Folks often ask for a typical plot set between these different methods.
While this is no such thing a typical GNSS day, we hope this will serve the need.

Here is a simple plot (using RTKLIB and one of our reference stations, with L1 only ) that clearly shows the typical positional variations and offsets you will typically observe with the same equipment and different corrections source.

• The Red is using SBAS/WAAS, corrections delivered by satellite and the same set used from Alaska to Mexico (this station happens to be in LA). It show the typical sub meter variations and the typical offset from truth of about a meter. [Not shown is any allowance for correction stream drop outs which are inducted by the local environment blocking of the geostationary SV, not an issue for planes but a big constant issue for cars.]
• The Blue is using traditional DGPS from a nearby station (50 km away), using RTCM message type 1. These corrections are what the Coast Guard now provides, or what some call “first DSRC generation” if you prefer.  This corrects the code, not the carrier phase, so it is a bit crude, but nearly 100% of devices today accept it.  The plot shows that the range of variation is about the same but now it is centered about the truth point.  [Not shown is how the correct offset is able to be maintained between multiple rovers using different SVs constellations, another major concern for vehicle safety uses.]
• The Green (the tiny dot in the very center) is an RTK navigation system operating in a fixed Ambiguity Resolved mode, using RTCM message types 1004 taken from a local station. It shows the typical ~100x reduction in the positional estimate variation, varying about a single centimeter. Yes the very tiny dot you see.  Expect as much as a few decimeters when you are moving about in an automobile due to frequent cycle slips.  Welcome to the future, it has taken 30 years to get here.

It is hard to give a scalar answer to the complex question “how accurate would I be with correction type A versus correction type B” -, but this image serves to give a fair insight into the variations  you can expect to observe. [See Note A]

Note A: at least until the sun wakes up.  For the past solar cycle the level of global TEC activity has been very low, making the need for localized corrections less vital.  If the level increases in the coming cycle (as is expected) or there is a day or a week with marked solar active, watch out!  Corrections become essential to maintain overall system accuracy in many application when this occurs.