Well, you mention the vast increase in game sales on Steam. The question isn't "did you sell more games at the lower price point." As a company, I'm not interested in customers--I'm interested in money. As the dot-bomb taught us, having lots of customers doesn't mean a thing unless you're making money off of them.
Someone below had an entertaining post about a city that had floated a bond to pay for a bridge and had set up a toll on the bridge. As the bond was paid off, they lowered the price of the toll and made more money because more people used the bridge more often. So they lowered it again and made still more money! They lowered it a third time and made less. So the whole "lower prices equals more money" doesn't work--like I said, if it were free, you should have an infinite amount of money. There is a point where you're leaving money on the table.
Setting a price is usually based on assumptions of how many people actually want this thing. For example, when movie studios started selling movies on VHS, they priced them around $80. They figured that only a small number of movie-buffs would actually want to own a particular movie. So the prices were set for the rental market--they knew that Blockbuster and local video stores were the ones buying the movies and that they were going to make money off the rentals. But when the video store bought 50 copies of a new release and, six months later, sold 40 of those copies cheap, people snapped them up. So the studios tried lowering the prices to catch those people and they made more money.
Now you look at that example and say, "See? You're making more money by selling for less!" I look at that example and I say, "Their assumption as to the size of the market was incorrect."
Assuming that developers are making more money by selling their game for $5 versus $20 (you state that sales increased, but did they increase more than 4x?), I'd be curious as to what assumptions they made causing them to set the price so high to begin with.