The "sunlight" test is so subjective as to be essentially useless IMO but I would be interested if someone were to test it against instrumentation to "see" if it works (pun intended
< ).
That said, based on what I have "seen" (pun intended again) a filter has to be pretty obviously nasty to show major restriction. I have had gauges on my rigs for the past, 10 years or so, and still haven't had to change a filter on the roadgoing vehicles. The Farm equipment is a little different but the only one that showed any significant restriction within a year was in my John Deere combine after a season. If you saw the dust in which a combine operates, you'd know why, and know that's equivalent to about 150,000 miles of dirt exposure (or more IME) for a passenger car. It was literally caked with dirt BUT was still only showing yellow (don't recall the actual number, but near 10").
The filter in the Honda Accord is about 11 years old at this point (installed in 2007) and has just under 54K miles on it now and the gauge has not moved. I've looked at it once since I installed the gauge in 2008 (tripping at each service and doing a WOT to reset) and there was some junk but nothing particularly serious (I vacuumed it off). Bear in mind I live in farm country down a gravel lane that's a half mile long, so my service is probably more severe than some-o-ya "city folk"
< ).
A few years back when testing the intake system on the late, great '05 F150 on a flowbench, we tested some throwaway filters I had collected from a nearby Ford dealer. Even the nasty ones were not all that far off a new filter.
Along the same lines, there is a white paper on the web (I printed it out but didn't bookmark it) by Marius Toma from 2015 where the University of Bucharest tested about 90 filters collected from the dealer for a popular brand of commercial Renault-powered vans. They recorded the miles at which the filters were collected... most of them changed at the factory interval. They used the intake system of the van to calibrate the restriction of a new filter and the built in restriction of the system and then began testing the collected filters. Turns out NONE of the of the collected filters exceeded the standard 10" restriction over a new filter, even one with 137,560 KM on it. Those changed at the factory interval were only 36% above the new filter for restriction and still 60% below maximum restriction. A similar test series to this done at another place and added chassis dyno testing to the equation and tested the most restricted filter (still below the point where replacement was indicated) and found the engine was generating virtually the same power as with a new filter (had not reached the restriction point yet where power was effected.. they noted the intake plumbing, not the filter, was most of the intake restriction).
And bear in mind that the first 10% of an air fitler's life is when it is the least efficient and most of the lifetime of contaminants that pass thru the filter do so in that first 10% of use. This is why you change the filter by restriction, not looks or an arbitrary interval that is calibrated to a worst case scenario (if you want to be generous) or to sell more filters (if you want to be cynical).
Bottom line, changing your filters too early and too often is counter-productive to longer engine life. The intake system is THE major source of outside contamination into an engine, so IMO, it's the MOST IMPORTANT part of making an engine last. It also helps the oil and oil filter last longer due to less contamination inputs.
The variables are that some engines have a great deal of filter area so can go a good long time before restriction effects power output, and some have less area so those will be effected sooner. Again, the restriction gauge tells you when in every case and in every operational situation (cleaner air or dusty conditions).