I'm rather a newbee at this, but I'll join the fray just for the fun of it.

It seems like JohnW and MajorTomski can both be right. John notes that "... torque ... is made up of four elements which cause or produce a twisting or rotating motion around at least one of the airplane's three axes". It seems to me that the key here is that the torque can be felt on any of the three axes.

1. Torque Reaction from Engine and Propeller
This would be torque (twisting moment) on the roll axis as the airframe's reaction to the drag of the prop in the air and the force to accelerate the prop, if the RPM is changing. Down thrust or side thrust would not affect this.

2. Corkscrewing Effect of the Slipstream
It seems to me that a turning prop must impart some type of turning (corkscrewing) motion into the propwash to preserve angular momentum. Assuming that the propwash continues back along the thrustline, it will impart torque on the roll axis through its effect on the flying surfaces. However, there may also be torque on the yaw axis because there is a lack of symmetry from the fin/rudder being above the trustline and nothing being below it. It seems to me that the tendency would be to push to the tail to the right, and this would require engine right thrust or right rudder to offset it.

3. Gyroscopic Action of the Propeller
If I remember my physics correctly, this is called Gyroscopic Precession, and would only occur when the plane was yawing or pitching, and it would produce a torque in the other axis (not the prop rotation, roll axis). I would expect this effect to small except, perhaps, in extreme spins (when the pilot would be occupied with other, larger, forces).

4. Asymmetric Loading of the Propeller (P Factor)
Asymmetric loading of the prop occurs when the angle of attack of the prop blades varies through a full rotation of the prop. For this to occur, the prop must be moving foward into the air at an angle that is offset from its axis of rotation. Thus, P-Factor will be most pronounced during takeoff when the tail is on the ground (ground effect issues aside) and during high alpha attitudes. This will produce a tendency for the plane to pull to the left. (This also occurs in boats where the prop shaft is not in line with the boat's motion.)

So, I don't see that there is any torque from the engine rotation that needs to be countered, and therefore MajorTomski is correct that offset will not counter torque on the roll axis. (Aileron trim might be used to counter engine torque on the roll axis.) However, P-factor and asymmetric loading of the tail surfaces (fin and rudder) can induce torque on the yaw axis, and engine thrust offset can used to correct this. So I agree here with JohnW.