Gradient Factor Presets by Manufacturer/Computer

[dmaziuk]

It's Workmann's fault really: the one fixed 2:1 M-value was not good enough for him, he had to replace it with a function of depth, and give each tissue its own one too. Really screwed up us recreating divers: now when checking for no-stop time we have to check the M-value at current depth, at the surface, and everywhere in between.

 

[CandiveOz]

I evolved my Gradient Factor grouping. It was suggested that I add a GF-hi Moderate grouping between 80-89. Also I highlighted those GF's that may be pushing risk limits too high or in some cases not applicable such as a GF-hi of 30 or below.

 

[CandiveOz]

With the new grouping, this is what the Presets look like. Note that the GF-lo's that are considered too deep are included in each colour group. Perhaps manufacturers should reconsider the red settings to be included as Presets and reserve these settings just for custom settings.

 

[CandiveOz]

You could consider using the delta or spread instead of absolute GF Lo value. The rationale would be that the base conservatism (your "in terms of risk") level is not the GF Hi but rather the GF slope at GF Hi = GF Lo.

I think you should not attach words like “aggressive” or “conservative” to dive plans, and especially to settings. Those words carry meaning to casual readers which isn’t actually there. You can get bent on a conservative plan, it is a matter of probability. Which probability counts as “conservative”?

I think you both have valid points. In essence, the use of gradient factors have highlighted two risks. First we know it is risky to dive close to the M Line. GF-hi mitigates this risk. We also now know (after the NEDU study and others) that having a wide GF spread, which creates deep stops, is also risky.

GF-hi is easy for divers to understand and set. But GF-lo is difficult because we don’t really know how much is too much. Maybe the diving industry and diving community have to rethink gradient factors. Perhaps the dive computer settings should have the GF-lo setting limited to a maximum spread.

I recut the data to rank the Presets by GF-hi and then by Spread. I highlighted GF Spread of 50-100 as being too wide. Unlike my previous sort, this one creates one big red block of questionable gradient pairs. Again, it is probably this block of pairs which should be left out of Presets and reserved for custom settings.

 

[tursiops]

I think you both have valid points. In essence, the use of gradient factors have highlighted two risks. First we know it is risky to dive close to the M Line. GF-hi mitigates this risk. We also now know (after the NEDU study and others) that having a wide GF spread, which creates deep stops, is also risky.

GF-hi is easy for divers to understand and set. But GF-lo is difficult because we don’t really know how much is too much. Maybe the diving industry and diving community have to rethink gradient factors. Perhaps the dive computer settings should have the GF-lo setting limited to a maximum spread.

I recut the data to rank the Presets by GF-hi and then by Spread. I highlighted GF Spread of 50-100 as being too wide. Unlike my previous sort, this one creates one big red block of questionable gradient pairs. Again, it is probably this block of pairs which should be left out of Presets and reserved for custom settings.

View attachment 513665

You seem hung up on GF "spread". That is just an artifact of a low GFlo and a high GFhi. The issue is that a low GFlo is arguably not so good, no matter what the GFhi is. The spread is not the point.

 

[dmaziuk]

We also now know (after the NEDU study and others) that having a wide GF spread, which creates deep stops, is also risky.

The point of "bad" deep stops is that they were reducing the shallow stop time. GF equivalent would be a GF Hi over 100%. When you don't do that, all you get from low GH Lo is extra deco time in the shallows. If you don't pad the shallow stops to compensate for deep ones, it's bad for you. Sure, and in the other exciting news the water was recently found to be wet.

The problem with the "spread" is that Eric designed his GF system for decompression dives with relatively deep first stops. As you push it away from its original design parameters toward no-top dives, you're getting into square peg in a round hole territory. It's rather obvious if you read the original paper: when your first stop is your last stop is the surface, the proportion for calculating the current gradient factor reduces to division by zero. There is no smooth transition between no-stop and yes-stop dive in ZHL+"spread" GFs, and the "spread" exacerbates it.

The main problem with all this is vendors selling computers to people not qualified to operate them. Including, but not limited to, dive computers.

 

[CandiveOz]

You seem hung up on GF "spread". That is just an artifact of a low GFlo and a high GFhi. The issue is that a low GFlo is arguably not so good, no matter what the GFhi is. The spread is not the point.

Hi @tursiops .Thanks for your feedback. We agree that a low GF-lo may not be a good idea. Where we disagree is the GF Spread. When GF-lo is equal to GF-hi and less than 100 (eg 85/85), all you are doing is a parallel shift of the M line to add conservatism to the model. Buhlmann even did this when he made ZH-L16 more conservative by moving from model A to C. He changed the “Mo” intercept in the medium compartments but kept the slope of the line the same in each model. The table below is taken from Baker’s paper Understanding M-Values.

When GF-lo does not equal GF-hi (eg 40/85) creating a spread, essentially you are changing the slope of the M line which deviates from the intent of the original model. You basically have created a new (untested) model. The following graph from Baker’s paper; Clearing Up The Confusion About “Deep Stops” illustrates this change.

With regards to GF-lo versus GF Spread for example 20/45 vs 45/45. Both pairs are considered very conservative and both GF-lo’s are low but, in my opinion, 45/45 is “safer” because although the M line has been shifted down, the slope is the same as the original model. When it comes to decompression theory, I don’t pretend to know more than Buhlmann and if he didn’t think it was appropriate to tilt the M line, should I?

So in response to you statement. The GF spread is the point because it alters the slope of the M line.

BTW I read in anther post that someone stole your dive computer. I sincerely hope you get it back. Good luck!

 

[CandiveOz]

The point of "bad" deep stops is that they were reducing the shallow stop time. GF equivalent would be a GF Hi over 100%. When you don't do that, all you get from low GH Lo is extra deco time in the shallows. If you don't pad the shallow stops to compensate for deep ones, it's bad for you.

Hi @dmaziuk . In general I agree with your comments. What you are describing is what I call a “GF twist” which generates deep stops while reducing shallow stop time illustrated below.

What I would also add is an observation from Bakers paper; Clearing Up The Confusion About “Deep Stops”. Baker uses 3 different profiles to illustrate deep stops; a conventional deco profile, Pyle stops and finally gradient factor deep stops. Note that the conventional deco profile has the shallowest first stop and shortest run time. The GF deco profile has a shallower first stop compared to Pyle but the longest run time. In fact its 20 minutes longer than the conventional deco profile. So if a diver is happy to spend an extra 20 minutes in the water, he probably would be better off doing it between 6 – 3 meters using a conventional deco profile rather than a GF deco profile.

 

[scubadada]

@CandiveOz is correct, the difference between Buhlmann ZH-L16A, B, and C is simply a shift of some of the compartments to the right, without a change in the slope of the line. A equal shift in the choice of the GF high and GF low would simply shift the curve farther to the right. A lower GF low shifts the bottom of the curve even farther to the right, compared to surfacing

 

[dmaziuk]

Hi @dmaziuk . In general I agree with your comments. What you are describing is what I call a “GF twist” which generates deep stops while reducing shallow stop time illustrated below.

But that's essentially the profiles NEDU study studied, as I understand it. It doesn't prove that "wide spread", or low GF Low within the "untwisted" ZHL+GF model, is bad per se. It strongly indicates that doing so is unnecessary and the resulting increase in overall decompression time is counterproductive. As long as you're adding more conservatism to base profile, so I don't find the "untested slope" argument convincing. Suboptimal, yes, but since the goal was to add more conservatism based on diver's personal considerations... well, that's what it does.