Poll: Algorithm Type - strong views/preferences?

[KenGordon]

For a single dive within NDL, I would consider A to be more conservative.

Where it becomes interesting is second, third, fourth etc dives. How does the computer achieve its NDL limit for the second dive? If I dive at 59 minutes instead of over an hour, do I get penalised? If so, how much?

Given that virtually all of the bubble models (RGBM and VPM) are proprietary and don't give details of how/why "penalties" are imposed or the extents of them, my strong preference is for a Buhlman

VPM isn’t proprietary.

Most of the things with RGBM in the name are gas content models when you read the manuals.

You could add the same penalties to a GF model by various means easily enough, making dives within a short time, repetative dives etc have lower GF (high).

What matters is whether you can plan the dive and there are more ways to do that than use a GF dive planner.

 

[RainPilot]

Speaking purely recreationally, if computer 1 with “A” algorithm gives 30 minutes BT and computer 2 with “B” algorithm gives 25 minutes, for 2 otherwise identical dives, How can computer 1 not be considered more conservative?

I suspect you meant it the other way around? If a computer gives me LESS NDL then I assume it is more conservative.

 

[uncfnp]

I suspect you meant it the other way around? If a computer gives me LESS NDL then I assume it is more conservative.

Thank you yes. The only excuse I can offer is that it was past my bedtime.

 

[dmaziuk]

Thank you yes. The only excuse I can offer is that it was past my bedtime.

It's part of the reason why NDLs out of context are such a bad indicator: I think "conservative" is in types of profiles you get penalized for on multiple dives etc. E.g. if algorithm "A" says you're completely clear after 3:00:01 and can go to max. NDL again, while algorithm "B" says you're completely clear only after 30:00:01 SI, I'd call "B" "conservative". I think "liberal" is a poor choice of word for "A", though. "Radical" maybe?

 

[Kharon]

I'm getting confused. I better head up.

Yep - think I'm narked.

 

[aquacat8]

Aaargh computer algorithm discussions makes me feels so stoopid— is there an old thread or an article someone can point me at that explains the differences for someone who really thinks a Seiko watch is looking good?

 

[dmaziuk]

Aaargh computer algorithm discussions makes me feels so stoopid— is there an old thread or an article someone can point me at that explains the differences for someone who really thinks a Seiko watch is looking good?

's easy, OK? -- nobody really has a clue, we're all just stayed in the holiday inn express last night. Different computers show different numbers, if you never get to 0 you needn't really care. Much.

Seriously, though: computer implementations are company s3kr3t, we don't really know what's in there. That's all of them except OSTC who release the relevant bits into open source every once in a while. Everything else is a matter of trust: Buhlmann and VPM are published and we generally trust Shearwater to have implemented them right. "PZ+" is something Pelagic did to Buhlmann and nobody here has any idea what it is. "RGBM" has a few musings published by Dr. Wienke with no clear picture of how it actually works -- but it is rumoured that low-end "recreational" computers don't have enough oomph to run it anyway so what's inside is some dumbed-down version that runs "close enough". DSAT has been tested pretty well though I'm not sure the details of the actual algorithm are public knowledge. We trust Pelagic programmers to have implemented it right.

And then there's Cochran whose mil. spec. Thalmann/Vval-18 has been tested and verified and all that, but the models they sell to us civvies apparently run something else. And they can't tell you what it is: it's a military s3kr3t and a company one.

HTH

 

[Rollin]

It's like you want to get a nail in a wooden wall. But, you are not allowed to get through the other side, nasty things will happen. And you have only one shot to do it right.
So you calculate the force of the hit with a one pound hammer, the resistance of the wall, the length and material of the nail. You test your hitting force, test it again, let your brother hit, and measure that, your father and mother too.
Then, you'll know how deep the nail will be if you hit it a certain way, over and over again. That is, if you had a cereal breakfast and normal night of sleep. You test again with more sleep and eggs and beacon breakfast, and measure the difference. Just to be sure.

Not really convinced of the outcome, you will hit the nail with 53,5% the force you calculated, test some more, and use a 47,8% shorter nail. And a smaller hammer.

Being an engineer, you even design a hammer that will hit with the same force over and over again.

The nail finally ends up in the wall, and the painting looks really good in the living room.

Next thing, your friends, the football club members and family use your calculated strike force to hit all kind of nails in all kinds of walls with all kinds of hammers.

That's how I see it. And different people made different calculations, all claiming the nail will not get through. Till it does, you being stronger than average or hitting a weak spot in the wall. But the chance on that is 1 in 10.000.

 

[aquacat8]

So it’s a crap shoot, unless you’re Bond, James Bond, then just glance at your Rolex and go back to knife fighting in 25 feet of turquoise water.

 

[dmaziuk]

So it’s a crap shoot

But it's very important that you pick one that will never limit all your future dives. Or else.