Using Karel's SBA 5.5 I runned a billion hands against 6 dks in order to obtain these figures for the dealer's final-hand probabilities.
Here they are:
BJ = 4.7473
21 = 7.2893
20 = 17.9486
In other words, the dealer will be 'super-pat'
29.9852%, a figure who deserves our respect.
Looking at page 161 from Dr.Humble's "The World
Greatest BJ Book", he give us the cumulative figure of 29.77%. The reason of this minor discrepancies can be related to the fact that they were extracted for single deck BJ, maybe,
an inference I'm doing based on his 4.83% dealer's BJ probability.
Lets play a little bit with these figures and paint the following scenery:
6 dks,75% delt out,7 players,11 rounds/shoe.
Could be well AC's weekends or?
Now to perform the calculations we need Bernoulli's binomial distribution of probability
formula, easy to find in every stat's book, and
write down an easy program in your favourite language to put the PC to work.
Here are the results I got for the 29.9852% figure:

K------ P ------ S
0 .0200 .02
1 .0930 .113
2 .2000 .313
3 .2570 .57
4 .2200 .79
5 .1320 .922
6 .0560 .978
7 .0170 .996
8 .0040 .999
9 .0010 1
10 .0000 1
11 .0000 1

K = nr. of times the dealer is 20,21 or BJ
P = probability of occurrence
S = cumm.probabilities

What is the probability of our dealer having 20, 21, or BJ, 9 times out of eleven rounds?
Answer: 0.1% or 1 vs. 999 odds.
And 6 times?
Answer: 5.6% or 1 vs. 17 odds.

So enjoy and don't be scared, our game can be
beaten. That's the most fundamental lesson I've learned from our host.
Thanks again Don!
Regards
Z