I assume you meant to say that your DVM was set on the 20K range and not 20X. If so, your total range of readings is within 6% of each other, which is quite accurate. In this application the resistance is there to suppress the Electro-Magnetic Interference from the high voltage, and the resistance value is not at all critical. The car would work fine with any resistance from 0 to 20,000 ohms, maybe more. For example, resistor wires can have as much as 12,000 ohms per foot.
From memory, those bow tie caps if original are stamped (on the lead entry end?) with their resistance. Pretty sure 9k ones were common, and maybe 5k, possibly others. I reclaimed a set from junk yard scouring over 25yrs ago when they were still reasonably common, and made up a set with same resistance. Still doing sterling duty today.
You can also dismantle and swap the small resistors in the lead end to get a matching set if you have plenty to play with.
I recently discovered my e type should actually have the Champion oval style - oh well
This sort of resistance variation will make no difference in performance.
In the 1960s, a 5% tolerance on resistors was considered to be “close tolerance” and a 10% variation in component values was the norm.
Resistor plug leads (as opposed to plug leads with attached resistors) themselves will vary with length and anything between 5k to 20k ohms isn’t going to appreciably tame a spark that goes into the kV. If that seems a bit odd, remember that the amount of spark energy which is delivered is not remotely constant but will vary greatly as the the dwell time (and therefore coil charging time) varies with rpm between idle and redline revs.