Not always. I used to work in the dyno room back in college and we had a Superflow hydraulic dyno. It's essentially like a large torque converter, with the engine connected to a turbine in a casing. The casing can rotate on the same axis as the crankshaft, with a pressure transducer between the housing and a hard stop. This does the same thing as the large scales on the old-school dynos. When the turbin housing is full of water, you get the max resistance on the engine and the load decreases as you drain out the water.
The Superflo was computer-controlled and would automatically run the engine with incremental speed increases up to a set max RPM. We had a small-block Chevy 350 V-8 on the dyno, with WOT tests going from idle to 6000 RPM in 500 RPM increments. When the test cycle began, the throttle opened up to WOT, while at the same time water was released from a small tank into the turbine housing, loading the engine up to the point where it was so loaded down it would turn at idle speed while at WOT! The water would be drained out a little at a time until the engine reached the max set RPM. Once the test finished, a dump valve would open and flood the turbine with water, taking the engine down from 6000 to 1000 RPM in about 1.5 seconds. Ouch.....
