Unfortunately, some divers are confused or irritated by the fact that SAUL is a probabilistic algorithm, particularly those looking to compare it directly with other existing algorithms in terms of what dive profiles are permitted.   The comparisons they would like to see are, essentially, between NDLs.  SAUL doesn’t have a set of NDLs to compare with.  Here’s why not.

First, a clarification:  It’s not that SAUL doesn’t use NDLs at all, just that they’re not pre-set ones.   In effect, the diver calls up an individualized set of NDLs by setting an acceptable degree of risk (of  decompression sickness).  One “size” does not fit all NDL needs.  Reasonable people can – and do – differ on how much risk they are prepared to take, both in general, and in specific situations. SAUL is sufficiently accurate that it can viably use a probabilistic algorithm, and it chooses to do so for the following reasons.

1.   In diving – as in other aspects of life – adults with access to adequate information are presumed responsible enough to make decisions for themselves (and for any children under their supervision).

2.  One important piece of the information divers need to keep in mind is that there is always some degree of risk in diving.  Over-reliance on fixed NDLs may mislead some divers into forgetting that fact.   An NDL does not mean that divers are always safe under that limit, always in danger above it.  Actively choosing an acceptable level of risk helps divers keep that in mind.

3.  The most important reason, though, is that a probabilistic algorithm is the most realistic way to deal with the occurrence of decompression sickness in diving.  This is true even just considering the general reasons outlined in an earlier posting   (NDLs, etc.). But it is the chaotic behaviour of bubbles themselves that makes a probabilistic algorithm particularly appropriate for any serious attempt to deal with the risks in diving.

Bubble behaviour can vary drastically depending on exactly where, and under what circumstances, the bubble arose, its size, whether it has entered the bloodstream and, if so, where it’s carried from there.  Most of these factors are matters of chance and unpredictable in advance.

I mention “where the bubble arose” as a factor because my research found that bubbles embedded in a soft elastic solid (i.e., muscle or cartilage tissue) will behave very differently from bubbles in a more liquid environment like blood.  In some cases, a bubble in a soft elastic solid may persist for long times, or grow, even when the tissue it’s in is undersaturated.

Luckily, predicting the behaviour of individual bubbles is not necessary in order to manage the danger of decompression sickness.  In general, algorithms try to do this by  calculating, in accordance with their particular underlying models, the accumulation and dispersal of excess nitrogen in the body.  Most algorithms then set NDLs based on the calculated nitrogen load.  It’s obvious that the frequency of decompression sickness is strongly correlated with the presence of excess nitrogen in the body.  And it’s also true that the presence of excess nitrogen in the body is correlated with an increase in the number and size of bubbles and, therefore, an increase in the cumulative  probability that one or more individual bubbles will cause problems.  This cumulative effect of bubbles on decompression sickness is already taken into account by calibrating algorithms with known data.  A more accurate algorithm (like SAUL) will do it better.

But it’s equally obvious that decompression sickness can, and sometimes does, occur when it seems to be “undeserved”.   Since bubbles are believed to be the initiating cause of decompression sickness, certain algorithms have purported to account for bubble behaviour in their calculations.  For reasons described above, this does not, and cannot work.  In my view the only proper way to take into account individual bubble behaviour is to recognize and work with its chaotic and therefore probabilistic nature.

(For anyone interested, my latest paper, Gas bubble dynamics in soft materials, has just been published by The Royal Society of Chemistry and has now been posted here under Articles.)

Once the degree of nitrogen saturation is accounted for, the additional effect caused by bubbles is effectively summarized by the bit of verse below.

on probability

Leave a Reply

Your email address will not be published. Required fields are marked *

Captcha Captcha Reload

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>