none of the above.
Wait times are fundamentally inaccurate.
They are based on the wait times of guests who have completed their waits.
Wether it be TP, where guests submit their wait times, or
MDE where the time is measured by guests who're given that red time card, those waits times only reflects the wait times of those guests, not the wait times of guests who come after them.
example : let's say that you've waited for 1 hour for the ride.
During that time, 500 guests (fictionnal and random figures) did ride that ride.
But during that time, only 250 guests have entered the line behind you.
The next guest will enter the line and see a posted time of 1h (and this time will have been set by you, basically) but since only half the capacity of the ride has entered the line, the next guest entering at 1hr will only wait for 30 minutes (half the guests, half the time) and find the wait times to be inaccurate.
Now this guests becomes the reference point for the wait time, and the posted time then shifts to 30 minutes, but right behind a 500 strong BTG enters the line on top of the 250 guests who enter the line at that time of day
you now have 750 people who enter a line which they think is 30 minutes because you've just helped Disney and MDE to measure those 30 minutes.
The last guest who will enter the line at 30 minutes posted will actually experience a 90 minutes wait.
Both TP and MDE are highly inaccurate
the accuracy of the system wil depend on its sample rate.
If there are enough TP Lines users to take samples every couple of minutes, then it's more accurate, if Disney hand out time cards every couple of minutes, MDE will be more accurate. The more samples, the more accurate
But in the end, sample times can only be measured at the load area, when guests have completed their wait
so, the longer the line, the more inaccurate the time.
Because a 20 minutes line cannot see that many guests enter the line in that 20 minutes time period
but a 120 line can greatly change things. Who knows how many guests will enter the line in the 2 hours it will take a guest to make a new time sample ?
Basically, on a 2hrs posted time, the posted time would have been accurate 2 hours ago, but now ? it's just a wild guest.
A 1hr posted time basically means : the wait time was 1hr ... 1hr ago.
and you can't fight that
it might change when MDX is able to take samples along the line. Measuring how fast guests move through the line. Then the system might be able to extrapolate how long it will take if a guest enters "right now".
and that's why, if Disney IT do their jobs, MDE could become much more accurate than TP lines
The real way TP lines could become a very accurate tool, would be for TP to come up with an app that uses "waypoints" inside the lines.
For example, send a measure when the guest enters the line, send another measure when the guest reaches a point (specific point in the line, recognisable by any guest such as SB/FP+ lines junction). The more "sample points" inside the line, the more accurate the wait time will be, because the app will be able to tell how fast a line is moving, and get more potential users in the same line at the same time.
How fast does the line move, and how long is the line "now" is the only way to accurately measure how long it will take to reach the endzone.
