I don't use constant water changes personally. However when I do conduct water changes, I add the new water to the display and open a valve on my sump which is located just above the normal water line. The new water obviously adds to the total system volume displacing water out of the sump and down the drain. I've grown to like this approach because I don't have to remove "X" gallons of water and replace with "X" gallons of new. It's a lazy approach but assuming the new water is well mixed in before reaching the sump, I have a very stable, low impact, but effective water change.
Back when I was at the University of Washington as an undergrad, I wrote a short paper comparing the two methods. I assumed that removing say, 20 gallons and then adding back in 20 gallons was the most effective way of exporting nutrients. I couldn't imagine that simply adding in new to the existing tank and thereby displacing the old could be as effective. The numbers ended up showing that there are marginal differences but that over time, they effectively equal out. So, when I built my new system, I engineered it to allow the "displacement" method. (Let me know if you want me to post images of the paper. It's pretty dry, but some might find it interesting.)
What I'm calling the "displacement" method is really the same as the automated systems that add and remove at the same time. I just don't see a need for 24-7 water changes on my system. I still prefer to be in the room when it happens. Sit back in my chair and watch the magic happen.