Hucho's compendium is helpful. He references work by Kunstner showing interactions between tug and trailer. Compromises are inherent. The optimal teared tug bobtailing takes a bigger hit when towing. A more SUV/wagon tug interacts better (FE wise) with trailer. His wind tunnel pix clarify what problem front transition effects solve: bow wave splash from too-blunt leading edge. Quantification by Peschke and Mankau is consistent and further shows that sloped trailers are better at keeping hook load down to control sway with speed (safety first) but a deflector proved better for FE. Waters showed that when it comes to front rounding a small percent does a lot of good but a lot does very little more (diminishing returns). That's all in Chapter 4. Hucho rocks!
Other research shows more drag coming from the rear of truncated boxes (think big eddy currents and vacuum effect) than the front. So funneling down rear flow offers more bang for the buck. Some thread on this site referenced a great summary of this by Kevin Cooper from CNRC in a 1976 article for Cycle magazine. Airstream and Don-Bur both apply that knowledge (plus some other trailers though I don't know any from Elkhart, IN).