Computer Algorithms, Objective Evidence of Whether One is "Safer"

[halocline]

Is there OBJECTIVE evidence, either statistical (incidents of DCS per diver using the algorithm) or medical (doppler studies) showing the relative effectiveness of the different common algorithms in computers.

Thanks!

I believe that the only true evidence would be statistical incidence of DCS associated with specific computer algorithms. Unless something has changed recently, there is none. The incidence of DCS in recreational diving is extremely low and as has been pointed out, there are so many factors for DCS other than dive profile that its very difficult to establish cause in many of the already-rare cases.

If your goal is to feel confident that you're diving as safely as possible, I would suggest that the best approach would be to study some basic deco theory so you understand the concept of tissue compartments and variable rates of N2 loading/off gassing. All the models are based on that. Equally beneficial would be to have a good basic understanding of non-dive factors that are shown to have an association with DCS. You don't need phd level-study, just some basic knowledge, and that can guide your behavior.

Dive computers are a great convenience, but just like anything else in diving, it's only gear, it has limitations, and it can fail. This is where dive planning and behavior based on knowledge is so valuable.

 

[guyharrisonphoto]

So, the thread has actually not revealed much. Many of the posts contain the usual, well thought out, commonly understood wisdoms about learning deco theory, the unpredictability of deco and the need to adjust for various factors. I have already done this, and none of it answered my question.

The point of the thread is to determine which baseline algorithm(s) are objectively “safe”. Once you settle on a baseline algorithm, all new computers that I know of allow you to adjust the algorithm to your personal needs/wants/dive conditions for more conservatism, but that goes to the subjective realm of estimates, theory, philosophy and concepts of risk tolerance, none of which I wanted for evaluating which baseline algorithm to begin from.

Some facts or statements, however, seem very interesting:

1. Per the paper that was linked, manufacturers do not test their algorithms by medical (Doppler) means or statistical analysis of DCS percentage of divers using their computers. Is this true?

2. No outside organization like DAN has compiled enough data from DCS incidents to know whether “hits” were taken by people using their computers in accordance with the algorithm (ie “undeserved hits”) so as to be able to compare the different algorithms

3. At least one group of people tried something; comparing the actual known risk (based on medical and statistical evidence) of certain dive profiles to what different computers would have allowed (or not allowed) for those same profiles as conducted in a pressure test following those profiles.

4. Per the linked paper, even computers supposedly using the same algorithm can give different results, and this is especially true in the multi-day diving context. Also, in that same context, the differences between the baseline algorithms seem incredible. The more conservative algorithms really do impose an immense penalty on multi-dive profiles that are “safe” at least per the parameters of the paper.

All this being said, given that most rec divers follow computers, and the (apparently) complete lack of objective evidence that any one of them is unsafe, why would anyone NOT select the most liberal baseline algorithm, and then work backward to add conservatism based on personal factors/dive conditions.

 

[tursiops]

And what do you mean by "baseline"? And how would you choose it? And why do you think you can be safer than safe?
If this is really a concern to you, don't dive; baseline is no on-gassing.

 

[guyharrisonphoto]

The point of the thread, and my meaning, seems very clear to everyone else.

"Baseline" is what profile the computer algorithm gives you without manually adjusting the computer for conservatism

How I would choose it is the entire point of the thread: the one that gives you the most bottom time and is still objectively "safe.".

It is not a "concern" as DCS is extremely low in rec divers no matter what computer is used. Given that most rec divers are using computers, that is an indication that computers as a whole are "safe" because they are not increasing the incidence of DCS. Or, maybe better stated, there is no evidence that a particular brand/algorithm is more responsible for the those cases that comprise the extremely low incidence of rec diver DCS than any other.

It is a point of curiousity, that maybe I have been short-changing myself by diving a baseline algorithm that unnecessaryly shortens bottom time for no objectively demonstrable benefit.

 

[Mike]

The real evidence would be the real world collection of data since 1965 of diving with dive computers and evidence that one brand has more divers taking chamber rides than another. I've never seen or heard hide nor hair of a whisper of one dive brand being more deadly or more apt to put you in the chamber. That's evidence enough for me that the more liberal dive computer is conservative enough, has it right and anything more conservative is simply wrong.

If we were living in a world where the reality was - "everybody knows that more people end up bent on Oceanics than any other dive computer" that would be one thing, since we don't live in that world and we live in a world where no more people end up bent due to owning an Oceanic than who own a Suunto that gives zero credibility to any notion that a Suunto is safer than another brand.

 

[lowviz]

So, the thread has actually not revealed much. Many of the posts contain the usual, well thought out, commonly understood wisdoms about learning deco theory, the unpredictability of deco and the need to adjust for various factors. I have already done this, and none of it answered my question.

The point of the thread is to determine which baseline algorithm(s) are objectively “safe”. ...//...

I won't annoy you with a MythBusters vid, but "There's your problem".

 

[JohnN]

The real evidence would be the real world collection of data since 1965 of diving with dive computers and evidence that one brand has more divers taking chamber rides than another. I've never seen or heard hide nor hair of a whisper of one dive brand being more deadly or more apt to put you in the chamber. That's evidence enough for me that the more liberal dive computer is conservative enough, has it right and anything more conservative is simply wrong.

If we were living in a world where the reality was - "everybody knows that more people end up bent on Oceanics than any other dive computer" that would be one thing, since we don't live in that world and we live in a world where no more people end up bent due to owning an Oceanic than who own a Suunto that gives zero credibility to any notion that a Suunto is safer than another brand.

I'm happy you gave the paper such a comprehensive and thorough read Mike.

Here is a reasonably close approximation of the dive from V-Planner:

V-Planner 3.91 by Ross Hemingway,
VPM code by Erik C. Baker.


Decompression model: VPM - B


DIVE PLAN
Surface interval = 5 day 0 hr 0 min.
Elevation = 0ft
Conservatism = + 2


Dec to 60ft (1) Air 50ft/min descent.
Level 60ft 53:48 (55) Air 0.59 ppO2, 60ft ead
Asc to 40ft (55) Air -30ft/min ascent.
Level 40ft 25:00 (80) Air 0.46 ppO2, 40ft ead
Asc to 20ft (81) Air -30ft/min ascent.
Asc to 15ft (81) Air -30ft/min ascent.
Stop at 15ft 10:30 (92) Air 0.31 ppO2, 15ft ead
Surface (92) Air -30ft/min ascent.


OTU's this dive: 13
CNS Total: 7.7%


159.6 cu ft Air
159.6 cu ft TOTAL




DIVE PLAN COMPLETE

My take-away is you'd be foolhardy to do this dive without a deco stop.

Of course you know better

 

[TSandM]

It's a good question. To my knowledge, the only computer that ever got a reputation for being unsafe was one of the very early ones, which was nicknamed the "Bendomatic". There have also been some questions about the Cochran computer algorithms. But the majority of recreational dive computers clearly operate within bounds which produce a very low risk of DCS. And I'm not sure if we even know what alterations to the algorithms really result in lowering that risk, except that shorter, shallower dives are going to be safer than deeper or longer ones.

 

[Lorenzoid]

I agree that this thread has not revealed much. I'm wondering if you knew the answer to your question when you first posted it. If there were conclusive data available, I suspect it would be cited all the time on SB by proponents of one algorithm over another.

All this being said, given that most rec divers follow computers, and the (apparently) complete lack of objective evidence that any one of them is unsafe, why would anyone NOT select the most liberal baseline algorithm, and then work backward to add conservatism based on personal factors/dive conditions.

Although there are a few outliers, for most computers it's hard to say which computer has "the most liberal baseline algorithm." From what I understand, there are a number of computers that are marketed mainly toward experienced or tec divers, such as the Shearwater, which allow divers a relatively wide range of adjustability and flexibility. I would say quite a few divers are doing pretty much just what you suggest.

 

[JohnN]

Although there are a few outliers, for most computers it's hard to say which computer has "the most liberal baseline algorithm." From what I understand, there are a number of computers that are marketed mainly toward experienced or tec divers, such as the Shearwater, which allow divers a relatively wide range of adjustability and flexibility. I would say quite a few divers are doing pretty much just what you suggest.

I'm not sure I can agree with that. If I'm reading the original study properly, and assuming VPM-B, +2 conservatism is the "gold" standard. That dive would have 10:30 of deco time.

The computers doing the same profile gave between +20 (dive still had 20 minutes of dive before any deco obligation) to -19 (19 minutes of deco).

That's quite a lot of variation