Thats because a lot of airports don't maintain them
I was in the DCA Essendon Flying Unit when a cyclone hit Townsville way back in the 1970's. The VASIS concrete bases were wrecked so a new VASIS was installed and we the flight tests in the DCA F27. First fly down run was on Townsville 01. We were shocked to find the TVASIS lights had been reversed. When approaching low the lights showed very high FLY DOWN. When approaching obviously high, the lights showed gross undershoot.
Stopped for a cup of tea and a ponder. Turned out that the contractor had laid down the light boxes for Runway 20 on the opposite runway 01. After that was sorted and the TVASIS flight tests were OK, we discovered that during the test runs for runway 20 we were getting spurious red lights. Closer inspection revealed salt water deposits from the nearby coastline were encrusted on the red filters causing refraction and erroneous light signals. Moisture on the lamp lens as in early morning dew or erroneous indications in mist or fog, can cause erroneous light signals meaning spurious combination of fly up or fly down lights.
A fatal crash to an RFDS King Air happened at Mount Gambier at midnight and misty conditions. The aircraft flew into the ground 4 miles on final in the approach configuration in line with the runway.. The Mt Gambier T VASIS was well known by local operators to give erroneous signals under those meteorological conditions. It could not be proved the pilot followed erroneous light signals in misty conditions but it was the most likely explanation since he had called visual and his landing lights were seen by a witness. After that accident Air Services issued a NOTAM stating RPT jet aircraft not permitted at Mt Gambier.
PAPI is also not immune to erroneous light signals under certain weather conditions.
Years before during operations at Nauru, TVASIS signals were often unreliable because obese kids used to jump on the TVASIS boxes displacing them enough to give an erroneous signal. Or the kids would fire stones from a catapult into the boxes and smash the lens. Often a whole TVASIS box was u/s. That made night approaches a bit tricky since one of the vertical boxes being out of action gave the pilot a different perspective on final. For example if the first two boxes of the three above the on course lights were inop the pilot could mistake that at night for the aircraft flying one dot high. By lowering the nose to get back on slope and the top light went out as expected, in reality you were now coming on final with an indication of being on the 3 degree slope but in fact you were coming in at two dots fly down since two of the vertical fly down lights were inoperative. At Nauru, a Boeing 727 nearly over-ran after landing a bit longer then expected on runway 30 because the pilot was fooled at night and thought he was on slope when he was actually two dots fly down all the way in. One tends to believe the TVASIS on a dark night and a vague feeling of coming in high is easily dismissed as imagination or tiredness.
The answer to that is to use the DME v Height as a back-up if the DME is suitably located. Better still, never completely trust any type of VASIS at some of the more remote destinations since the maintenance could be non-existent and flight tests few and far between - if any at all.