Some wedge theory comments for what its worth:
The Rankine "failure plane" is defined as 45-phi/2 from vertical for a vertical wall, level backslope, and no wall-soil interface friction. The Coulomb "failure plane" will lay back more or is flatter for the same situation due to wall friction but will result in a lower calculated pressure. Pick your religion at this point. Both require that the same phi angle material be on both sides of the "failure plane" or "wedge" to provide the assumed frictional resistance from the soil shear strength thus the zone has to be somewhat larger than the failure plane envelope.
Any change in wall batter, backslope, or additional surcharge loadings will tend to cause the "critical failure plane" to lay back further with an infinite backslope and Coulomb analysis approaching 45˚. In most cases, backfilling a 1:1 zone (45˚) is the safest method to limit pressure to the simple granular fluid pressure model although it could be less. In fact, one can just go a couple of feet behind a wall and then up at a 45˚ angle with granular material and satisfy the lower design pressure state criteria. Concrete walls commonly use granular material to the rear of the footing as that area is excavated and must be backfilled anyways.
Keep in mind that there are different measurement points, one for the wall stem (or internal pressure) and one for external stability pressure (back of footing). If one is just trying to reduce stem pressure, the zone can be smaller than if on is trying to reduce external pressures as well. This is why you may see a wall backfilled with granular material to the rear of the footing and then up at a 1:1 from that point back to keep all pressures low.
