The problem with pH predictions is that while it is theoretically possible to do them it is practically speaking impossible to obtain the data you need to put into the model (assuming you have the right model) without doing a lot more work than would be required to just make a small test mash and check what pH you will actually see when you make the beer. This, therefore, is what I always recommend doing. Doing it, of course, requires that you obtain a pH meter. If you don't want to do that I'd try brewing the beer with 2% sauermalz in the grist and then again with 3%. Stick with the sauermalz addition that gives you the best beer.