RMS power is measured with a resistive (non reactive) load, with in phase current/voltage. It bears no relation to the power consumed by most of todays electrical devices, which tend to take current at the peak of the waveform only, or after the waveform peak has occurred.
A hot wire ammeter is very good at measuring non sinusoidal currents. However, for power measurement, its the magnitude of the current and voltage value at a given points in the voltage waveform which will give the true power value.
For instance, lets assume a trailing edge dimmer circuit which fires 120 degrees into the sine wave. The instantaneous power at the moment of trigger will be the load current (not necessarily resistive) and the voltage 120 degrees along from the start of the sine wave. As the voltage falls the power will change tending to zero at the moment the voltage waveform reverses polarity. Additionally any phase shifted load may return power to the supply reducing the apparent wattage.
So the way I see it is that its very tricky to get accurate power readings without a fair degree of sophistication. The original electric 'Ferrari' meters (I think) used induced eddy currents and some clever magnetic stuff to get accurate power readings. Modern meters use dedicated chips to do the same job.
Still thinking about rolling my own