A flow test measures how much air a part passes at a given pressure difference, and how much resistance it adds while doing it. On a bench the part is connected to a flow source and the reading is taken as a number, usually with the part tested on its own rather than as part of a system.
That number describes the part. It does not describe what the engine will do with it. The two are often confused, and the confusion sells parts that read well on a bench and deliver very little once installed.
A bench pushes a steady stream through the part. An engine does nothing of the kind. Air in an intake moves in pulses set by valve events and, on a charged engine, by the compressor. Exhaust gas arrives in bursts and is hot, expanding and changing density as it travels.
Steady figures are still useful, because they compare one part with another under the same conditions. What they cannot do is reproduce pulsation, temperature or the way a part interacts with everything upstream and downstream of it. A bench result is a controlled comparison, not a simulation.
Reducing restriction is generally good, but the gain only becomes real where the part was the limit. If the filter, the throttle body or a squashed hose joint is the narrowest point in the tract, a free-flowing pipe ahead of it changes very little. The restriction simply moves.
There is also a cost to over-sizing. A larger bore holds more volume that has to be filled and emptied on every change in load, and it can slow the response of a system that was actually well matched. Flow capacity is one property among several, and it is rarely the only one that matters.
Restrictions hide in places that are not obvious from the outside of a part. A step where two diameters meet, a gasket that protrudes into the bore, a weld bead on the inside of a joint, a hose end that intrudes, or a bend that has been crushed rather than formed all cost more than the nominal diameter suggests.
On an exhaust, the same applies after the collector, along the mid-section and at any change of section. Measuring each join as well as each pipe is the only way to know which piece is setting the limit. The smallest effective section in the chain is the one that decides the result.
It cannot tell you about temperature. Density changes with heat, so the same volume of air carries different mass at different points in the system. A figure quoted without saying where and how it was taken is difficult to compare with anything else.
It cannot tell you about noise, resonance, fitment, durability or how the part behaves under vibration. Nor does it capture turbulence. Two parts can pass the same quantity of air while one does it smoothly and the other with separation and swirl inside the bore, and the engine will not treat those two the same way.
Treat a flow figure as one input. Establish where the restriction is, decide whether the part in question is that restriction, and then check the change against the rest of the system. If the intake side is being reworked, start from the air intakes end and work back through every joint to the throttle.
On the exhaust side, the same logic applies along the run rather than at one point. Comparing sections against each other is often more informative than comparing a single part against a claim, which is why the exhaust pipe range is matched piece by piece to the layout of the car rather than sold as one universal answer.
