They are very different in that the SLS landing location is known ahead of time with minimal variation. When you don't even orbit once, it's pretty easy to minimize dispersions and know something will land in the Indian ocean.
There is no "skill" with how the LM-5B stage re-enters, it is totally random and up to chance. Since it's in orbit for days at a much higher altitude, it's way harder to predict where it will land - just look at the current predictions, can't even nail it down within 1 orbit!
I think we both agree that orbiting an object is the same as flying a straight line on a curved 2D plane in space time. That means the difference is only the distance on that 2D plane. Whether 1 orbit or 10 orbit isn't really relevant.
SLS is like shooting a bullet to a target 100 meters away. LM5 stage is like shooting the target 1000 meters away (10 revolution of earth). Due to the drag of thin air, there is more uncertainties in case of LM5's final entry point and landing site. This is the same as shooting a bullet, the further it goes the higher CEP.
SLS has a smaller CEP due to its shorter range, say 1 meter. LM-5 has a CEP of 10 meters. SLS's landing region is in south atlantic? LM-5 is south pacific? LM-5 has higher CEP but also larger landing range. That larger CEP is enough, isn't it?
The final point is equally random regardless the CEP. You just need to be sure that the final landing point is within your predetermined range according to your calculated CEP. The skill is to be confident in your calculation of the location of that range and its size, NOT about knowing exactly where it lands.