Very good. Now go to your room.
Probably half the good luck here was due to the location of the battery, which is a layer below the passenger compartment. It makes them very bottom-heavy, so it was much more likely to land wheels down rather than roof down or some other side down.
Agree that the crashworthiness seems pretty impressive on first glance. But in uncontrolled one-off experiments like this, happenstance plays an unknowably large role. You might have to throw a dozen Teslas with crash dummies off that cliff at about the same speed to get just one survivable stop at the bottom. No way to know until it’s tried.
I’ll point out that the Fire Chief quoted in the OP’s cite said accidents there are common, survivors are rare. But rare is not the same as “unheard of”. IOW, folks in non-Teslas have also taken a similar plunge and lived.
I’ll also point out the car stopped just a bit above what appears to be the high tide line. Had it carried slightly farther it would have been in the surf depending on the tide state. That would be a much less survivable situation through no fault or benefit of the car.
Separately ...
A marketing counter to the idea that heavy batteries help survivability in cliff plunges is that I bet heavy batteries do not help much in lake plunges. Despite roughly the same air volume in the passenger cabin vs. a competing brand ICE car, replacing an air & gasoline-filled fuel tank that floats with a literal half-ton of dense batteries that don’t float even a smidgen can’t be helpful. The proverbial lead balloon comes to mind.
AFAIK, the first Tesla lake plunge is still in the future. But it’s gonna happen and it’s gonna make the news. And for all the accolades, deserved or otherwise, that Tesla will get from this accident, IMO they’ll lose a bunch more when that lake plunge occurs & the car stays on the surface for 3 seconds flat before submerging into the murky gloom below. Glug glug glug; sucks to be those folks.
Cf. one aw-shit offsets one thousand attaboys. In a fair world, the exchange rate would be a lot closer to 1:1. But it isn’t and so it isn’t.
When we first heard the news of the accident we wondered what was going on in the car, my comment was that the kids were probably scared and crying out “slow down daddy”. When I heard it was intentional my heart broke for those kids and Mom.
I’d expect the EV to be more stable, really. There’s more than enough buoyancy from the trapped air to handle the extra mass. But being concentrated at the bottom, it’s less likely to flip.
But the more important case is road flooding. Not many people drive into lakes, but lots and lots of people drive through roads where the water is way deeper than they expect. And in these cases, Teslas (and I expect most EVs) do rather well. Here’s one example of a Tesla driving through a flooded tunnel past a bunch of stalled ICE vehicles.
The fact that an EV won’t stall in water is another key advantage.
I’m shocked the car didn’t burst into flames and explode as if filled with gas bombs and igniters; the usual fate in movies and TV. Seriously l hope the kids and wife recover completely.
Musk claims (correctly) that the battery and the drive motors are sealed, but I have to wonder how long those seals will hold out under the increasing water pressure of a sinking Tesla, and what happens when some 100 kWh of stored energy suddenly short-circuits?
On the matter of driving in floodwaters, in some of those videos where the floodwaters are fairly deep, you can see the Tesla happily driving along with water beyond the level of the hood, still firmly planted on the road. Which, as you say, is a point in its favour, but it raises the question: just how far would the Tesla sink if it fell into a lake?
Musk has actually claimed on Twitter that the Tesla is capable of “boat mode” for a short period of time, with the wheel rotation providing propulsion. I kid you not, he actually said this. This strikes me as the kind of abject nonsense that Musk is capable of uttering from time to time. There are numerous videos of Teslas driving on flooded roads submerged up to their hoods and beyond. This implies that if trying to “float”, they’d be submerged much more than that – meaning probably most of the way up to the roof. So they’re trying to act not so much as a “boat” but more like a submarine! And even if it stayed almost-afloat for more than a second or two, I’d like Elon to explain to us how rotating submerged wheels are going to provide propulsion in water, which requires some mechanism to drive significant quantities of water backwards to generate thrust. It’s like Musk has never bothered to understand how a propeller works.
Not fast enough to matter. The car will happily float as long as there is not too much water in the cabin. And the cars are well-sealed, so there should be a fair amount of time before water intrusion causes the car to sink. If the occupant isn’t out of the car at that point, they probably aren’t going to get out.
The battery is a seriously strong box of metal, designed to at least fail safely in a conventional accident. I’m sure it can survive an extra atmosphere of pressure without exploding. But that’s 30 feet of water. I don’t think people usually get out of cars that far down.
An open wheel is going to fling water in all directions while spinning. But in a car, the top portion is enclosed by the wheelwell. Only the bottom part is open, flinging water to the rear, which should provide some net forward thrust. Not much, but maybe enough to get through a puddle.
ETA: Ninja’d by @Dr.Strangelove
From a passenger survival POV I doubt it matters much what happens when high-ish pressure water gets past the seals into the power electrics.
By then the people have either gotten the doors open and are on the surface well away from the car or they’re trapped in the cabin behind doors they can’t open against the water pressure in a rapidly depleting air bubble at some ever-increasing depth. With none of the gear or training submariners have in the same dire circumstance.
Here’s a video that provides some insights (action part starts around the 18:00 mark). It’s one of those car nuts trying to prove that a Tesla Model S is also a boat, or something. The Tesla was pulled into a small artificial pond by a tow cable. It sure filled with water pretty fast – even though I believe they made special efforts to reinforce the seals – and the guy sounded rather anxious to be pulled out.
I only skimmed a few parts, but at 16:45 he said they added “4 to 5 thousand pounds of weight” to the car. That doesn’t sound remotely like how a normal car would behave in a lake. That’s almost double the stock weight.
That was to intentionally make it sink, as part of the stunt. But I was responding to your point about it being well sealed. It took in a considerable amount of water over a span of just a small number of seconds, even though it was only partially submerged most or all of that time, and even though they tried to strengthen the seals.
Musk is now claiming that the CyberTruck will be able to “really” float and, retrofitted with amphibious propulsion gear, will be able to cross rivers and small lakes. He’s going to try this out across the water between SpaceX’s Starbase and South Padre Island in Texas. He’s welcome to try it – I just don’t want to be in there with him! 
Surely the designers didn’t bother to seal the car in a way to stop intrusion that it would never experience under normal conditions (after all, you don’t want the car to be totally sealed all around–passengers need to breathe). The car didn’t even sink with the 5000 lbs of weight. Which means that normally, the car would have to take on >5000 lbs of water weight before sinking becomes a serious risk. That’s going to take quite some time even if the seals are poor.
And to be clear, I’m talking “quite some time” relative to accident timeframes–a small number of minutes. Whether a car can be driven around like a boat indefinitely is another matter.
I remember that horrible story, BUT ISTR that all of the kids had been sedated, mostly with Benadryl, and the parents were both intoxicated too.
IASTR that the body of one of the children was not found for a long time.
I, too wonder what made authorites conclude so quickly that it was deliberate. Maybe the wife and/or kids told them? We’ll find out in time if they release it.
He also put those First Responders at risk while they were rapelling down the cliff to rescue them.