There has always been that conundrum about real-world testing vs simulation. Simulation is only as good as your math - and must be mistake free for your results to be valid. But the real problem with simulation is that you tend to get out what you expect to get out. On the other hand, real world testing, while not infallible (such as string testing with too long a rocket or too short a string) will give to real results whether they are what you expect or not.

It reminds me of the current battle over CG special effects vs real world models and pyrotechnics. These days, CG is cheaper and gives you just exactly what you expect it to give you. On the other hand, models and pyro can give rise to what one SFX model maker referred to as "happy little accidents" - unexpected results that look a lot cooler than what the artist expected/envisioned. As another (totally unrelated) example, if Fleming had been using computer modeling in his experiments, he never would have discovered penicillin - such results would simply not have been programmed in in the first place.

Bringing this back on-topic, a wind tunnel doesn't have to be large and elaborate to work and be useful for model rocketry. A simple one can be built using a honeycomb to BT-20 tubes, several large pieces of cardboard or foam-core and a portable fan (a box fan would work, a small vortex style fan would be even better). I know that Estes published the plans for a small wind tunnel in one of their tech publications - look for Classic Collection TR-TN or TR-5. TR-11 (Aerodynamic Drag of Model Rockets) also has a lot of useful info as well.

Greg