Modified ratio deco

[boulderjohn]

No one is using ratio deco as a primary means of creating an ascent profile.

I don't know what UTD is doing now, but that was very much the case when I was with them. We were required to do all dive planning with the UTD ratio deco program, and nothing else was allowed. This was true even though we were diving at altitude--we were told altitude was not a factor in decompression, so we didn't have to take that into consideration.

I took a Ratio Deco class from Andrew Georgitsis, and every time we did a profile problem, a UTD instructor in the room would do the same problem on a computer with an established program, including Buhlmann, VPM, and RGBM. Every time he did that, the UTD Ratio Deco profile would be different, and Andrew would celebrate that difference, because it proved the superiority of the UTD Ratio Deco. Since the UTD profile was assumed to be perfect, the difference showed the inferiority of the other algorithms.

 

[boulderjohn]

The 2017 Spisni et al study was rather pointless and suffered from severe methodological flaws. They picked one arbitrary, undocumented ratio deco approach and then compared it to Bühlmann ZHL-16C with 30/85 gradient factors. Which is of course completely backwards: if what you want is ZHL-16C with 30/85 then you can pick ratios to approximate that model. Curiously they didn't even explain where they obtained the particular "ratio decompression strategy (RDS)" profile used in the study. Was it from UTD training materials or did they just pull numbers out of their butts? There are no details in the methods section and nothing relevant in the references so who knows? Overall this is just very sloppy irreproducible science, to the extend that I'm disappointed it passed peer review.

With UTD, we had one and only one way of calculating dive profiles, so that is the way they did it. There was no choice to be made.

Andrew Georgitsis set up the rules for the study. When he announced the study, people were excited, but that excitement waned among some of the more scientifically literate as the details emerged. My friend Lynne Flaherty, an ER doctor who was then a big supporter of UTD (that changed), was disappointed because she felt the rules for the study had been arguably set up to make ratio deco look good. She felt it should have been set up to make a fairer comparison.

Andrew made a HUGE tactical mistake by publishing a video ahead of the study. In it, he said that the purpose of the study was to prove the superiority of ratio deco. Of course, anyone with any knowledge of scientific method knows having such a purpose leads to questions about your methodology--you are supposed to have an objective purpose. He said that ratio deco would win a comparison easily if they used pure Buhlmann for the comparison, so they selected the 30/85 GFs to give Buhlmann a fighting chance. I would argue that if he had used a higher GF low, the results would have been even worse for ratio deco.

 

[grantmac]

Let me get this straight:
You'll trust a computer to run your car, fly the plane when you travel and control your ECCR.
But it's going to give random erroneous data which is less accurate than your task loaded brain will approximate using far less precise inputs?

 

[Nick_Radov]

I don't know what UTD is doing now, but that was very much the case when I was with them. We were required to do all dive planning with the UTD ratio deco program, and nothing else was allowed. This was true even though we were diving at altitude--we were told altitude was not a factor in decompression, so we didn't have to take that into consideration.

Yikes. OK well I had no idea that was going on. That isn't anything I would endorse.

I took a Ratio Deco class from Andrew Georgitsis, and every time we did a profile problem, a UTD instructor in the room would do the same problem with an established program, including Buhlmann, CPM, and RGBM. Every time he did that, the UTD Ratio Deco profile would be different, and Andrew would celebrate that difference, because it proved the superiority of the UTD Ratio Deco. Since the UTD profile was assumed to be perfect, the difference showed the inferiority of the other profiles.

It's funny, when I took some GUE courses years ago from him and other instructors we used DecoPlanner (Bühlmann algorithm) first, then learned how you could apply a different ratio deco approach to get similar results as the program. I guess UTD lost the plot somewhere along the way?

 

[Nick_Radov]

Let me get this straight:
You'll trust a computer to run your car, fly the plane when you travel and control your ECCR.
But it's going to give random erroneous data which is less accurate than your task loaded brain will approximate using far less precise inputs?

Trust but verify. The most likely dive computer failure mode is user error, like accidentally configuring the wrong gasses before the dive or selecting the wrong gas during the dive. If your ratio deco estimate is way off from what the computer is telling you then that's a sign you need to stop and check everything.

A dive computer can certainly give you a more precise output but is it really more accurate? The error bars on this stuff are so wide that it's hard to be sure. A lot of people confuse precision for accuracy but those are fundamentally different concepts. I am unconcerned with precision in a field which is still as much an art as a science.

 

[boulderjohn]

It's funny, when I took some GUE courses years ago from him and other instructors we used DecoPlanner (Bühlmann algorithm) first, then learned how you could apply a different ratio deco approach to get similar results as the program. I guess UTD lost the plot somewhere along the way?

At the time I did that class, I was also in contact with Jarrod Jablonski of GUE. He told me exactly what you described, implying that UTD had, indeed, gone down a different path.

 

[grantmac]

Trust but verify. The most likely dive computer failure mode is user error, like accidentally configuring the wrong gasses before the dive or selecting the wrong gas during the dive. If your ratio deco estimate is way off from what the computer is telling you then that's a sign you need to stop and check everything.

A dive computer can certainly give you a more precise output but is it really more accurate? The error bars on this stuff are so wide that it's hard to be sure. A lot of people confuse precision for accuracy but those are fundamentally different concepts. I am unconcerned with precision in a field which is still as much an art as a science.

I'd argue that compared to a task loaded brain it's both more accurate and more precise. It's also presenting information in real time rather than requiring input.
Technical dive training should be focusing on how to leverage this incredibly powerful tool we bring with us rather than using it as a colourful bottom timer.

 

[mr_v]

I briefly looked into ratio deco and decided against it pretty much for the reasons summarized in post #8.

My SOP is having several dive plans (e.g., normal, deeper + longer, bailout), diving the plan and relying on several computers which in some cases come from different manufacturers. Usually I surface ~2-3 mins earlier than planned due to slight PPO2 variation on a higher side compared to the plan.

If you think you can do math better than computers at > 7 ATA, then go for it. My math skills don't improve with depth.

 

[Nick_Radov]

I'd argue that compared to a task loaded brain it's both more accurate and more precise. It's also presenting information in real time rather than requiring input.
Technical dive training should be focusing on how to leverage this incredibly powerful tool we bring with us rather than using it as a colourful bottom timer.

If you're so task loaded that you can't do a simple ratio deco calculation (using whatever method you prefer) then you're diving beyond your ability. Time to take a step back.
Technical dive training should be focusing on how to develop thinking divers. Dive computers are excellent tools (I love mine) but you can't have blind trust in them and need to be able to recognize when the output seems wrong, due to either device failure or (more likely) diver error.

 

[lostsheep]

If you're so task loaded that you can't do a simple ratio deco calculation (using whatever method you prefer) then you're diving beyond your ability. Time to take a step back.
Technical dive training should be focusing on how to develop thinking divers. Dive computers are excellent tools (I love mine) but you can't have blind trust in them and need to be able to recognize when the output seems wrong, due to either device failure or (more likely) diver error.

If you can't be bothered to add a computer check, including which gasses are configured for the dive, you are diving beyond your ability.

A thinking diver should be thinking about everything, including their computer. Blind trust in ratios is as asinine as blind trust in your computers.

Interestingly enough, I saw a recent conversation involving local GUE divers, primarily trained and using ratio deco, where they didn't understand what their computer was telling them and why there was a discrepancy. There are plenty of letters in GUEEDGE where a computer check would fit. To this groups credit, they did catch it.

RTFM