So, if the trial was always using OBF, how did pushing the interim back help the final alpha allocation? It should have hurt it.
I'd have to play with OBF to completely answer your question but perhaps some simpler scenarios will provide some insight:
1) If you keep the alpha for an interim constant then as you move the interm to the right (towards the final) then the alpha you can assign to the final goes up. So if the moving an OBF interim to the right drives down the final alpha it is because OBF allows greater and greater alpha to an interim as it moves to the right.
2) If you keep the alpha for an interim constant and move the interim to the right the power of the trial goes up.