Deep Stops Increases DCS

[UWSojourner]

I don't agree with that.

Simon Mitchell is heavily vested in these arguments, in a different kind of currency - community stature and lasting scientific recognition.

He is a working Dr., and hyperbaric specilaist who has chosen to publish papers and make public opinions. He travels the world giving talks on these and other ideas, based on the strength of his name and reputation. Like others in science, he embarked on a direction and a set of ideas early in his career, and is pursuing that now.

The currency of Simon Mitchell is community stature and lasting scientific recognition. Those are the valuable commodities to academics and take life time to achieve. A career and future prospects hinges on these matters too.

Simon Mitchell is not a neutral or unbiased in any way. He is more heavily invested in the direction this takes, than you or me.

Well, good morning.

I agree with you that everyone likely has a currency. Your nice description of Dr. Mitchell's - community stature, lasting scientific contribution, public education - can all be summed up as integrity. The currency of integrity buys trust. I have no doubt that what I hear from Dr. Mitchell is what he believes to be true. And that if the tide of scientific evidence on deep stops comes back in, I expect he'll ride it in and let us all know of any new evidence.

Financial interest is fickle. It will follow the evidence if the cost isn't too high, but may not if it is. That's why it's good practice to require disclosure of financial interest as a potential conflict of interest. Nobody ever had to disclose integrity as a conflict.

 

[David Doolette]

Added later:



Really? David seems to think its OK now.

But how interesting that you suddenly seem to understand these important differences... now - when it suits you. Your youtube "presentation", is devoid of such distinctions and uses that to incorrectly make conclusions that are not there.

.

Ross,

You changed the graph, I believe after Simon made the suggestion. In the version I am looking at now, the two double headed arrows should fall between the solid red tissue gas pressure line and the respective dotted lines (orange ambient pressure line for supersaturation, and green inspired inert pressure line for offgas gradient)

David Doolette

 

[rossh]

Ross,

You changed the graph, I believe after Simon made the suggestion. In the version I am looking at now, the two double headed arrows should fall between the solid red tissue gas pressure line and the respective dotted lines (orange ambient pressure line for supersaturation, and green inspired inert pressure line for offgas gradient)

David Doolette

I probably made ten changes since I got your input. Thanks the help. I'm going to keep going on this and see if we can resolve some of these issue with more pictures - worth a thousand words I believe.

 

[UWSojourner]

I wonder what you have to disclose?

I've already "disclosed" that I have no financial, or familial, or any other interest in any dive company, product, or person.

You never say your name.

No need to. You've posted it over and over and over. You can call me Kevin if you don't like UWSojourner.

You never say your interest here.

Diving is a hobby.

And you have been trying to cut my throat ever since.

Given your bizarre, misinformed, paranoid, vitriolic posts "ever since", some against arguably the most well respected decompression scientists on the planet, I think by any objective standard you've been cutting your own throat.

You turn up to these thing once a year, with perfect timing, for two months, and make trouble.

I go online from time to time as time permits -- usually in lulls from other activities. If its "perfect timing" then that must be Providence.

I wonder what your day job is?

I switched careers back in 2010 to become a full time harasser of Ross Hemingway . Jeeze, not everything is about you and this "debate". My career is unrelated to the dive industry and has nothing to do with these forums.

But what don't we know about you, that you should be revealed, for full disclosure?

Nothing comes to mind. My life is pretty ordinary -- family, some hobbies, responsibility, etc.

 

[shoredivr]

Irrelevant, deleted

 

[The Chairman]

Guys, guys, guys... let's cut the personal vendetta crap and get back to the science and how it affects Joe Diver.

 

[UWSojourner]

I've been waiting to make a correction partly because the discussion switched to published papers. And I tried to let Ross make it himself, but what can you do. I just want to have a post that I can reference if he tries the make the same claim again (so I don't have to deal with it).

Ross has attacked the heat map as flawed. In particular he asserts quite strongly that it is not based on supersaturation whereas his charts (jagged line charts) are based on "exact supersaturation". Here's his argument.

Ross's Argument

Ross's argument comes with graphics. His graphics make reference to a heat map posted here. See Ross's chart below.

The argument Ross makes goes as follows:

"The colors in the heat map claim to represent relative levels of supersaturation. Light Green should represent 50% of Red per the legend. My software shows that when the diver surfaces from VPM-B+3 on this dive they have 7 kPa of supersatuation in compartment #16 (C16). The most heavily supersaturated compartment, C9, has 55 kPa of supersaturation. Therefore, C16 should be the color that represents 12% of the highest supersaturation. This can be derived by dividing C16 by C9, or 7 / 55 = 12%. A 12% compartment should be very dark green-black (see legend). But it is NOT! It is colored some shade of light-green which implies a 50% relative supersaturation to the highest compartment. Error! Therefore, these colors are not based on supersaturations. Error!"

The same basic argument with different colors/compartments is repeated below the first.

Problems With Ross's Argument

Ross's argument depends on 2 pieces of information and 2 implicit assumptions.

1. The supersaturation of C16 at the surface is 7 kPa.
2. The supersaturation of C9 at the surface is 55 kPa
3. It is appropriate within the goals of the heat map to divide C16 by C9 to set the relative color of the heat map.
4. The color scale is ALWAYS a fraction of the most heavily supersaturated compartment in the profiles shown (more on this later).

Unfortunately, #1, #2, #3, and #4 are all not true. Ross misunderstands, either conceptually or intentionally, the purpose of the heat map when he divides compartments within the same profile. The purpose of the heat map is to compare supersaturation patterns between profiles. When developing the heat map I wasn't interested in how superaturation in a single profile's compartments compared to each other. I was interested in how the compartments between 2 or more profiles compared across profiles. This error was spotted and pointed out here by Victor.

The errors in Ross's calculation of supersaturation are documented below.

Ross's "Supersaturation"

I can repeat Ross's numbers and charts. This only makes sense. We're both using the same tissue model. So what's Ross missing in supersaturation?

Ross makes the claim that C16's supersaturation upon surfacing is 7 kPa and that C9's is 55 kPa. I can get those numbers as well by the following calculation for the dive discussed here:

As you can see, Ross used 7 and 55 and my calculation is very close to his.

It's easy to see what Ross overlooks when looking at the chart below.

The charts Ross shows in his software as "supersaturation" are really compartment inert pressure (N2 and He) less ambient pressure. It fails to account for a small, but important, part of supersaturation (and therefore bubble formation potential) from the other gases O2, CO2 and H20. Normally, even in his beloved VPM, this is modelled as a constant added to the inert compartment pressures N2 and He. But here, as in the RBW debate a few years ago, Ross seems unaware of what something as basic as supersaturation really means. Certainly his assertion that his charts represent "exact supersaturation" might be seen as exaggerated.

Summary

Ross's error probably doesn't matter much in his software since the charts are more of a "ooh, cool" thing. I'm only pointing it out because HE has made particular assertions of error in the heat maps based on his misunderstanding of supersaturation. In very heaviliy supersaturated compartments the small missing pressure isn't as big a deal. But he chose a compartment to make his argument where it does make a difference. In that context, the fact that he failed to include 13.6 kPa (I believe that is VPM's value for other gases) of pressure is an error of about 200%. Of course, making the 3 other errors didn't help him either.

The heat maps are accurate in their visual portrayal of the relative handling of supersaturation across profiles. Perhaps Ross would ask the Mods on SB to allow him to remove or correct his posts here:

Inaccurate claim 1 Inaccurate claim 2 Inaccurate claim 3 Inaccurate claim 4
Inaccurate claim 5 Inaccurate claim 6 Inaccurate claim 7

I may have missed others.

Sorry for the long post, but one more point. The heat map, as disclosed previously, will not always scale based on the highest supersaturation of any profile's compartment. I did that initially but it inappropriately implied (visually) that more supersaturation was present than actually was when compartments were only lightly supersaturated. For example, assume 2 profiles. C16 in the first profile has 0.02 kPa supersaturation and in the second has 0.01 kPa supersaturation. That's VERY LOW. Clearly both compartments are JUST BARELY supersatured. Without some adjustment the first profile would scale "red" (highest supersaturation) and the second light-green (50% of the highest). This would be consistent coloring, but undesireable in my view (and generally would make VPM look worse). To adjust for this, when the highest compartment falls below the supersaturation implied by some multiple (e.g. 90%, etc.) of the N2 M-value at the surface, I scale the colors based on this higher value. That does not inappropriately disadvantage any profile since the colors are all relative anyway (and generally makes VPM look better than without the adjustment), but it scales back the coloring as compartments converge toward an unsupersaturated state. Remember, again, the colorings are relative. I was just trying to see how the "complex on gassing and off gassing patterns" actually compared between profiles. I had no idea it would highlight anything, but it did.

 

[rossh]

{amended}

I'll look into this more shortly......

... later on.

Confirmed yes. It was exactly as you said. The pOtherGas was left out. In fact I had removed it earlier, as I was concerned about making values a little too high.

Anyway all fixed now and it agrees with the number shown.

I have amended all the diagrams here that were affected.

This change only has a incremental effect, and no nothing major has changed.


However, the criticism i made of your heat maps, still stands, its now a 30% vs a 50-75% indication error. The rest is still true - the heat maps have no dimensions, and do not show plain supersaturation. This comparing of two unrelated objects, is not useful supersaturation information.


New programs are available now for download.

.

 

[igor p]

I've been waiting to make a correction partly because the discussion switched to published papers. And I tried to let Ross make it himself, but what can you do. I just want to have a post and I can reference if he tries the make the same claim again (so I don't have to deal with it).

Ross has attacked the heat map as flawed. In particular he asserts quite strongly that it is not based on supersaturation whereas his charts (jagged line charts) are based on "exact supersaturation". Here's his argument.

Ross's Argument

Ross's argument comes with graphics. His graphics make reference to a heat map posted here. See Ross's chart below.
View attachment 378853

The argument Ross makes goes as follows:

"The colors in the heat map claim to represent relative levels of supersaturation. Light Green should represent 50% of Red per the legend. My software shows that when the diver surfaces from VPM-B+3 on this dive they have 7 kPa of supersatuation in compartment #16 (C16). The most heavily supersaturated compartment, C9, has 55 kPa of supersaturation. Therefore, C16 should be the color that represents 12% of the highest supersaturation. This can be derived by dividing C16 by C9, or 7 / 55 = 12%. A 12% compartment should be very dark green-black (see legend). But it is NOT! It is colored some shade of light-green which implies a 50% relative supersaturation to the highest compartment. Error! Therefore, these colors are not based on supersaturations. Error!"

The same basic argument with different colors/compartments is repeated below the first.

Problems With Ross's Argument

Ross's argument depends on 2 pieces of information and 2 implicit assumptions.

1. The supersaturation of C16 at the surface is 7 kPa.
2. The supersaturation of C9 at the surface is 55 kPa
3. It is appropriate within the goals of the heat map to divide C16 by C9 to set the relative color of the heat map.
4. The color scale is ALWAYS a fraction of the most heavily supersaturated compartment in the profiles shown (more on this later).

Unfortunately, #1, #2, #3, and #4 are all not true. Ross misunderstands, either conceptually or intentionally, the purpose of the heat map when he divides compartments within the same profile. The purpose of the heat map is to compare supersaturation patterns between profiles. When developing the heat map I wasn't interested in how superaturation in a single profile's compartments compared to each other. I was interested in how the compartments between 2 or more profiles compared across profiles. This error was spotted and pointed out here by Victor.

The errors in Ross's calculation of supersaturation are documented below.

Ross's "Supersaturation"

I can repeat Ross's numbers and charts. This only makes sense. We're both using the same tissue model. So what's Ross missing in supersaturation?

Ross makes the claim that C16's supersaturation upon surfacing is 7 kPa and that C9's is 55 kPa. I can get those numbers as well by the following calculation for the dive discussed here:

View attachment 378854

As you can see, Ross used 7 and 55 and my calculation is very close to his.

It's easy to see what Ross overlooks when looking at the chart below.
View attachment 378855

The charts Ross shows in his software as "supersaturation" are really compartment inert pressure (N2 and He) less ambient pressure. It fails to account for a small, but important, part of supersaturation (and therefore bubble formation potential) from the other gases O2, CO2 and H20. Normally, even in his beloved VPM, this is modelled as a constant added to the inert compartment pressures N2 and He. But here, as in the RBW debate a few years ago, Ross seems unaware of what something as basic as supersaturation really means. Certainly his assertion that his charts represent "exact supersaturation" might be seen as exaggerated.

Summary

Ross's error probably doesn't matter much in his software since the charts are more of a "ooh, cool" thing. I'm only pointing it out because HE has made particular assertions of error in the heat maps based on his misunderstanding of supersaturation. In very heaviliy supersaturated compartments the small missing pressure isn't as big a deal. But he chose a compartment to make his argument where it does make a difference. In that context, the fact that he failed to include 13.6 kPa (I believe that is VPM's value for other gases) of pressure is an error of about 200%. Of course, making the 3 other errors didn't help him either.

The heat maps are accurate in their visual portrayal of the relative handling of supersaturation across profiles. Perhaps Ross would ask the Mods on SB to allow him to remove or correct his posts here:

Inaccurate claim 1 Inaccurate claim 2 Inaccurate claim 3 Inaccurate claim 4
Inaccurate claim 5 Inaccurate claim 6 Inaccurate claim 7

I may have missed others.

Sorry for the long post, but one more point. The heat map, as disclosed previously, will not always scale based on the highest supersaturation of any profile's compartment. I did that initially but it inappropriately implied (visually) that more supersaturation was present than actually was when compartments were only lightly supersaturated. For example, assume 2 profiles. C16 in the first profile has 0.02 kPa supersaturation and in the second has 0.01 kPa supersaturation. That's VERY LOW. Clearly both compartments are JUST BARELY supersatured. Without some adjustment the first profile would scale "red" (highest supersaturation) and the second light-green (50% of the highest). This would be consistent coloring, but undesireable in my view (and generally would make VPM look worse). To adjust for this, when the highest compartment falls below the supersaturation implied by some multiple (e.g. 90%, etc.) of the N2 M-value at the surface, I scale the colors based on this higher value. That does not inappropriately disadvantage any profile since the colors are all relative anyway (and generally makes VPM look better than without the adjustment), but it scales back the coloring as compartments converge toward an unsupersaturated state. Remember, again, the colorings are relative. I was just trying to see how the "complex on gassing and off gassing patterns" actually compared between profiles. I had no idea it would highlight anything, but it did.

This aproach is very inapropriate as show very small differences as being enormous and big differences as being small or equal to smaller ones and vice versa.

 

[UWSojourner]

This aproach is very inapropriate as show very small differences as being enormous and big differences as being small or equal to smaller ones and vice versa.

I don't think so. The colors are relative so they don't show big and small. They show relative supersaturations in profiles of the same run time. Remember, it's a comparative exercise. How do profiles handle supersaturation compared to the other profiles. If all the colors between two profiles were the same, they would handle supersaturation the same. The differences in color show relative differences in how they handle supersaturation.

Again, absent the NEDU hard data points (A2 deep stop 5% DCS risk; A1 shallow stop 1.6% DCS risk) I would think you'd both be looking at the deep stops toward the left of the heat maps and all the "Red" in A1 and say "Wow, that's not good." It's only because we know what profile was actually better that the comparison is meaningful. And as a comparison its appropriate.