This rendition of an OSU tsunami shelter prototype shows a large multi-story building on stilts, likely to lessen the impac of the base of a tidal wave crashing through its vicinity. (Source: Oregon State University)

Shake, rattle, roll and splash -- major seismic event could ravage the pacific northwest twice over.

There
have been no shortage of powerful and often, sometimes
catastrophically, deadly earthquakes in the past ten years. Sumatra,
Haiti, Japan, Chile and more have suffered to various degrees from
the results of plate tectonics and the roiling seas of magma far
below the surface of the planet. Though, in the US, California has a
reputation for being earthquake-friendly, it is a far cry from the
only threatened west coast state.

Based on data collected from
the Cascadia Subduction Zone, which lies off the west cost of North
America and runs from northern California up to British Columbia,
Oregon State University marine geologist Chris Goldfinger and team
says the chance of a quake of high magnitude, 8 or better,
is unsettlingly
high in the next fifty years. Using telltale signs of seismic
activity, they have mapped out a time line of major events for the
last 10,000 years. As it turns out, the pacific northwest is about
due for a major earthquake.

According to their findings, the
Cascadia has already gone past the 75% mark as far as a major event
within a generally rhythmic period of time. Over the past 10,000
years, they have found evidence of 41 large events, spaced at roughly
500 year intervals. Should no event occur in the next 50 years, the
chances jump to 85%. There is no doubt, feels Goldfinger, that the
event is coming -- it's just a matter of time.

At present, he
states there is an approximately 37% chance that a magnitude 8 or
greater quake will hit the southern section of the Cascadia
Subduction Zone, which runs from northern California to near Newport,
Oregon, in the next 50 years. Further north, the chance of an event
is less -- 10% to 15% -- but with a better chance of being much
stronger, magnitude 9 or greater.

Not all of the west coast is
oblivious to this sleeping giant. Not only would a sizable off-shore
event cause the standard stand-in-a-doorway building rumbling action,
it would most certainly create a powerful tsunami in its wake. The
last recorded high magnitude quake from the Cascadia was in 1700.
Though no records exist from the Americas, Japanese historians
recorded the ocean-traversing tsunami that reached their shore,
crashing down at 30 or more feet in height.

The town of Cannon
Beach, Oregon, is working with engineers from OSU to create an
earthquake
and tsunami shelter for its residents using advanced
construction techniques and an eye for vertical space to stand above
the wave. If completed, it may be the first tsunami shelter built
outside of Japan.

quote: But if a coin flip turned up heads 9 times, then you have a higher probability that the next try will come up tails. But your chance stays independant.

Either you are using incorrect terminology, or you are falling victim to the gambler's fallacy.

For independent events like coin flips (assuming a balanced coin), the probability for tails is always 50%. Doesn't matter if you got 9 heads or 9999 heads. The probability for tails was, is, and always will be 50%.

If you are talking about estimated probability, then the opposite of what you say is true. You take the average of the previous flips to work out a likelihood of coin being unequally weighted. For 9 heads, that works out to a 99.8% probability that the coin is two-headed. Thus your best estimate for the probability of the next flip being a head is 99.8%, with only a 0.2% probability of it coming up tails.

For dependent events like earthquakes, the probability increases the longer you go without an earthquake. But that's because of stress building up in the earth's crust. It has nothing to do with how probability works. Probability just allows you to statistically model the frequency of earthquakes, and based on that estimate the probability given how much stress has built up (how many years it's been since a big quake).

BTW, I'm pretty sure you're falling for the gambler's fallacy. The reason coin tosses trend towards 50% as you approach infinite tosses isn't because getting 9 heads means there's a higher chance for a tail. It's because the subsequent tosses will tend to be 50% and they swamp out the imbalance of the 9 heads by sheer sample size alone.

Right - that's where the Central Limit Theorem comes in.

So everybody is clear, the Central Limit Theorem doesn't describe the probability of a single coin flip. It describes how, as the number of flips approaches inifinity (gets bigger), the overall sample mean of those flips converges to within an epsilon neighborhood (keeps getting closer to) of the true mean/probability.

"Nowadays, security guys break the Mac every single day. Every single day, they come out with a total exploit, your machine can be taken over totally. I dare anybody to do that once a month on the Windows machine." -- Bill Gates