In shorthanded cash games, It's common to see stacks flying across the tables and being quite a few buy-ins down is common, and should also be fine, as long as it's not the only direction in which they're going (that is, that after the aforementioned 100k hands you have a nice positive winrate).
15 sounds way too high to be common, but since you're talking about 30k hand sequences in a 100k hand database , it doesn't sound that bad to me, again. 6-7 buy ins down sounds normal for short handed (NOT normal as in - every few sessions (that would be abysmal) , but, it happens from time to time), so 10-15 ....3 times or so in 100k hands, well, I think that's quite likely.
Once again, as long as you're a solid winning player usually, and after a 100k hands, you should have a nice positive winrate. If not, THEN it's a good time to worry
There are many programs that calculate the likelihood of a certain size downswing. If you write your winrate and your SD (standard deviation) as it appears in your Holdem manager (or pockertracker, or whatever you're using for it), I can provide a more mathematically correct answer.
For now, plugging some normal values into my variance simulator I get this:
bb/100 (in big blinds, not big bets, since the default bb/100 written in HM is in big blinds anyway, and so is the default SD written under the "stakes" report) - 14
SD = 80
results:
In the course of 100,000 hands, that player has a 20% chance of having a 1250 bb downswing (almost 13 buy-ins...),
and a ~17% chance or so to experience a 15k break even period.
Now just change that player's SD in short handed cash games to 100, and now he has a 40% chance to encounter a 15 buy in downswing during the course of a 100,000 hands!!!
So these stats are important to answer your question.
Note also that the number of hands is very important....because without limiting it... heh, well, in an infinitely long number of hands, you'll encounter any finite downswing of any size with probability 100%
Hope that helps with answering your question