Modern evidence (2011 NEDU study) points to the oldest model-based algorithm in use, Buhlmann with flat gradient factors, usually known as ZHL-16, being the best. "The best" means it can either give less deco time for the same safety, or give more safety for the same deco time - you decide. Gradient factor 100 is the fastest safe deco, lower GF like 50 or 70 (rarely 30) is extra margin of safety and less tissue saturation. You can choose any GF you want and keep your tissue supersaturation as low as your air and dive plan allow.
This distinction is important in decompression diving, whether it's technical or unplanned deco that may happen in rec diving. If all you plan to do is NDL diving, computer and algorithm choice won't affect your deco efficiency, simply because you won't be following your computer for ascent, but rather your divesite's depth, your dive plan, and rec diving practices. The only difference the computer will make is telling you to go up sooner or later, and you can always decide to go sooner on your own. It's actually recommended in rec diving to have a few minutes of NDL left.
That's the theory. The practice is a bit more complicated. Some people will "ride the computer up" - ascend just a bit when their NDL is 1 minute, till it's 3-4, repeat until the stop. This is an unsafe practice and some computers will cut the diver off more aggressively than others, but that's protecting you from yourself. Otherwise, the cheapest computer that has a deco mode and doesn't flood (i.e. not Mares) plus reasonable conservatism can keep you just as safe as a Shearwater.
I've heard a lot of people swear by Suunto, telling stories about how others got up with mild DCS, Suunto users didn't. Too many to keep my normal skepticism against such claims unshaken, so there might be something to it. This sounds contrary to the poor deco efficiency of the RGBM model that Suunto use - but Suunto computers are quite conservative, which might help.
Their algorithm is also a complete black box, so it's likely that Bruce Wienke, the only person in the universe who has any idea what a Suunto computer is doing (because it's definitely not the same algo as in other RBGM computers, I've been down with both and they differ too much), has inserted some secret sauce rules against unsafe dive profiles that have nothing to do with the RGBM. Not sure if this "magic" can be replicated with simply setting ZHL GF lower than anyone ever does.
It's personal opinion and many would disagree, but I would not give much consideration to algorithms other than ZHL-16 with GF and Suunto's whatever-it-is. ZHL+GF is the most efficient, science-based, and adjustable algorithm; Suunto has a strong safety record going for it, however they got it. Other algos, they're kinda there, but I can't think of a strong reason to use them.
As a dive leader, it's generally a good idea to use a similar computer to that of your clients. Saves a lot of explaining. OTOH, if you go on to be an instructor with a non-technical agency, you may have to push your margins in things like CESA demonstrations, and a less conservative computer that keeps you just safe enough to stay on this side of the chamber can be good too (the job's not all it was cranked up to be...) Suunto will just lock up and say "Err" if you go far enough into the danger zone.
Another practical concern is how easy the computer's screen is to read. Suunto lose out here: their only color screen model is the Eon, which is large and very expensive for what is is, a big-screen DX. With other makes, there are cheaper options. Some are close to your price range - the Oceanic VTX can be had for
$520. Color screens require frequent charge-ups, but they're really better. Unless you want to wear it as a watch in non-diving settings, though; the only color screen computer easily mistaken for a watch is the $1.5k Garmin Descent. Is daily wear a factor?
Air integration is useful if you plan to have your own regulator set. AI is much more accurate than analog gauges and the data is right there on your wrist. There's not a lot of PDC features that affect safety in rec diving, but air integration with its
accuracy and audible low-air alarms is one. It's not perfectly reliable, so you still need an SPG (and you need it for teaching anyway), but it usually works. It's an expensive feature, and I guess that $550 D4i comes without the transmitter; not much point to the "i" then.
So confused.
