Your analysis is exactly opposite from the mathematical truth. You even mention a "narrower frequency distribution" for the eight-deck game, which is correct. But, somehow, you then interpret that to imply greater variance, when, in fact, just the opposite is true. If all the true counts are squeezed into a narrower width, then there is clearly LESS variance. To take it to the absurd, suppose all the true count could ever be was -1, 0, or +1. Results and bet sizes would hardly vary at all! It's precisely because, with fewer decks, there is a wider distribution of true counts that the variance for these games is, obviously, greater.

But, if you would stop trusting your erroneous and faulty intuition and just look at the Chapter 10 charts of BJA3, you would see all of this clearly.

Don