Duke moved the prop mass back. This would have reduced the polar moment of inertia of the plane, increasing its pitch rate. Moving mass rearward, even when it is increased to maintain the chosen c.g. can still achieve this, although the minimum turn radius will increase. Interestingly,...
If the plane balances with a greater weight on a shorter nose or has a shorter nose and shorter tail to balance it, then its polar moment of inertia can be smaller. This basically means that the plane can be pitched more quickly (higher angular acceleration). The moment of inertia is proportional to the first power of mass and the second power of the radius (distance from point of rotation); so the radius dominates. That is something of value to combat flyers, but usually not needed by stunt folks, whose planes pitch quickly enough to challenge reflexes, but also need some pitch damping to allow smooth recovery without overshooting, or overcorrecting causing point-losing wiggles.
For example, since for concentrated masses, the moment of inertia I = mr2, we can change only two things: the mass or the radius ('of gyration'). Suppose our engine/prop has a mass of 10 mass units of some sort and its center of mass is located 10 appropriate length units ahead of the plane's center of mass.
So m1 = 10, and r1 = 10. The moment of inertia contribution from the engine/prop, to be countered by the elevator's torque, is I1 = 10 x 102 = 1000 units.
If we double the engine/prop's weight and re-balance the plane, we will find that it will balance with the engine half the previous distance forward of the chosen c.g.
So m2 = 20, and r2 = 5. The moment of inertia contribution from the engine is now I = 20 x 52 = 500 units.
By mounting even an engine/prop combination twice as heavy, we have cut the inertial moment in half. The plane will actually pitch quicker, even though it will require much more lift (higher aoa) for maneuvering.
So, in short, Duke probably had a reason for that. The shorter shaft moved the prop in and probably saved some mass. This would have made for quicker turns. As we've seen above, this can be done even with heavier weights. There is another related consideration - FWIW. For light engines like Fox .35's, adding shaft extensions may reduce the polar moment of inertia by moving the largest mass (engine) back, but it can and usually does move the c.g. forward, rather than back. I think a lot of classic stunters actually got heavier and gained turn radius due to the mistaken use of these devices. The point of c.g. benefit for extensions on classic sized/proportioned planes is somewhere around engine weights of 10 oz. 'but of course, "we don't need no steenkin' math."
SK