Have I understood the basics of decompression theory, GF99 and SurfGF?

[Dan]

I ask CharGPT:
What would you recommend for medium conservative recreational dives?

Here’s its answer:
“Short recommendation (medium conservative, recreational)

• GF Low = 55
• GF High = 85
  Use ZHL-16C (what Shearwater uses) with those GFs for a sensible “medium” conservatism for recreational no-deco profiles and occasional short deco events

Why this pair

• 55 (GF Low) follows Fraedrich-style caution against very low GF Low values (<55). It avoids forcing an overly-deep first stop that can keep you at high ambient pressure for longer (which can load slow tissues).
• 85 (GF High) keeps final ascent conservative enough to give a safety buffer but doesn’t massively extend total deco time the way a much lower GF High would.
• Together 55/85 is a pragmatic compromise: noticeably more conservative than many factory “medium” presets (45/95), but still usable for recreational diving without huge penalties.

If you want to tweak

• Slightly more conservative (age, cold, heavy exertion, strong current): 60/80
• Slightly less conservative (young, fit, warm water, short dives): 50/90
• Very conservative / tech-style planned deco: consider 65/75 or use full planned stop tables — but that’s getting into technical territory.

Practical reminders (always)

• For true no-deco dives GF Low is rarely triggered — but the GF pair matters if you ever stray into deco.
• Always do a 3–5 minute safety stop at ~5 m (15 ft), longer if you feel rough.
• Adjust conservatism for factors that increase DCS risk: cold, heavy exercise, dehydration, long surface intervals, multiple repetitive dives, age, heavy workload.

If you tell me your typical max depth and usual bottom time (or your Shearwater model), I can give a concrete GF pair tuned specifically to your profile.”

I use Teric

It’s response:
“For a Shearwater Teric doing typical recreational, no-decompression dives and wanting a medium-conservative profile, here’s a solid setup you can enter in the Custom GF field:

Suggested Gradient Factors

• GF Low: 55
• GF High: 85

Why 55/85 works well on a Teric

• 55 (Low): Avoids the overly-deep first stop that research (e.g., Fraedrich) warns about, yet still gives a modest buffer if you accidentally incur a short deco obligation.
• 85 (High): Adds a bit of extra margin on the final ascent compared with the default 95, without stretching total run time much.

How to set it on the Teric

1. Menu → Dive Setup → Decompression
2. Make sure Algorithm = Bühlmann ZHL-16C.
3. Set GF Low = 55 and GF High = 85.
4. Leave “Conservatism Preset” on Custom so these numbers stay active.

Extra tips for recreational use

• Do a minimum 3–5 min safety stop at 5 m / 15 ft every dive.
• Bump GF High down to 80 or GF Low up to 60 if you’re cold, tired, or doing multiple deep dives in a day.
• Remember that on true no-stop dives, the Low value is rarely triggered; the main benefit is if you accidentally go into light deco.

That 55/85 pair is a well-regarded “medium-conservative” choice for a Teric when you want a bit more margin than the factory Medium (45/95) but without adding unnecessary time.”

 

[tursiops]

Fraedrich's analyses show that GF Low should not be set to < 55 so SW presets are really wrong on that side. OTOH if their definition of "recreational" is "no-stop", then GF Low doesn't matter anyway unless set to something ridiculous like 10 IIRC.

In the same article he says not to exceed GFHi=80. So what is the rationale in the posts above for 55/85? Seems like cherry picking various results to find the cherry you want. 55/80 would be Fraedrich's and would ignore David Doolette. Hmmm, who would I rather believe?

 

[tursiops]

I ask CharGPT:
What would you recommend for medium conservative recreational dives?

Here’s its answer:
“Short recommendation (medium conservative, recreational)

• GF Low = 55
• GF High = 85
  Use ZHL-16C (what Shearwater uses) with those GFs for a sensible “medium” conservatism for recreational no-deco profiles and occasional short deco events

Why this pair

• 55 (GF Low) follows Fraedrich-style caution against very low GF Low values (<55). It avoids forcing an overly-deep first stop that can keep you at high ambient pressure for longer (which can load slow tissues).
• 85 (GF High) keeps final ascent conservative enough to give a safety buffer but doesn’t massively extend total deco time the way a much lower GF High would.
• Together 55/85 is a pragmatic compromise: noticeably more conservative than many factory “medium” presets (45/95), but still usable for recreational diving without huge penalties.

If you want to tweak

• Slightly more conservative (age, cold, heavy exertion, strong current): 60/80
• Slightly less conservative (young, fit, warm water, short dives): 50/90
• Very conservative / tech-style planned deco: consider 65/75 or use full planned stop tables — but that’s getting into technical territory.

Practical reminders (always)

• For true no-deco dives GF Low is rarely triggered — but the GF pair matters if you ever stray into deco.
• Always do a 3–5 minute safety stop at ~5 m (15 ft), longer if you feel rough.
• Adjust conservatism for factors that increase DCS risk: cold, heavy exercise, dehydration, long surface intervals, multiple repetitive dives, age, heavy workload.

If you tell me your typical max depth and usual bottom time (or your Shearwater model), I can give a concrete GF pair tuned specifically to your profile.”

I use Teric

It’s response:
“For a Shearwater Teric doing typical recreational, no-decompression dives and wanting a medium-conservative profile, here’s a solid setup you can enter in the Custom GF field:

Suggested Gradient Factors

• GF Low: 55
• GF High: 85

Why 55/85 works well on a Teric

• 55 (Low): Avoids the overly-deep first stop that research (e.g., Fraedrich) warns about, yet still gives a modest buffer if you accidentally incur a short deco obligation.
• 85 (High): Adds a bit of extra margin on the final ascent compared with the default 95, without stretching total run time much.

How to set it on the Teric

1. Menu → Dive Setup → Decompression
2. Make sure Algorithm = Bühlmann ZHL-16C.
3. Set GF Low = 55 and GF High = 85.
4. Leave “Conservatism Preset” on Custom so these numbers stay active.

Extra tips for recreational use

• Do a minimum 3–5 min safety stop at 5 m / 15 ft every dive.
• Bump GF High down to 80 or GF Low up to 60 if you’re cold, tired, or doing multiple deep dives in a day.
• Remember that on true no-stop dives, the Low value is rarely triggered; the main benefit is if you accidentally go into light deco.

That 55/85 pair is a well-regarded “medium-conservative” choice for a Teric when you want a bit more margin than the factory Medium (45/95) but without adding unnecessary time.”

People should STOP quoting ChatGPT (etc) for answers to health-and-life-endangering questions. A little human intelligence and a lot of human skepticism and judgement are needed, not spewing out possible nonsense.

 

[Dan]

People should STOP quoting ChatGPT (etc) for answers to health-and-life-endangering questions. A little human intelligence and a lot of human skepticism and judgement are needed, not spewing out possible nonsense.

Which part of the statement you are disagreeing with and spewing out possible nonsense?

I’ve been diving with GF 40/85 for ~ 1000 dives w/o any problem, no headache, nada, never miss a dive on dive trips (got 7 Iron Diver medals from all 7 dive trips on Aggressors for not missing a dive in their liveaboard trips), etc. A few times I hit a light DECO, which cleared up as I ascended to safety stop.

That GF Conservative 40/85 is Med. Shearwater GF Low Conservative setting is 40/95.

I made sure my SurfGF < 70 before sufrfacing.

 

[tursiops]

Which part of the statement you are disagreeing with and spewing out possible nonsense?

I’ve been diving with GF 40/85 for ~ 1000 dives w/o any problem, no headache, nada, never miss a dive on dive trips (got 7 Iron Diver Medals from Aggressors), etc. A few times I hit a light DECO, which cleared up as I ascended to safety stop.

You may dive however you'd like. That is irrelevant to others. You are one anecdote, not even a data point without mor info.

My problem is spewing out verbatim ChapGPT output; at least you said it was ChatGPT. Why should I go through that wall of text and discredit it or OK it, line by line. YOU should do that before you post it. If you can't, your are guilty of pssible spreading of misinformation.
It is mixing recommendations from different sources. And at the end, which you did NOT quote, it probably says something like, "this may be wrong." If it doesn't, it should.

 

[Dan]

. And at the end, which you did NOT quote, it probably says something like, "this may be wrong." If it doesn't, it should.

I asked the question and posted its answer completely (not missing a word). You can verify it yourself by asking ChatGPT the same question.

You are still dodging. Which parts of the statement are false or nonsense? Educate me.

 

[tursiops]

Educate me.

Here is what Google AI says about ChatGPT. Do your own fact-checking, or don't post.

No, ChatGPT is not always correct; it is a language model that generates responses based on patterns in its training data, not a reliable source of facts, and can produce inaccurate or misleading information, a phenomenon known as hallucination. Therefore, you should always critically evaluate its responses and fact-check any information, especially for critical or technical topics like legal, medical, or financial advice. [1, 2, 3, 4, 5, 6]


Why ChatGPT makes mistakes

• Based on patterns, not facts: ChatGPT doesn't "know" things in the human sense but rather predicts the most statistically likely next word based on its training data. [4, 7, 8, 9]
• Outdated information: Its knowledge is based on its training data, which can become outdated over time. [1, 6]
• Lack of independent verification: ChatGPT cannot independently verify the accuracy of the information it generates. [3]
• Limited context: The accuracy of its responses depends on the context and complexity of the task. [10]

How to improve response accuracy and reliability

• Fact-check everything: Always cross-reference information with reputable sources. [1, 2, 11]
• Be specific with prompts: Provide more context in your prompts to help guide the model. [6]
• Ask for sources: Request sources or explanations for the reasoning behind a response to encourage more fact-checking. [6, 11]
• Use the latest models: The latest models, like GPT-4o, have improved internet access and can cross-check responses, though access may be limited for free users. [6, 12]
• Understand its limitations: Recognize that ChatGPT may not always be right and should not be trusted for critical advice without verification. [1, 3]

AI responses may include mistakes.
[1] https://fresnostate.teamdynamix.com/TDClient/1922/Portal/KB/PrintArticle?ID=164927
[2] https://vocal.media/01/why-chat-gpt-isn-t-always-right
[3] https://www.chatlab.com/blog/how-accurate-and-reliable-is-chatgpt-for-users/
[4] https://help.openai.com/en/articles/8313428-does-chatgpt-tell-the-truth
[5] https://www.scribbr.com/frequently-asked-questions/chatgpt-credible-source/
[6] How Accurate is ChatGPT?
[7] How reliable is ChatGPT when it comes to answering complex technical questions accurately?
[8] https://pmc.ncbi.nlm.nih.gov/articles/PMC11472406/
[9] https://dataconomy.com/2025/04/28/6-techniques-to-fix-chatgpt-annoying-habits/
[10] Does Chat GPT always give the right answer?
[11] [12] How accurate is ChatGPT? Should you trust its responses?

 

[Dan]

You are still dodging.

 

[tursiops]

You are still dodging.

LOL.
And you are still refusing to take responsibility for possibly posting nonsense! How do you know, if all you do is repost something from ChapGPT?

 

[boulderjohn]

Cool, I dive 70/85. Does that make me an expert?

Well, I think you have things reversed. Copying experts (which I do myself) does not make you an expert. I was talking about people who are experts on their own otherwise demonstrated expertise, like Dr. Simon Mitchell.