Average Depth Diving?

[jonnythan]

"Since it has worked for you so far, its clear that the math is accidentally close in the range that you’ve used it, you’ve been damn lucky that you’ve not stepped outside the range that it works"

Did you consider the possibility that the person you're addressing has actually run these exact profiles through a computer to confirm his "accidently close" math?

If you would be willing to accept that possibility for a brief minute, where is the luck?

 

[lamont]

"Since it has worked for you so far, its clear that the math is accidentally close in the range that you’ve used it, you’ve been damn lucky that you’ve not stepped outside the range that it works"

Did you consider the possibility that the person you're addressing has actually run these exact profiles through a computer to confirm his "accidently close" math?

If you would be willing to accept that possibility for a brief minute, where is the luck?

The problem is that you must get your answer to 20 decimal places from a computer or it just isn't accurate enough. Having a human do an estimate to within 10%, and then take steps to ensure a 10% safety margin on a mathematically model which is probably nowhere near 10% accurate in modelling human physiology is playing russian roullette.

 

[jonnythan]

The problem is that you must get your answer to 20 decimal places from a computer or it just isn't accurate enough. Having a human do an estimate to within 10%, and then take steps to ensure a 10% safety margin on a mathematically model which is probably nowhere near 10% accurate in modelling human physiology is playing russian roullette.

I think I'm beginning to understand the problem here

 

[rjack321]

I could likely hip shoot most of my decompression needs without looking at a computer or a set of tables, just from experience. But I would no more do that than dive your series of linear approximations.

Do you not trust your own judgement and experience?

Some of us have ample experience with this approach for both "no-deco" and deco dives and thus have a better understanding of our own bodies than some hypothetical compartment constructs might provide. BTW, I haven't felt the fatigue of subclinical DCS since I gave up my computer (and tables) and started diving gauge and timer only. Which was about 150 dives ago.

But frankly that does not stand the test of modern science, it’s just voodoo.

Its all voodoo.

 

[rjack321]

The problem is that you must get your answer to 20 decimal places from a computer or it just isn't accurate enough. Having a human do an estimate to within 10%, and then take steps to ensure a 10% safety margin on a mathematically model which is probably nowhere near 10% accurate in modelling human physiology is playing russian roullette.

Only 20? The right side of your gun is feeling very heavy...

:cwmddd:

 

[Charlie99]

We don't even have to use a computer to check some profiles. Just run the example profile first posted by BiggDawg and crank it through the PADI RDP and Wheel.

Do the numbers yourself:

• Dive at 100 feet for 8 minutes, then at 50 feet for 8 minutes, and you surface in a pressure group of O.

• Now, is that the same as 75 feet for 16 minutes?

Certainly not, since the latter would have you in the F group. And, if you planned a repetitive dive as if you were an F, when really you were an O, well you are just asking to get bent.

Of course, there are plenty of examples where the average calculation would have you in a higher group than the tables.

Ummmm. Which table are you using?

It clearly isn't USN-based. The letter groups look like you might have been using PADI Wheel or RDP, but not quite doing it right.

Anyway, here's your example run through a multilevel planning device, the PADI Wheel. (Actually, I don't have a wheel handy, so I'm using a flat table equivalent that I generated ---- I'm pretty sure that you will get the same result using a real wheel).

16 minutes at 75' is pressure group G. 8 minutes at 100' followed by 8 minutes at 50' is pressure group G. Hmmmmm. Same result.

Granted, if I run the 8 at 100' + 8 at 50' profile on the square profile planning device, the PADI RDP, all of the time has to be treated as being at 100' and results in an dive-ending pressure group of K, but that is just because the RDP is poorly tracks multilevel profiles. Clearly, DSAT/PADI consider pressure group G as the "more correct" answer, since that's what the wheel comes up with.

Run the two profiles through a decompression program and you will find the loadings for the 2 level profile are less than for the single level equivalent.

------------------

Since this is the profile you used as an example of how depth average is dangerous, please explain how you came up with your calculations and conclusion.

Charlie Allen

 

[rjack321]

The example is bogus from the get-go.

The neither the Padi wheel nor letter groups are used. They are not germane to the "averaging" approach.

The answers to your questions are in here:
http://74.52.40.164/showthread.php?t=139239

 

[Thalassamania]

The luck is in the fact that the math worked out. It's like Uncle Pug's, "add the depth the time" formulas. The point is that you can't work even a small fraction of the possible combinations and permutations, this is true for any model, but more so for a mathematical structure that is not really integral to the model (itself being an abstraction of at least one level from reality), but that has been fit to it ex post facto and is thus an abstraction carried out yet one additional level away from reality.

This is not to say that the system will not work within a narrow suite of endpoints. It may well do so, but you have no idea what the actual endpoint are or when the model will fail to work, and that is the problem. Confidence grows as is successfully used and thus the universe to which it is applied is expanded and confidence grows, etc. But the very real possibility of a catastrophic (in mathematical terms, at least) failure sits there waiting in unexpected and unpredictable loci, one which may be entered as a result of nothing more than not padding the model enough with a subjective safety margin.

 

[Thalassamania]

There are some real potential advantages to some of the averaging techniques, not the least of which is the promise of being able to "ride the curve" and thus decompress more efficiently. I'd love to see systems (tables, computers, whatever) developed from first principles to make them work. It's just that we're not there yet and I'm kind of attached to my bowel, bladder and sexual function.

 

[jonnythan]

It's lucky for us divers that the math works out.

The part that doesn't involve luck is making use of it!

The leap is proving to yourself that the math does indeed work out by comparing to tables and computers.