No, reversal of your reversed way of thinking will make it good. Try this: the method used to move the observer at 2 m/s downwind in the realwind is a very long treadmill. Do the rest of the numbers of the realwind scenario to get the observer matching the speed of the realwind.
Like this:
10m/s (+4m/s) = 6m/s Difference in wind to observer at 4m/s downwind
6m/s (+4m/s ) = 2m/s Difference in flow to observer at 4m/s downwind
Wind to flow = 4m/s
Constant difference between wind and flow of 4m/s, independent of observer velocity
10m/s (+6m/s) = 4m/s Difference in wind to observer at 6m/s downwind
6m/s (+6m/s ) = 0m/s Difference in flow to observer at 6m/s downwind
Wind to flow = 4m/s
Constant difference between wind and flow of 4m/s, independent of observer velocity
10m/s (+8m/s) = 2m/s Difference in wind to observer at 8m/s downwind
6m/s (+8m/s ) = -2m/s Difference in flow to observer at 8m/s downwind
Wind to flow = 4m/s
Constant difference between wind and flow of 4m/s, independent of observer velocity
10m/s (+10m/s) = 0m/s Difference in wind to observer at 10m/s downwind
6m/s (+10m/s ) = -4m/s Difference in flow to observer at 10m/s downwind
Wind to flow = 4m/s
Constant difference between wind and flow of 4m/s, independent of observer velocity
There. We've made the transition from realwind to belt with the steps in between put in so you can see how it works. Does that help?
No, it is quite obvious that the relationship will hold for all values in real wind.
You have shown that. It is constant in real wind at 4m/s.
The flow in the treadmill is the other way around, so it cannot be constant That is the point.