[Syonyk]

You don't have to comb the NTSB records - they're aggregated by various groups.

A lot of them involve loss of control of some form or another - typically in instrument flight conditions, often enough by someone who either doesn't have an instrument rating or is badly out of practice. A VFR only pilot in the clouds has a lifespan measured in minutes.

Running out of fuel for some reason is depressingly common - and while an off-airport landing isn't automatically fatal, "pilot failed to monitor fuel in flight" is a pretty stupid reason to crash.

And avoid light twins. They're a lot more demanding when an engine fails, and typically don't handle off airport landings very well. There don't tend to be many injuries with twins - either you handle everything properly and land safely, or the aircraft leaves a small smoking crater in the ground.

A lot of it is simply looking at the sky, forecast, and deciding "You know, this just isn't a good day to put a small airplane in the sky." I consider night VFR to be fairly risky too. Clouds are invisible, visual illusions in sparsely populated areas are common, and it's hard to find a nice flat area to land if your engine quits at night.

[Natsu]

Sometimes this just makes me think that every pilot needs some IRF experience and a fuel gauge that has an alarm that says 'land now, you moron' or such.

[hbrav]

Small airfcraft fuel gauges are notoriously unreliable. If it's reading low, yes, be alarmed. But sometimes the float gets stuck, so if your estimate for how much fuel you should have burned puts your fuel lower, trust that.

At least in the UK, getting a private pilot's licence did involve a couple of hours of simulated IFR. Not enough to do anything complex. But enough to turn 180, and be able to follow detailed controller instructions.

[Syonyk]

In the US, you need a couple hours of simulated instrument time as well. One difference in the US is that you can legally fly VFR at night - even if it may be unwise in a lot of areas. A lot of countries (I believe the UK is one of them?) require an instrument ticket for flying at night.

If you're up over, say, Iowa at night? Tons of farms, flat, lights everywhere? It's fine. Pretty, you can still see stuff, and short of an invisible power line for an off airport landing, it's almost like flying during the day.

Out in Idaho, in the mountains? You're pretty dumb to fly VFR at night. There's no light anywhere, lots of hard rock, and plenty of clouds that like to appear with no warning.

But, yes, fuel gauges suck. The problem is that even if you put something like a fuel totalizer in, you still only know how much has gone into the engine, not how much is left in the wings. I understand a loose or missing fuel cap will drain a Cessna's wing tank in about 10-15 minutes.

[salawat]

Huh. Wait, aircraft use float based fuel gauges still? I thought everything had gone electro-resistive since you'll get the same measurement regardless of orientation of the tank if you place your sensors right.

I remember coming across that as a specific design challenge to overcome. Floats don't read the same when flying upside-down.

[Syonyk]

Most of the US General Aviation fleet is from the 1950s to 1980s. They still use floats.

Those who fly upside down on purpose tend to either not care about the fuel readings when upside down, or fly something modern enough. But I can't imagine an airshow performer is paying any attention to the fuel gauges. "I have a 15 minute routine, I have an hour of fuel onboard, and I'm surrounded by an airport, which is a good place for an emergency landing."