Son of Deep Stops *or* Waiting to be merged with the mother thread...

[leadduck]

Correct me if I'm wrong, but I seem to remember that Suunto's "RGBM" isn't really RGBM, but rather straight Bühlmann with some kind of fudge factor added to make it more conservative for short SIs and fast ascents. Am I totally lost here, or have I got a decent idea of the realities?

Check this out: http://njscuba.net/z-gear/deco_weinke_rgbm.pdf

f in eqs. (89)-(91) is like a gradient factor; not like Baker's GF_low/GF_high, but Wienke applies a factor per compartment, think GF_slow/GF_fast. The factor is computed internally. It's a penalty for previous bad behavior. Reverse profiles, repetitive diving, and multiday diving, reduce f ((94)-(99)). Maybe he used a bubble model to fit (94)-(99).

I guess you could do this to map VPM-B to ZHL+GF: Run a set of profiles with VPM-B+3 and for each find GFlow/GFhigh that reproduces a similar profile, then write a mapping function from (depth & bottom time) to GF. Now GF is not a user setting anymore, but is calculated internally.

 

[cerich]

I can't disagree with a word of that.

Nevertheless, I would be happy to post the pure Bruce and Tim recreational RGBM chart that I was given in my course. I'd happily share had it not been liberated by a bunch of Boy Scouts on the Diver II on the Deep 6.

Somebody has that table, please post it here...

here is a link to a email BW wrote in regards to some questions, in it he says " collectively have logged many 10,000s of technical and recreational dives with only a 2 reported cases of DCS So, RGBM DCS incidence rate is virtually zero, especially on the technical envelope where it matters most as a model test. "

Reduced Gradient Bubble Model

I personally, based on my experiences with BW would believe nor trust nothing he says however.

 

[rjack321]

here is a link to a email BW wrote in regards to some questions, in it he says " collectively have logged many 10,000s of technical and recreational dives with only a 2 reported cases of DCS So, RGBM DCS incidence rate is virtually zero, especially on the technical envelope where it matters most as a model test. "

Reduced Gradient Bubble Model

I personally, based on my experiences with BW would believe nor trust nothing he says however.

Lies, damn lies, and statistics

From your earlier cited Norwegian symposium, bottom of page 100.
"From a sample of 10,738 dives, dived with Bühlmann ZHL-16 or Wienke RGBM algorithms, 165 DCS cases were recorded, almost equally distributed between the two (1.35% vs. 1.75%)."

http://www.ntnu.no/documents/10301/340015/Proceedings+of+Validation+of+Dive+Computers+Workshop.pdf

Probably not statistically significant but once AGAIN a bubble model has worse outcomes than a straight dissolved model.

 

[Patoux01]

So, RGBM DCS incidence rate is virtually zero, especially on the technical envelope where it matters most as a model test. "

Mark Ellyatt?

 

[rjack321]

Mark Ellyatt?

I know I shouldn't have laughed but touche

 

[dmaziuk]

There has been a lot of statements by Ross to the effect that the scientific community with the exception of some other posters in this thread does not consider VGE as important to DCS and Ross argues VGE means nothing.
Here is some interesting reading...http://www.ntnu.no/documents/10301/340015/Proceedings+of+Validation+of+Dive+Computers+Workshop.pdf

Page 93 is a good discussion on VGE..

"specificity": what fraction of true negatives test negative
"predictive value positive": what fraction of positive tests are true positives

Seems to me one person argues the test has low predictive value positive while the rest are arguing that its high specificity weeds out true negatives.

 

[cerich]

"specificity": what fraction of true negatives test negative
"predictive value positive": what fraction of positive tests are true positives

Seems to me one person argues the test has low predictive value positive while the rest are arguing that its high specificity weeds out true negatives.

well it tells me that the ongoing assertions by one poster in this thread that the scientific community has dismissed VGE except for another poster in this thread is not quite accurate.

 

[rjack321]

well it tells me that the ongoing assertions by one poster in this thread that the scientific community has dismissed VGE except for another poster in this thread is not quite accurate.

You mean that its been an accepted metric of DCS stress for over 40 yrs except in one Canadian computer programmer's mind? You don't say?

 

[Shearwater]

You mean that its been an accepted metric of DCS stress for over 40 yrs except in one Canadian computer programmer's mind? You don't say?

To the best of my knowledge, Ross is actually an Australian living in Manila. He lived in Canada for a while. I don't know his citizenship status.

Bruce

 

[cerich]

You mean that its been an accepted metric of DCS stress for over 40 yrs except in one Canadian computer programmer's mind? You don't say?

which is funny when you look at how the Canadian DCIEM tables were validated...lots of VGE tracking and thermal considerations/testing, one of which said programmer says doesn't matter, the other same programmer seems to feel is not something that a algorithm can(should) account for.