I am wondering why probe start ceramic metal halide lamps with an actual starting resistor and starting electrode were not developed.

Sealing the auxiliary electrode into the ends, also why they couldn’t do it with HPS.
A starting aid wrapped around the tube was the closest they got!

Lamp manufacturers could certainly seal multiple conductors to ceramic burners, one even commercialized HPS lamps with multiple feedthroughs -  see picture below of the Tungsram TCF 400:



There are multiple reasons as to why probe-start CMH lamps were never released on the market. The first reason is because it is simply more expensive to add an extra feedthrough for the ignition probe. The added complexity in the burner design results also in a higher risk of premature failure. Moreover, the added ceramic and metal at one end of the arc tube only increases thermal losses there, which translates into a lower lumen efficacy. Finally, by the time CMH lamps were introduced (i.e., 1994), electronic ignitors were cheap standard components in MH circuits, so going back to a probe start configuration did not make sense, especially since this would require a high ballast OCV, certainly a problem in Europe, and a low gas fill pressure, which would impact the lumen maintenance negatively.

Well, passing multiple physical wires through a seal is indeed possible, but all presented were rigidly the same potential wires, so no voltage difference at all.
But is it still possible to pass them so they are able to sustain few 100's V between them, as it is necessary to allow the ignition?
I have quite strong doubts there.

Obviously the costis very clear thing there.

But is it still possible to pass them so they are able to sustain few 100's V between them, as it is necessary to allow the ignition?
I have quite strong doubts there.

Yes, it's possible and this was actually built and tested - I have an experimental HPS lamp with starting probes connected the same way as in a standard HPMV lamp.

That would be interesting, Max. Do you have a photo of it..?

I checked my files and I don't have a picture of it. I'll make a photo next time I get the chance to retrieve it from the Collection. It's an interesting lamp for sure, beside its unusual electrode configuration it also has a wide burner to provide enough space for the more complex end seals.

Tungsram did produce HPS lamps with probe-start design and starting resistor similar to what is described, and I wonder if they were not in fact commercialised because in the GE historic lamps collection there are dozens of these across a wide range of wattages, mainly dating from the early 1970s.  Most of them seemed to have a ring-shaped probe surrounding the central main electrode, and only in arc tubes with external amalgam reservoir at the opposite end.

I think one of the biggest problems would have been amalgam condensation shorting the probe to the main electrode during life, if not run with the probe always at the upper end.

The same problems would plague a CMH lamp, since the salts are conductive at operating temperatures and these lamps usually have the molten salt pool reservoir exactly where the feedthrough would be.  Using a bi-metal switch as on larger lamps would not be a complete solution on HF gear due to the not inconsiderable dielectric constant of the ceramic burner itself at such high temperatures.

Moreover the probe start design is technically inferior  - in particular the USA manufacturers had spent the past years pushing consumers to upgrade to the superior pulse-arc lamps, and it would have been a backward step to offer probe-start CMH.  Nevertheless it was certainly explored, there are quite a few patents which detail such kind of CMH arc tubes. 

I agree, the condensation of MH salts and of mercury is also a major issue in ceramic burners provided with an auxiliary probe. All in all, going for a probe-start design when it comes to CMH lamps is not a good choice, there are too many issues that need to be solved for a very limited benefit in return, i.e. , the removal of the ignitor in the circuit. Besides, Philips already found a way to start CMH lamps on standard probe-start lamp control gears, they replaced the argon fill by a neon-argon Penning mixture (c.f. CDM AllStart lamps).

@James, very interesting information! I did not know that GE was doing such experimental work with HPS lamps as far back as the 1970s as you describe.
You also mentioned GE having a historical lamps collection. Are there any photos of this online or elsewhere? It would be interesting to see some of their HID prototypes and other lamps that might not have been put into production. Thanks!