I recall reading that dynamic augment was an issue with multi-engine steam locos, and was particularly bad when the two units walked into synchronization (PRR T1 for example). So I just thought, why wouldn't it help to have the front engine quartered one way, while the rear engine was done so the opposite way? Why I think this would work is because if one side of the loco happened to be in sync, the other side would be completely out of phase, and would work to correct that. Suddenly vice versa, and the reverse process starts. Thoughts? Equations? IRL proof?