My guess (and that is all it is) is that WP probably won't care all that much about your Rat experiences and the negative impact on GPA. If anything, a lower GPA at VMI would show difficulty in adapting to the stresses of such an environment. That's not something USMA is going to like. This is just an opinion.
If you go to VMI and take courses similar to what plebes take at USMA, and if you get good grades, I'd imagine you'd be in a similar boat as any re-applicant with similar stats. The ROTC and VMI experience couldn't hurt, but the grades have got to be there.
Also, and this is coming from a VMI grad who got a similar letter from USMA back in the time before electricity, it probably isn't the best idea to go to VMI with the thought of not enduring it. I mean, it's not wrong to have VMI as a plan B (you certainly won't be alone there), but to go in with the idea of using it a holding place to try for USMA again won't sit well with the folks down there. You may think that doesn't matter now, but I wouldn't advertise an intention to go to WP if you ultimately do end up at VMI.
Indeed, I think if you do decide to go to VMI, you should go with the idea that you want to be there and to graduate from there. If you should eventually decide you hate it, or that even after going through the Ratline you still can't deny the fact that you really want WP, then go ahead and try for USMA; however, I think that once you endure the Ratline, you may find yourself less interested in other schools. Again, my opinion.
On the other hand, Gen. Patton was at VMI for a year before he went to USMA. I had a BR who went to USNA after Rat year (he is the only one I was personally acquainted with who did end up at one of the SAs). People have their different reasons for attending what schools they do.
So.....in summary, my view is that if I really wanted USMA above all else, that is, that was the only way that I was going to be happy, I think I'd go to a civilian college (not an SMC) and focus on strengthening whatever weak area was identified in my file. From there I would dedicate all of my energy to compiling a successful package for re-application.