Randomness is omnipresent. It's everywhere in your life and in your trading.
On any given trade, it's hard to determine if I had good or bad luck, if I have any skill, or if I'm in the right place at the right time.
But once you have monthly returns from over several years of trading you can begin to run statistical analysis to get a better idea if you are good or lucky.
If nothing else, the concept of randomness keeps me grounded or sober, in that "I am powerless over the markets."
A great place to learn about the role of randomness in your life and trading is Nassim Taleb's Fooled by Randomness. I recently listened to the audiobook version after having read the hardcover book 6 times. I found that I (think I) heard things for the first time.
Monte Carlo simulations show you how starting dates can alter your performance. If your first trade was on September 1 and that path led you to a 20% drawdown, you might find that someone trading the same system but starting two months later was up 25% by EOY, while you finished at breakeven.
FYI - if you are working on a trading system, make sure that you include data from companies that went out of business, were taken over or merged, or instruments that were delisted, else you have only a list of survivors.
For example, if your system would have had you long Enron, you'd like to know how that trade affected your P&L, as well as that of CMGI or any of the other dot.com stocks that blew up circa '99 and '00.
If you would have traded Pork Bellies back in the day, you should run that data through your simulator also although the contract has been delisted from the CME in mid 2011 because of lack of interest and trading volume.
Doing this will at least simulate the "worst case scenario" in your backtesting.