Well
This is all an abuse of statistics.
Now, I'm no maths genius, so please correct the glaring ommisions here.
As pointed out above, the 1 in XX chance of hitting someone means that the essential algorithm is reversed.
ie, if you were to transpose yourself from the lottery example into the satellite example, you would not be a person on the earth, you would be the satellite.
Calculating how likely it is to hit me is beyond my number crunching skills (I'm more of a logic cruncher).
However to take the lottery as a comparison.
There is 1 draw per week, with very long odds. However, _millions_ of people enter, so there is regularly a winner. If millions of satellites were to enter the atmosphere every week, then in all likelihood, someone would be hit each week.
By the same reasoning, if only a single person (and no one else) were to buy a ticket this week, then they would be less likely to win the lottery than be hit by this satellite (according to the above numbers that I can't confirm)
This is how you'd make the two equivalent. Essentially, a single, given lottery ticket winning is the valid thing to compare to a single satellite, not the entire lottery.
You see the results of this in the satellite odds themselves. The 1 in 3200 is the total system involved. The odds to hit someone, anyone. If you were to calculate what the odds were for someone, anyone, to win the lottery. I'd fully expect that they'd be in the same order as these. Maybe a bit lower, given that there is generally a winner each week or two.
Here ends the lesson, corrections sent to my good friend, Mr Don Kiddick or his special friend, Ben Dover.