I've used a couple different kinds of pogo pins. The fancier one comes in two parts. The inner part is a complete unit, with contained spring and moving pin. The unit slides into the outer cylindrical housing, clicking into place. So they're replaceable.
The cheaper ones were a single part. So if the pin goes bad, you have to desolder w/e the pin connects to in order to replace it.
Ironically, the fancier pins are so well made, I have never worn out or damaged one of them. But it's still nice that you can change out individual pins and/or different spring forces and head shapes. The stiffer springs are quite stiff. You might need that to pierce through oxidation or flux residue. If working with clean ENIG pads, you can use the lowest force springs.
The cheap pins wore out fairly quickly, and they had more wiggle to them, radially.
Fancy pins:
https://www.mouser.com/datasheet/2/193/55-11775.pdfThey're rated for 0.05" center to center. I have never tried to solder them into a pcb like a thru hole part. The housings are about 0.034-0.038" in OD, per the datasheet. I imagine it is just possible with a custom pad shape. What I do is just solder the housings flat on their side to the pcb for a SIL 0.05" pitched interface. If you use two short rectangular pads on either end of the housing, you can flow and adjust one end at a time to straighten the pins and get them parallel. But it takes a steady hand, for sure.
Personally, I think the single point spear-head is the best for flat, clean pads. It has good piercing power and gives the most room for error in jig alignment. If you have any solder lumps on the pads, the multiple point serrated heads will register better on a lump in case of being slightly off but still have good piercing power. Rounded head pins are not that great on flat pads, IMO. I haven't found the point of them.

On the PCB I purchase, I can often see a single sharp divot where the test probes were placed, which I think is from a high force spear point.