Originally posted by reh321 EVF is always a benefit of MILC; see what sensor sees.
Both parts of that statement are false.
First, you don't see what the sensor sees but the monitor's low fidelity rendition of it. (And actually the sensor in live EVF view is seldom run at exactly the shutter speed, aperture, and ISO parameters that the actual photo will use.)
Second, even if you saw what the sensor sees, it's not always a benefit. The sensor clips both the highlights and the shadows (the display worsens the clipping). But even if you saw the full DR of the sensor for true WYSIWYG, it's still not the whole scene. I'd rather have a what-you-see-is-what-is-possible (WYSIWIP) view that does not clip the ends of the tonality spectrum. (That's not to say I would not welcome a quick glimpse of an EVF just after I take a shot but I often don't want it before.)
Finally, an EVF will always lag even if the refresh rate is infinite and the sensor read rate were infinite because the whole system must wait for a sufficient number of new photons to accumulate to form a changed image in the sensor before it can be read, processes, and displayed.
There's ALWAYS some drawbacks to both EVF and OVF that drive different people to pick different designs.