Our fuel gauges are as accurate as they come, but I wouldn't want to rely on them. Pointers can stick. There are only two gauges with three-way toggle switches, so you don't see all tanks in one glance.
Our plane has six tanks and we switch manually between them using two separate fuel selectors (one for each wing with three positions each). To keep track of our fuel state at any time, we log a detailed record of the contents of each tank, the dip-checked amounts, the fuel used and the fuel remaining before and after each flight, plus the fuel used from each tank between switches, in a separate "fuel log".
NO WAY we wanted to risk getting caught out not knowing *exactly* what our fuel state is or getting in trouble like some unfortunate pilots in the Comanche's long history have. Without keeping some record, it is just too easy too make a bad mistake!
Some Comanche pilots like to avoid trouble by just flying on the main tanks with both switches selected ON, effectively reducing the need to switch. One problem with this is that you don't know from which tank the engine is actually feeding. Another problem is that the mechanical fuel pump's efficiency can lead to fuel crossfeeding from one wing to the other via the central sump. Things like unusual attitudes, turns, and having the electrical fuel pump on further complicate matters, until you end up having no idea at all how much fuel is in which tank while you are flying. Since we don't want to refuel after each flight, which is pretty ridiculous after a one hour flight when you have 10+hours endurance, this way of operating is not suitable for us.
We have a wonderful instrument called the Shadin Miniflo fuel computer. Besides being an invaluable aid in leaning the mixture correctly, it plays an important role in our fuel planning and logging.
Before a flight we plog our expected fuel usage at the planned power settings for the planned legs of our route, expected time to climb, cruise, descend, etc. and we set a procedure for the order in which the tanks are switched and the times to run on a tank given our load balancing needs (This is useful because we have a wide wingspan and 30 minutes flying on a tip tank makes quite a big difference).
We then fly following our planned switches and we note the amounts used at each switch.
This all may sound like a lot of work, but the workload is minimal and just part of our normal operating procedure at the time of the walkaround (dipping and logging) and writing down the amount used at the times a tank is switched. We like the way it automatically adds awareness of our fuel safety. Double-checking the fuel computer against dip-checked amonts and uplifted amounts gives us several opportunities to spot errors in resetting the fuel computer or calculating errors in the log.
In November I got ramp checked. Among other things I was asked to produce proof that I had planned my flight with sufficient range. This was a bit of a laugh because there was about eight hours remaining in the tanks when he asked. When I showed him the fuel log he was so stunned that I wondered if he had ever seen one before. The C414B I have been privileged to get acquainted with had eight tanks to keep track of, so the need for some kind of procedure is not unique to the Comanche.
Most light aircraft have only two tanks, but I know of several types that have long range capacity and extra tanks so for these fuel mismanagement risks might be mitigated by strict procedures as well.