Can We Agree On How To Measure The Similarity Of Dive Profiles?

[agilis]

I'm sure you are not as obtuse as you pretend to be in your pleonastic posts. I write one thing, you seem to infer something else. Perhaps it's my poor writing skills, or your attempt at providing amusement.

In any case, I am reasonably familiar with basic programming, though my training and involvement took place long ago, in the punch card and main frame eras. I'm not sure why you might conclude that I "don't know about" calculating relationships between sequences. I did not write anything to suggest such a benighted perspective.

Perhaps someone who held such militantly ignorant opinions treated you cruelly, took your parking space, cut in line in front of you at the bakery, and in your grief you have transposed their opinions to me, despite my innocence in that connection.

I was really doing nothing more than pointing out that graphing relationships in as incredibly broad a context as dive profiles is of vanishingly limited utility when attempting to develop useful generalities related to human physiological reactions while scuba diving.

You might reread and rewrite your last sentence. You begin by setting up the sentence structure for a metaphor "...would be about as utilitarian...as" but fail to follow through after "angels", changing horses in mid-stream by omitting the expected comparison and inexplicably questioning the historical existence of what was intended as an absurdity that paralleled your irrelevant DNA type examples, and then citing a dim textbook classification of error.

Please excuse me from any further discussion. I have to adjust the thermostat.

 

[DiveNav]

The dive is similar until one stars applying different decompression models to it. If one decompression model calls for 30 minutes more deco than another - IMHO those decompression models are NOT similar.

Maybe I missed something ..... are we trying to compare dive profiles? or decompression algorithms?

 

[dmaziuk]

Maybe I missed something ..... are we trying to compare dive profiles? or decompression algorithms?

My impression is we're trying to compare two bent dead divers based on their (former) individual fitness, exertion, hypothermia, and phase of the moon...

@kr2y5: clearly, the answer to the thread's topic is "no".

 

[kr2y5]

Maybe I missed something ..... are we trying to compare dive profiles? or decompression algorithms?

In this thread, the question I was specifically interested in is, whether and how we can measure the similarity of dive profiles, in a way that may be practically useful. This discussion is inspired by a number of other threads, in which other, related questions (like that about deco strategies) were raised. Here, I hoped to specifically address just this one question: measuring profile similarity.

Here's how I hoped it might go:

(a) Throw a whole bunch of ideas on the table, evolve them into a few concrete proposals, debate their pros and cons, and their practical utility.

(b) Weed out proposals that obviously make no sense for one reason or another.

(c) If we have a few proposals left on the table that seem useful for different reasons, claim success and celebrate. Otherwise, if not satisfied, go back to (a), wash, rinse, repeat until we get bored.

Regardless of the outcome, I've been hoping that we can learn something in the process, which is why I think it'd be better to debate things that are concrete, rather than assert possibility or impossibility of something purely in abstract, without a compelling justification or reference.

One direction that was suggested earlier is that, if we pick some reference model M that expresses some value V of interest that we believe to be physiologically significant (in this case, it was tissue supersaturation, if memory serves me well) for each point in the dive profile, then we can measure profile similarity in terms of how much the values V differ throughout the dive (whether by maximum, integral, or whatever).

Obviously, a metric defined this way would only be useful to the extent that predictions of model M and the value V of interest are believed to be of any practical utility.

This approach generalizes in the obvious way if we have multiple such values of interest to track. In the absence of other proposals, I guess the most interesting question to me is, what other parameters (besides tissue supersaturation) might be of interest, and whether some of them are believed to be more predictive than others, and why.

 

[dmaziuk]

In the absence of other proposals, I guess the most interesting question to me is, what other parameters (besides tissue supersaturation) might be of interest, and whether some of them are believed to be more predictive than others, and why.

That's not what you asked, though. You asked about dive profiles. The primary data for that is pressure and time, usually plotted as depth on y axis and time on x.

For "other parameters" dive computers normally record water temperature, AI computers record tank pressure, and I think some dive computers include heart rate monitors of unknown accuracy. I don't know of any other parameters you could reliably collect using today's common instrumentation. So realistically temperature is a parameter that should be readily available and AC is the one available sometimes, and anything beyond that is anecdotes.

 

[kr2y5]

That's not what you asked, though. You asked about dive profiles. The primary data for that is pressure and time, usually plotted as depth on y axis and time on x.

For "other parameters" dive computers normally record water temperature, AI computers record tank pressure, and I think some dive computers include heart rate monitors of unknown accuracy. I don't know of any other parameters you could reliably collect using today's common instrumentation. So realistically temperature is a parameter that should be readily available and AC is the one available sometimes, and anything beyond that is anecdotes.

I meant values of interest that can be calculated from a dive profile according to some model, and correspond to something physiologically meaningful. AFAIK, tissue supersaturation falls into that category. Take model M that consists of a bunch of tissues, along with whatever parameters we care to feed into it, plug in the profile, parse the result. Is there any other physiologically meaningful value of interest that can be modeled?

 

[boulderjohn]

AI computers record tank pressure, and I think some dive computers include heart rate monitors of unknown accuracy. I don't know of any other parameters you could reliably collect using today's common instrumentation. So realistically temperature is a parameter that should be readily available ....

And each of these is of questionable value.

Tank Pressure: what value would that have?

Heart rate:. I know some computers check that, but I have no idea what value that provides or how it is incorporated into an algorithm. What does an individual heart rate tell us about the diver? My heart rate starts slow and does not go up a whole lot from there. When I am doing a normal exercise routine, my heart rate jumps all the way up to my wife's resting rate. When I am pushing really, really hard and gasping for breath, my heart rate will get up to about the level many people reach when they are warming up.

Temperature. What is important is the temperature of the diver, not the temperature of the water, and short of placing a probe in an orifice and connecting it to the computer, I am not sure how you will get that. The first time I dived in Hawai'i, with water temperatures in the 70s, I foolishly wore only a 3 mm suit and was shivering by the end of the dive. The first time I dived in Puget Sound, with water temperatures in the mid 40s, I foolishly wore my heaviest dry suit undergarments and a thick hood, and I was sweating by the end of the dive.

 

[agilis]

And each of these is of questionable value.

Tank Pressure: what value would that have?

Heart rate:. I know some computers check that, but I have no idea what value that provides or how it is incorporated into an algorithm. What does an individual heart rate tell us about the diver? My heart rate starts slow and does not go up a whole lot from there. When I am doing a normal exercise routine, my heart rate jumps all the way up to my wife's resting rate. When I am pushing really, really hard and gasping for breath, my heart rate will get up to about the level many people reach when they are warming up.

Temperature. What is important is the temperature of the diver, not the temperature of the water, and short of placing a probe in an orifice and connecting it to the computer, I am not sure how you will get that. The first time I dived in Hawai'i, with water temperatures in the 70s, I foolishly wore only a 3 mm suit and was shivering by the end of the dive. The first time I dived in Puget Sound, with water temperatures in the mid 40s, I foolishly wore my heaviest dry suit undergarments and a thick hood, and I was sweating by the end of the dive.

In addition to body temp and heart rate there are other important variables. To list just a few: alcohol/drugs/ medications in subject system, stomach contents being digested, age, existing chronic illnesses, more recently contracted diseases, lung function, circulatory system issues, ear and sinus problems, condition of teeth and fasciomaxillary area, and degree of fear/anxiety. These are just off the top of my head. There are many more.

 

[kr2y5]

Do we believe that the impact of all those factors mentioned above on the outcome of the dive, including anxiety, stomach contents, and the frequency of bowel movement, might be at the level comparable to the impact introduced by the differences in the dive profile? I don't know the answer to this question, but it has been my impression based on what I've read that we don't believe this to be true. And if this is, in fact, true, what is the point of calculating decompression schedules, and why would a couple of minutes here or there make any difference at all? We can't simultaneously hold these two opposing beliefs: that (a) infinite human variability affects the dive outcomes so much that you can't deduce anything of value whatsoever out of a dive profile without considering all the specific individual circumstances, and (b) that when the time comes, you need to haul your ass and stick to the ascent schedule by the minute, or else your risk of getting bent significantly increases.

 

[boulderjohn]

Over the decades, what research has been done has been done by using a lot of test subjects, with the assumption that all of those personal variables will even out. We have not even mentioned perhaps the most important of the personal factors--patent foramen ovale (PFO). In all of those tests, we must assume that the same percentage of test divers had undetected PFOs (or other issues) as are found in the general population. We can then as individuals make our own judgments.

When I dive, I dive knowing that in terms of whatever test populations have preceded me, I am older than that average, I am probably reasonably average in terms of fitness (even considering my age), and I know I do not have a PFO. I have a number of other health issues that I can throw into the mix as I make my decisions.

So when I am choosing decompression algorithms (and this is very much a key subject for me), I take all of this into consideration. That is why a word like "similar" has absolutely no meaning to me in comparing dive profiles. I will want to look at the specific details, and then I will decide--based on MY criteria--how similar they might be, and how important that similarity is for me.