well, first, there are two principles involved here. None of it all that complex, but quite interesting (at least in *my* mind...)
aside from the *cool* factor, I am assuming that you are asking about the physics behind it. Why and/or *how* it can work. Please correct me if I am wrong.
1) balance and counter-balance. *generally* speaking, if you were to make a rocket with only ONE motor, and it was canted, it would be unstable as there is a sidewards force acting on the bottom of the rocket. *IF*, however, you were to place a second motor, canted in the opposite direction, this sidewards force is cancelled (or balanced) by the second motor. Assuming both motors have identical thrust curves, they will exactly counter (or balance) each other, resulting in a straight flight.
2), the above was speaking generally and didn't take into account a rather interesting *specific* regarding canted motors. To understand the stability of a rocket, you also need to understand the principles of an "Object in Free Fall" An object in free fall (let's ignore atmospheric properties for this discussion) will not rotate or spin unless acted upon. If I were to apply pressure to the object on some point and direction that was NOT at the center of gravity (CG), then the object would begin to rotate (spin) in that direction, and about the center of gravity. HOWEVER, If I were to apply pressure right *at* the center of gravity (or somewhere on its surface along a vector pointing TO the CG), then the object would *move* in the direction that I am pushing, but would *not* rotate or spin. It would remain *stable*
The same consideration can be made in designing rockets with canted motors. If the angle of the motor results in a line (vector) that passes *through* the CG of the rocket, then even if only 1 motor is burning, the rocket will be stable because the motive force will be pushing on the CG. The tendancy would be for the rocket to go off to the side slightly, but would still be pointed up and flying true. The only problem with this "thought experiment" is that as the motor burns it uses up fuel and thus becomes lighter resulting in the CG moving forward so the motor would no longer be pointed at the CG (but the effect is miminal and the larger the model, the less the effect)
To experiment with this principle, take a 12" ruler and (be quick now) place it on your finger at the 5" mark (not balanced) and thrust your hand upward. You will see the ruler spin into the air (you pushed on a point NOT at the CG). Now, balance the ruler at the 6" mark (the CG of the ruler) and thrust your hand upward. You will see that the ruler no longer spins. Not a perfect test as many factors can induce spin, but you will see the effect well enough.
Hope this helps more than it confuses...
jim