Living With Your Computer's Algorithm

Part 2 of 3 · Part 1: How Dive Computers Work · Part 3: Reading Your Computer in Real Time

You don't need to rebuild the algorithm

There are two families of decompression model. Dissolved-gas models (Bühlmann ZHL-16C, the de-facto standard) track inert gas in your tissues and let you tune conservatism with a gradient-factor dial. Bubble models (RGBM in Suunto, Mares and Scubapro; VPM-B in planning software) go further and model the gas bubbles themselves.

How any of that actually works is the job of the decompression series, not this article. If you want the machinery, read how decompression models work for compartments and M-values, gradient factors for the conservatism dial, and bubble trouble for what the bubble models add and whether it pays off. This article is about the questions you actually have with the computer on your wrist: why is it so cautious, how do I live with a bubble-model unit, and what happens when my team's computers disagree.

Why is my dive computer so conservative?

The single most common gripe, and there is no one answer. Usually it is some mix of these.

It is protecting a diver it cannot see. The model assumes a generic, well-behaved diver in ideal conditions. It has no idea you are cold, working hard, dehydrated, tired, older than the test subjects, or carrying a PFO, so it pads its numbers to cover the diver it cannot measure. That padding is exactly the margin the model can't personalise.

Its default settings are cautious on purpose. Manufacturers ship conservative defaults because a stingy computer is a safer liability position than a generous one. On a Bühlmann computer that shows up as the default gradient factors; on a Suunto or Mares it is the default "personal" and "altitude" settings. These are yours to change.

It remembers your last dives. On repetitive and multi-day diving it carries residual loading forward, and bubble-model computers carry bubble nuclei forward too, so your second and third dives get noticeably shorter than the first.

It punishes certain profiles. Reverse profiles (going deeper on the later dive), sawtooth up-and-down profiles, fast ascents, and short surface intervals all trigger penalties, especially on bubble-model computers.

Conservative is the safe direction to be wrong in. The whole system is biased toward stinginess, because the cost of a slightly short dive is nothing and the cost of a bend is a lot.

What you can do about it: set your conservatism deliberately rather than fighting the computer in the water. On a Bühlmann unit, choose sensible gradient factors (something like 40/85; resist the urge to crank GF Lo low in the belief it is safer, because it isn't). On an RGBM unit, set the personal and altitude levels with intent and avoid the profiles that draw penalties. What you should not do is "beat" it by surfacing early or ignoring its alarms.

Living with an RGBM or bubble-model computer

If you dive a Suunto, Mares, or Scubapro, you are on a bubble model, and it behaves differently from a Shearwater in a few ways worth knowing.

• There is no gradient-factor dial. The conservatism is baked into the bubble physics. You adjust it through "personal" and "altitude" settings, not GF, and the steps are coarser.
• It is harder on repetitive, reverse, and sawtooth dives. Because it carries bubble nuclei between dives, your later dives of the day shorten more than they would on Bühlmann. This is the usual reason an RGBM diver feels shortchanged next to a buddy on a Shearwater.
• To buy back time, change the dive, not the alarm. Longer surface intervals, shallow-to-deep ordering across the day, and avoiding yo-yo profiles all help. Drop the personal setting toward less conservative only if your fitness, conditions, and experience genuinely warrant it.
• Don't expect it to match other computers. It won't, and that is normal. If you truly need gradient-factor control and predictable, tunable schedules, that is a fair reason to move to a Bühlmann-plus-GF computer, but a sensibly configured RGBM unit dives just fine.

Diving a mixed-computer team

Put a few divers together and their computers will rarely agree, especially if some run Bühlmann and others run RGBM. The rule is simple and one-directional: everyone follows the most conservative computer. Hold every stop until the stingiest unit in the team has cleared it.

On a no-stop dive that means turning when the shortest NDL in the team reaches zero. On a decompression dive, a dissolved-gas computer and a bubble-model computer can diverge by several minutes, so run the plan in planning software (Subsurface and others) before you splash, agree the schedule, and brief it. The question of whose computer the team follows is one to settle on the surface, never at 6 metres with a spool in your hand.

So which algorithm is best?

Less than you would think. When Los Alamos compared Bühlmann and RGBM across thousands of identical profiles, the predicted bends rates came out within a whisker of each other, close enough that the difference is not real. The deep-stop idea the bubble models pushed hardest is the one that actually failed its outcome test, which is the story in bubble trouble.

So don't agonise over the badge on the case. What actually moves your risk is a sensible conservatism setting, a slow well-shaped ascent (safe ascents), staying warm and hydrated, and not grinding out maximum dives day after day. The algorithm is the junior partner here. Your habits are the senior one.

The honest takeaway

Your computer runs cautious because it can't see the diver wearing it. It assumes a generic, average body and pads its numbers to cover everything it doesn't know about you, so it leans safe by default. Learn its conservatism controls, gradient factors on a Bühlmann unit, personal and altitude on an RGBM one, and set them on purpose. Dive shallow-to-deep with unhurried ascents, give yourself real surface intervals, and in a mixed team follow the most cautious box. Do that and the whole "my computer is too conservative" question mostly dissolves.

References

• Bühlmann AA. Decompression: Decompression Sickness. Springer-Verlag, 1984 (revised 1992).
• Wienke BR, O'Leary TR. Understanding Modern Dive Computers and Operation. Springer, 2018.
• Yount DE, Hoffman DC. On the use of a bubble formation model to calculate diving tables. Aviat Space Environ Med. 1986;57(2):149–156.
• Doolette DJ, Gerth WA, Gault KA. NEDU TR 11-06: an evaluation of decompression algorithms for use in a fleet dive computer. US Navy Experimental Diving Unit, 2011.

Train with me

Choosing and configuring a dive computer, and diving mixed-computer teams, come up in every technical and CCR course I teach. Enquire about training →