Like every design, there is always room for adjustment.

Personally, I would prefer the fan 2/3 out, rather then 2/3 in.
The shape of the fan and the shape of the shroud place the best distance at 50/50 fan in/out.

If the fan is in or out too far, it can either allow too much air to not be effected by the fan to pull air through the rad at low rpm, or it may reduce the amount of air that can pass by at higher vehicle speed, or higher rpm, or it may do both. Exactly how much too much is will depend on the car, the fan, the radiator, the fan shroud, and how many air "leaks" are involved.

Lets see if I can paint a picture of the whole air flow deal.

Picture the radiator support on your car as a solid wall. If you move that wall forward, all the air has to go around the edges of the wall. Now, we cut a square hole in the wall (opening for the rad), and move the wall forward. Now, some of the air hitting the wall, can go through the opening (and through your rad), and some of it still goes around the wall. Once the air has passed through the wall, it sort of just bellows out and moves away.
For fun, lets put sort of a funnel in front of the hole in the wall (the front of the hood, the fenders and headlight area, and the front bumper). When we move that wall forward, now all the air that enters the funnel has to go through the hole in the wall opening (rad opening) or it backs up until it can overflow the sides of the funnel. If the funnel isn't sealed to the wall very well, we will loose some air flow through those seals as well.

Well, we can't just leave well enough alone, can we? Lets all a tunnel and another wall (the inner fenders, hood, and the firewall). When we restrict the area where the air can flow, the air tends to move faster then the air outside the restrictions. fortunately, our tunnel still has the bottom open, until we stick a rather large rectangle block (the motor exhaust, frame rails and suspension) there. The restricted space around the bottom now will restrict the amount of air that can pass through our hole in the wall, but what passes through is moving pretty fast (and takes the heat away with it). The faster the wall is moving, the faster the air passes through our hole in the wall. Life is good!

It all works pretty slick until the wall stops moving. Now we have to create a way to get the air moving. Introducing the fan. If we mount that fan inside the tunnel (behind the rad) and pull the air through the rad, the tunnel effect still works pretty well, because the air still has to pass through the hole in the wall to get pulled through. If we mount that fan out in the funnel, and there is no air trapped there, is sort of directs the air through the hole in the wall, but a large amount can just circle around the fan. It still works, but is way less efficient then pulling the air from inside the wall.

So, the fan by itself helps a bunch, but what if the motor creates a lot of heat, and/or is larger or there are headers and other stuff restricting the air that can pass out of the bottom of our tunnel? If we surround the fan (fan shroud), and seal it to the rad and to the hole in the wall, we can really pull some air into our tunnel and force it back out the bottom. We dramatically increase the efficiency of our puller fan.

This fan and shroud may be great when the wall isn't moving, but what happens when the get the wall moving? The fan and shroud can become the restriction, if the fan can't move the air fast enough. The very thing that makes it work so well when the wall is not moving, causes problems when the wall is moving, and the faster the wall is moving, the bigger the issue the fan and shroud become. At some point, the fan blocks all the air it can't move, and the heat starts to build. Somewhere we need to make concessions for the fan and shroud efficiency, because it still has it work if the wall is not moving.

Some of the compromises made are the difference between the blade diameter and the shroud opening. They (those way smarter then me) can determine how much difference those two diameters can be to fit their design perimeters. They can reshape the edges of the blades, and reshape the edges of the shroud. They can change the position of the blade in the shroud, or they can make provisions in the shroud that allow higher velocity air to pass through the shroud (some front drive car have neat little doors that can lift open to allow air to pass through the sides of the shroud, away from the fan blade). The largest problem is that they made their designs to fit a set of standards that may, or may not agree with our set of standards, and on top of that, we have probably altered the components they based the designs on.

Everything on an automotive cooling system is a set of compromises. The motor creates heat, and a certain level of heat needs to be maintained for the motor to function at its best performance level. We have to dissipate that heat when it reaches a certain temp, and we generally do that through heat transfer into coolant, and then from coolant transfer through moving air, and we have to do that through vastly varying outside air temps and conditions. We have simply covered some of the air flow part of this complex process. Gene