Remez is much more interesting when you can find a feature that has a low probability of happening by chance. Before asking for the significance, you should look at the probability that this would happen in a random similar text. In other words, we want to define what you find to be "interesting", generate random data similar to what we find in the Chumash, and see if we find similar results. If we do, the remez becomes less compelling, because we would have expected something similar to happen in any case.
How can we generate lists of numbers similar to these? To start with, we don't have to worry about the last digit, which is always 0. The second to last digit is also 0 for everybody except Gad, so let's exclude that as well. Generating lists of numbers for the first digit would be very hard because it would have to involve data about the population growth and size, which was far from natural due to sextuplets, What happened to the bechorim (first-borns)?, and other things. In any case they're not so relevant to this question, because Yehuda, who had unusually large population, still fell short of 80,000 people.
We can approximate that in populations of people of around this size, the second and third digits would be randomly distributed. There are 24 numbers in these positions - two from each shevet. So, we can generate lists of 24 random digits and see how likely it is that there's a digit that doesn't appear.
Running this code in python:
import numpy as np
from collections import Counter
from scipy.stats import poisson
simple = True
def probability(ints):
ctr = Counter(ints)
counts = [ctr[_] for _ in range(10)]
if simple:
return 0 in counts
else:
return np.prod([poisson(2.4).pmf(_) for _ in counts])
for i in range(20):
print(probability(np.random.randint(0, 10, size=24)))
print()
print(probability([6,5,9,3,5,6,4,6,4,4,7,4,0,5,2,2,5,4,2,7,1,5,3,4]))
I first simulate 20 lists of 24 numbers and print whether or not there's a digit that doesn't appear. I get about half True and half False. Then I print whether there's a digit that doesn't appear in the real numbers, and of course I get True.
For a more complicated analysis, you can set simple = False. Then it will print the probability of getting the counts, using a Poisson distribution with a mean of 2.4 for each digit. The probability for the real numbers is about 4e-09, which is on the low end but within the range I find for the simulated numbers.
Obviously, Hashem gave us these numbers for a reason, but I don't think the absence of 8 is that reason.
