Saying that you are gauranteed to lose a 80/20 hand 20% of the time is just being silly. Uou dont think that over 10 80/20's that you will have lost exactly 20% of them?
Even over 999999999 80/20's there is stil a chance that you will lose ALL of them. Nothing in poker is gauranteed

You don't have to see it to believe it, since it's a mathematical construct. The CLT dictates that a sufficiently large number of identically distributed independent random variables each with finite mean and variance will be approximately normally distributed. How approximate is good enough? Since parametric statistical tests that assume a normal distribution (anova, t tests, etc) are pretty robust, perfect normality isn't a requirement.

Also, "parts of a curve" don't have distributional qualities. A set of data is characterized by some statistical distribution. Some distributions fit data better than others.

Finally, it doesn't matter what actual values are in a normal distribution, since the mean is defined as zero.

An 80/20 hand is defined as one with a probability that you will win 80% of the time. Probability is mathematically equivalent to the frequency of something occurring. Given an infinite number of trials, it is guaranteed that a success will occur 80% of the time and a failure 20% of the time.

No sorry. There is still the possibility that you lose all of them. Correct? Then you cannot guarantee anything. If you ran AA vs QQ 99999999999 times you cannot guarantee that the times the AA will win is the region of 80%. It's simple logic. There is still that chance that you lose ALL of them.
What you are trying to say is that more than likely, a huge possibilty in fact, the aces will win roughly 80% of the time.

The guy with the nl 25 10k in a month challange,
only broke even after his first 17k hands...

This is the most absolute basic of probability concepts. The probability of an event is exactly identical to the frequency of that event happening over an infinite number of trials, by definition. In a small number of trials, say 10, AA might lost to QQ more then 20% of the time. This is because the process is "stochastic." The more times you repeat the trial, the more likely it is to converge on the true frequency. In your example, the probability of AA losing to QQ 99999999999 in a row is (0.2)^11= 0.00000002, or 1 time in 48 million. How do I know this? Because of the math, which is "simple logic."

Here's an experiment you can do yourself that might help convince you that what I say is true. Take a die. If you roll it, the likelihood of getting a "1" is exactly 1 in 6. This isn't TOO far off the likelihood of QQ beating AA (1 in 5). Roll the die 6 times. Write down how many time a "1" comes up. Now keep doing that. The more times you roll the die, the more close to 0.16667 n (which is equal to number of times 1 is rolled divided by number of times any other number is tolled) will be. If math doesn't convince you, perhaps empirical evidence will. If neither of these two approaches is convincing, please join me at my poker table.