What is the difference between a Dummy Load and an Antenna?

Foundations of Amateur Radio Recently I started an experiment I plan to run for a year. Using a WSPR beacon and a dummy load I'm transmitting 200 mW, 24 hours a day across all bands supported by my hardware, in this case it covers 80m, 40m, 30m, 20m, 17m, 15m, 12m, and 10m. The aim of the experiment is to determine if, and to what extent my dummy load can be heard outside my shack. Why? Because I've not seen anyone do this and because a dummy load is widely believed to not radiate, despite evidence to the contrary. Together with the transmission side, I've also configured an RTL-SDR dongle, initially with the telescopic antenna it came with, now, since my HF antenna isn't being used by the beacon, I'm using it instead. It's about five metres away from the beacon, outside. It's a helically wound whip resonant on the 40m band built by Walter VK6BCP (SK). It's what I've been using as my main antenna for the past seven years or so. While I'm telling you this, my beacon has been heard by my dongle 1,182 times across all eight bands. Some of those reports were from inside the shack, some from outside, some while I was monitoring a single band, and for the past week or so, I've been monitoring all the bands supported by "rtlsdr_wsprd", 18 in all. Purposefully, this includes some bands that I'm not transmitting on, because who knows what kinds of harmonics I might discover? The receiver changes band every half hour, so over time when I monitor a band will shift across the day, this is deliberate. I don't know when a stray transmission might suddenly appear and this will give me the best chance of hearing it, short of using 18 different receivers. At this time, my beacon hasn't been heard by any other station. I'm not expecting it to, but that's why I'm doing this experiment in the first place. I'm not in any way reaching any sense of "DX on a dummy load", but it got me thinking. My beacon can be heard, albeit by me, from five meters away. So it's radiating to some extent. I've already discussed that this might come from the patch lead between the beacon and the dummy load, or it could be the dummy load itself, or some other aspect of the testing configuration. Regardless of the situation, there is a signal coming from my beacon that's wirelessly being heard by a receiver. That's the same as what you'd hope to achieve with any antenna. So, in what way are an antenna and a dummy load different, and in what way are they the same? Whenever someone asks this, the stock answer is that an antenna radiates and a dummy load doesn't. My experiment, 20 days in, has already proven that this distinction is incomplete, if not outright wrong. Even so, if we take it on face value, and we say, for argument's sake, that a dummy load doesn't radiate and an antenna does, then how do we materially distinguish between the two? How does an antenna compare to a dipole, Yagi or vertical antenna and where does the isotropic radiator fit in this? The best I've come up with so far is a spectrum line comparing the various elements. Let's say that at one end of the spectrum is a dummy load, at the other is an isotropic radiator, to refresh your memory, that's the ideal radiator, it radiates all RF energy in all directions equally. Somewhere between the two ends is a dipole. We might argue if the dipole sits equally between a dummy load and an isotropic radiator, but where does a Yagi or a vertical fit in relation to the dipole? Also, if you turn a Yagi in the other direction, does it change place? So, perfect this notion is not, but here's my question. What's the measurement along the axis between the dummy load and the isotropic radiator? It's not SWR, since the ideal antenna and a dummy load share the same SWR, unless this line is a circle that I don't know about. It might be Total Radiated Power expressed in Watts, but that seems counter intuitive. It would mean that in order to determine the effectiveness of an antenna we'd need to set-up in an anechoic chamber, basically a warehouse sized room where incoming radiation is shielded to some predetermined standard. Do we measure gain using a VNA and call it a day, or is there something else going on? Remember, we're attempting to quantify the difference between a dummy load and an antenna. In case you're wondering, I'm asking the question. In the 15 years I've been part of this community, I've never seen any coherent response. The Internet seems to return a variation on the radiation vs. not-radiation pattern, but so far I've not seen anyone quantify this, or perhaps I haven't understood it while it was staring me in the face. I even checked the syllabus for the three license classes in Australia. The single reference that the regulator appears to specify is that at the introductory level you are required to, wait for it, recall that when testing a transmitter, a non-radiating load, or dummy load, is commonly used to prevent a signal from being radiated. Very illuminating. Obviously my dummy load is of the wrong type, the radiating variety. Which begs the question, if there's an ideal radiator, is there a theoretical ideal dummy load that doesn't radiate in any way, and if so, how far away on this line is it from my actual dummy load? Over to you. What are your thoughts on this? Better yet, got any references? I'm Onno VK6FLAB