I think it can be helpful and educational (and dare I say fun) to sit and have discussions about antennas and the math behind them. However, In some cases like this, I think it is a little like arguing about what sports car is better based on the published specifications when, in fact, the best thing to do is get out and drive.
I would suggest just picking something “good,” get out there, and dont look back for a while. With experience and/or a desire to tinker at a later point, you may try different things, but just picking something and going with it will pay far more educational dividends than constantly changing. In this hobby and many others I have been in, people look for the perfect killer antenna or recipe or whatever. The truth is, it doesnt exist.
I currently use a resonant EFHW with a home built 64:1 unun. In just getting out and using it, I have learned a lot about it and have been doing well with it. Is there something better out there? Maybe. Is there something better out there for me? No, not currenty.
I would like to stress that for a SOTA antenna, theoretical efficiency is only one of the many variables for its actual usefulness and performance. In my SOTA antenna projects, I now tend to start with what is doable from the perspective of handling and then test whether it works sufficiently well. For instance, my two SOTA vertical designs:
are mainly inspired by an optimized handing on small or crowded summits. They are less performant than a full-size dipole (center-fed or endfed), but in practice, I can manage more QSOs on a challenging summit because I am QRV within 1 - 2 minutes. Still, I can work @SP9AMH with QRP2QRP regularly.
A dipole / EFHW with traps might be a little less efficient than a linked dipole. But when in a hurry on a summit, the antenna with traps allows me trying for a nice DX catch on 20m at the end of an activation in winter in a few more minutes, while I would not have the time to take down the antenna for a band-change.
As for the “any antenna design works” argument, I think we should add that this is true as long as you do not make any mistake building and tuning it: If you just look for a low SWR, you may never notice very bad construction mistakes that cause a lot of losses, for instance
long coax for sections with a very different impedance from 50 ohms,
shorts or cold solder joints,
capacitative coupling into nearby metal or otherwise conductive objects,
So a WSPR performance comparison against a reference design is always useful. Low SWR just assures an impedance match to the transceiver, so that the finals do not get blown away. 80 % of your antenna can be actually a dummy load, while you can still achieve a 1:1 SWR, as any full dummy loads proves.