That's your opinion. It's not what observation suggests.

Incidentally, let's say that at time 0, you have a population of 100. At time 1, 10 have died. Assuming that at time 0 everyone would drink water, which is your own assumption, there is absolutely no correlation between drinking water and dying (I wrote the series down for you so you can think about it). If you're trying to stun the rest of us, try at least to find an example that makes sense

Drinks Dies Lives
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 1 0
1 0 1
1 0 1
1 1 0
1 0 1
1 0 1
1 0 1
1 1 0
1 0 1
1 1 0
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 1 0
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 1 0
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 1 0
1 0 1
1 0 1
1 1 0
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 1 0
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 1 0
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1
1 0 1