Typical row/value for a column col1, can be explained like this ...
sel col1, count(*) cnt from mytable group by col1 ;
now check what is the maximum repeated value for cnt (ie the most recurring count value) that's the typical row/value for col1.
for example if the above query gave the following result set,
col1 cnt 12 3 11 2 5 5 30 3 27 3 22 2 21 3
you can see that cnt=3 has 4 records, cnt=2 has 2 records and cnt=5 has only 1 record. So since cnt=3 has maximum number of records (4), the typical rows per value for col1 is 3.
Note that this can get complicated at times... for example what if there were 4 records with cnt=2 ? ... in this case we would stick to the maximum value of cnt, ie it would be still cnt=3 that would be chosen. (worst case scenario).
You can read more about this in detail in the Database design manual, "maximum and typical value frequencies"... there are also case studies provided with examples...