No, I am talking about far more problematic confusions than va vs. ba (or even ca).  I became aware of this when working on some texts that had been published in Devanāgarī, but which contained numerous nonsensical readings.  When I looked back at the original manuscripts, I discovered that many of the glyphs had been misread.  

    I later discovered additional evidence of this even when reading some Tibetan translations (lacking Sanskrit originals), such that when I back translated nonsensical passages into Sanskrit then accounted for misread glyphs, all of a sudden the passages became intelligible again.  

Below are some examples of what I mean by potentially confused and conflated glyphs.  As you can see, such glyphs would be presented unambiguously in Devanāgarī, but if the original manuscript was misread (even by a copyist), the rendering of the text in Devanāgarī would completely mask the confusion of the glyphs, which would only become clear again when rendered back in the original script.

IMO, no one should be reading anything written pre-14th century using Devanāgarī for this reason.  Such texts *really* need to be rendered in their original scripts, or at least as close to them as possible.

Best,

Paul