DecoPlanner) with the altitude set to 9500ft/2900m to increase the conservatism further. I say bubble model fans may disagree because the NEDU deep stop profile is nearly consistently 10ft deeper than either the GF20/65 or VPM-B +4 profiles between 50-115 minutes (i.e. the "intermediate stop" part of the ascent schedule).
Personally, I claim not to favour deep stops or bubble models (I typically plan with GF 60/85, but will accept anything more conservative my buddies propose), but ask me whether I'd rather plan and execute a dive with GF 100/44 or 20/65, and I'd choose the latter any day. I suspect most people in this thread would do likewise.

Maybe we should be getting shallower even earlier than we think.

All profiles plotted in Excel below

 

[huwporter]

The shallow stops profile fits a Buhlmann GF 100/44 plan very closely. Yes, the numbers are the right way around: GF-low is more than twice the value of GF-high! I know nobody that plans dives with anything like that.
The deep stops profile gets an ok match (bubble model fans may disagree) to either Buhlmann GF 20/65 (at sea level), or VPM-B +4 (most conservative setting in Subsurface and GUE DecoPlanner) with the altitude set to 9500ft/2900m to increase the conservatism further.

As a counterpoint to this; here are the Buhlmann/gradient factors for each stop of the actual profiles in the study (note: Buhlmann wasn't used to generate these profiles, this is just an analysis of what Buhlmann theory thinks of the profiles that were actually tested):

Shallow stops:

First stop is ~GF93, surface is ~GF43

Deep stops:

First stop is ~GF42, surface is ~GF70

Interestingly, Buhlmann kinda thinks both these profiles are 'OK' (i.e. they both surface less than GF100) however the cold and work levels must have pushed up the aggressiveness to the point of observable DCS, and the lower early GFs of the deep stops profile doesn't compensate for the higher GF at the surface.

(Note: Ascent time between stops is added to the shallower stop since ascent time is not included in the NEDU stop times published in the paper.)

 

[ajduplessis]

Except that the bubble models claim exactly the opposite - i.e. that their bubble/deep stop profiles are safer despite having a deeper average depth.

Except none of the models availble today will give you a 20min deepstop for the dives as used in the study, unless you inflate the parameters way outside the realm of daily norm.

The reverse also hold true, you cant just go shallow (last stop) immediately either.

The answer is somewhere in between.....but both sides argue one is wrong when in reality neither really knows.

 

[Dr Simon Mitchell]

Personally, I claim not to favour deep stops or bubble models (I typically plan with GF 60/85, but will accept anything more conservative my buddies propose), but ask me whether I'd rather plan and execute a dive with GF 100/44 or 20/65, and I'd choose the latter any day.

Hello Elmo,

I put it to you that you are only saying this because some instructor / course somewhere has deeply ingrained the notion that the deepest stops on a decompression dive should be deeper (perhaps much more so) than prescribed by the original Buhlmann algorithm (represented by a GF Lo of 100). And I conclude it is an "instructor or course" that has imbued you with this belief because you won't find any proof of it anywhere in the diving science literature. Indeed, as I have described elsewhere in this thread, there is a growing body of evidence that is pointing away from deep stops as deep as those prescribed by bubble models or ratio deco (and probably 20/65). That is not to say that all the way back to a GF lo of 100 is the correct way to go, but I would not dismiss the idea out of hand. Buhlmann obviously thought it was correct. And importantly, if you read the NEDU study carefully, you will find that the NEDU shallow stops profile (and therefore the 100/44 GF combination you are deriding) actually produced the least cumulative supersaturation across representative fast and slow tissues of over 500,000 different ways of decompressing from that dive (which would have included the 20/65 approach).

Maybe we should be getting shallower even earlier than we think

Yes indeed. But in the absence of really solid data defining exactly how shallow, it is sensible to back away from deep stops cautiously (which I believe you are doing appropriately with your 60/85 approach).

Simon M

 

[Dr Simon Mitchell]

Except none of the models availble today will give you a 20min deepstop for the dives as used in the study, unless you inflate the parameters way outside the realm of daily norm.

You need to re-read David Doolette's explanation of why they "inflated the parameters way outside the realm of daily norm". Until you take the time to get your head around this, you will never understand the NEDU study. You can find the explanation here:

Deep stops debate (split from ascent rate thread) - Page 14

The answer is somewhere in between.....but both sides argue one is wrong when in reality neither really knows.

Actually, I think it is fair to claim there is now a fairly convincing evidence-based signal that in significant decompression dives, deep stops prescribed by bubble models or ratio are too deep. In contrast, the side arguing for deep stops has no evidence whatsoever that they are necessary. Nevertheless, I agree that there are things that "neither side really knows". In particular (as I alluded to in my response to Elmo above), we don't have strong evidence defining how far we should back away from current deep stops practices.

Simon M

 

[Patoux01]

Well it seems obvious in hindsight. . .

Bruce Wienke of the RGBM Model had a seemingly compelling argument in this monograph on deepstops:
http://tecvault.t101.ro/Deep_Stops-BW.pdf

The first four paragraphs in the above linked article are particularly damning and now "cringeworthy" (the claim that deepstops can result in the overall shortening of decompression times -no wonder we haven't heard from BRW in a while. . .) all in the context of the unexpected results of the NEDU Deepstops Study.

Actually, hadn't he also predicted deep stops to fail? from your link

With higher incidence of surface decompression sickness, as expected, the Australians devised a simple, but very effective, in-water recompression procedure. The stricken diver is taken back down to 30 ft on oxygen for roughly 30 minutes in mild cases, or 60 minutes in severe cases. Increased pressures help to constrict bubbles, while breathing pure oxygen maximizes inert gas washout (elimination). Recompression time scales are consistent withbubble dissolution experiments

 

[ajduplessis]

You need to re-read David Doolette's explanation of why they "inflated the parameters way outside the realm of daily norm". Until you take the time to get your head around this, you will never understand the NEDU study.

Simon M

Yes agreed, one needs to do a lot of explaining to argue that a dive profile outside of the norm is not a good idea whislt deriving a value (percentage) to indicate possible risk increase .

 

[victorzamora]

Except none of the models availble today will give you a 20min deepstop for the dives as used in the study, unless you inflate the parameters way outside the realm of daily norm.

I've heard this one a lot, but I've honestly never understood it. We all know that working at depth makes you on-gas more, right? We all know that being cold on deco makes you off-gas less, right? The algorithms chosen HAD to compensate for the high workloads at depth and the cold deco. This would make deco NOTABLY longer than "normal" tech-diving models would suggest.

Why longer? To make the deep vs shallow profiles different. Why different? That's literally what they're testing.

One of the commonly brought-up points in previous threads has been that the NEDU study had far too much deco. The proof was running VPM or Buhlmann with "normal" conservatism and claiming that that'd be "enough" or "appropriate" deco. The proof that the NEDU study clearly did NOT have too much deco, however, is in the pudding: Real-life human beings got bent at a rate notably higher than what I'd consider acceptable in my diving.