Captain Radials discovers WSPR, invites “methodology critique”…

and almost accidentally does science

Every now and then, an influencer does something rare: he gets tired of the usual “S-meter Olympics” and tries to build a fair A/B test.

Captain Radials (freshly promoted to Captain Statistics) announces he will use WSPR and ChatGPT to analyze antenna performance. He alternates antennas in 2-minute slots, keeps the station constant, checks ALC/SWR/power, and records everything so nothing is hidden.

Honestly? That is already a massive upgrade over “trust me bro, antenna B sounded more sparkly.”

The part he did right (yes, I can say nice things)

• Fixed timeslots (WSPR forces discipline).
• Same band, same radio, same power, same window (good: you’re not “testing propagation” on accident).
• Checks ALC / SWR / power output (so you don’t secretly “win” by feeding one antenna harder).
• Publishes the raw session (the internet can’t accuse you of selective editing… only selective thinking).

So yes: as a process, this is far cleaner than most online antenna “tests.”

The one flaw: he tested the antenna… plus the entire installation

Here comes the snag, and it’s a big one:

You can’t A/B “vertical with radials” vs “horizontal delta loop” if the vertical is installed as a proper vertical system… and the delta loop is hugging the lawn.
That’s not an antenna comparison — that’s a launch-geometry comparison.

Because “antenna” is never just the radiator. It’s the whole system: height, surroundings, ground conditions, nearby metal, feedline routing, local noise sources, and the glorious chaos of backyard physics.

If antenna A is a vertical with a real radial field, and antenna B is a low-to-the-ground horizontal delta loop near fences, gutters, and whatever metal surprises the garden provides… congratulations: you’ve discovered which one got the better living conditions.

40 meters: a lawn-hugging horizontal delta loop is basically an NVIS machine

Let’s put numbers on it, because physics is rude like that.

On 40 meters, wavelength is roughly 43 m. Now look at common “easy” backyard heights:

• 2 m height is about 0.05λ (yes, five hundredths of a wavelength)
• 6 m height is about 0.14λ
• ~21 m height is about 0.5λ (half a wavelength)

A horizontal loop (including a horizontal delta loop) down at ~0.05–0.15λ tends to launch a lot of energy at high elevation angles. That’s not “bad.” That’s just what it does. It’s great if your goal is regional coverage (NVIS-ish behavior). It’s less great if your scoreboard is “who gets spotted DX.”

Meanwhile, a vertical with a decent radial system typically produces much more low-angle radiation (assuming the radial field is not decorative). So in a WSPR “DX visibility” contest, you can end up comparing:

• A low-angle launcher (vertical + radials)
• A high-angle launcher (low horizontal delta loop)

“My vertical got more DX spots than my low horizontal delta loop” is not a revelation.
It’s a polite way of saying: “I built a cloud warmer and got cloud-warming results.”

What a more honest comparison looks like: raise the delta loop to ~½ wavelength

If you want to make “vertical vs horizontal delta loop” claims with a straighter face, the delta loop needs a fighting chance at the same game.

On 40 meters, ½ wavelength is about 20–22 meters above ground. That single change transforms the loop’s behavior:

• Lower effective takeoff angles become possible (more energy at DX-relevant angles).
• Ground coupling and near-field losses drop (less “the lawn is my attenuator”).
• Pattern changes dramatically: lobes/nulls move, and the antenna stops behaving like a pure “regional blanket.”
• It becomes a different antenna system — in a good way, if the goal is DX visibility.

And yes, at ~0.5λ the loop can also develop multiple lobes. That’s not a bug. That’s the reality of electrically large antennas. But at least you’re now comparing two systems that both have access to useful low-angle radiation, instead of one antenna being sentenced to NVIS duty by height alone.

If you keep the delta loop low, you are not “testing horizontal vs vertical.”
You’re testing “NVIS-ish geometry” vs “DX-ish geometry.” Different mission. Different winner.

So what did the test actually prove?

It proved something valuable — just not what most viewers will repeat in the comments.

The test proves: Vertical + radials in that location beat a low horizontal delta loop in that location during that window. That’s real data. That’s useful. That’s honest.

But it does not prove: “Verticals are always better” or “This loop is worse than that vertical.”

Because the moment height and environment change, you didn’t A/B antennas anymore — you A/B’d two different launch geometries with two different missions.

Vertical vs horizontal polarization in DX: almost meaningless (and here’s why)

This is where the internet loves to cosplay as an RF textbook: “Vertical for DX! Horizontal for local!” — and everyone nods like the ionosphere reads your YouTube comments.

In HF skywave, polarization is not a loyal companion. It’s more like a drunk friend who disappears, comes back wearing someone else’s jacket, and insists nothing happened.

Why? Because the ionosphere is a magnetized plasma. As your wave passes through it, it experiences Faraday rotation — the polarization can rotate, sometimes continuously, depending on conditions. Add multiple hops, scattering, reflections, and the fact that real antennas are rarely “purely” polarized… and by the time your signal arrives, it’s often elliptically polarized and effectively “random-ish.”

By hop two, “vertical vs horizontal” is mostly a bedtime story. The ionosphere put your polarization in a spin cycle.

So if WSPR results show “vertical beats horizontal,” the likely reasons are usually:

• Takeoff angle differences (height and ground interaction),
• Pattern differences (lobes/nulls, directionality),
• Noise pickup differences (verticals often hear more local man-made noise),
• Installation quality (radial field, common-mode control, nearby junk coupling).

Polarization purity is not what your DX signal arrives with. It arrives with… whatever it survived as.

How to fix the test (without turning your garden into a government lab)

If you want to make “antenna A vs antenna B” claims with a straighter face, do one (or more) of these:

• Normalize the geometry: keep feedpoint height and surroundings as similar as possible, and don’t compare “proper vertical site” vs “loop stuffed in a corner.”
• Run the delta loop at multiple heights: low (NVIS-ish), mid (compromise), and as high as feasible — ideally approaching ~0.5λ on the band being tested.
• Be honest about the claim: “This installation outperformed that installation” is still useful — and far more accurate.
• Pair spots by reporter: compare only cases where the same receiving station heard both antennas in near time proximity. Otherwise, reporter geography becomes your hidden variable.
• Separate day/night and regimes: don’t average different propagation modes into one “winner.”
• Control common-mode: add chokes where needed. Because sometimes the “winner” is your coax playing antenna cosplay.

RF.Guru Verdict

Captain Radials deserves credit for trying to remove human bias and use a repeatable framework. WSPR-based testing is disciplined, boring, and therefore dangerously close to truth.

But the single biggest trap remains: placement. Height and environment are not footnotes — they are the plot.

If the vertical “wins” because it’s installed like a real vertical system…
and the delta loop “loses” because it’s trapped near the ground…
then the lesson is not “buy antenna A.”
The lesson is: “install antenna B like you mean it.”

Anyway.
Kettle’s on. ☕
Don’t forget: the most powerful antenna accessory is still… a tall support and fewer nearby metal surprises.

Interested in more quiet RF heresies?
https://listmonk.rf.guru/subscription/form

Questions or experiences to share? Feel free to contact RF.Guru.