As the world waited for today's close flyby of the asteroid 2012 DA14, a small asteroid broke up over Chelyabinsk Russia this morning, blasting in city windows with its sonic boom and injuring hundreds. Meanwhile, NASA's Near-Earth Object Program has revised the odds of asteroid 2011 AG5 hitting the Earth in 2040 to 1 in 500. Just how serious does the risk have to be, before we should do something about it?

The answer is not as straightforward as you might think.

When you assess the potential risks posed by a nearby asteroid, you need to consider a whole host of factors, including the probability of impact (which is complicated by our constantly changing estimates), the size of the asteroid (which determines the scope of damage), and the estimated date of collision (the farther away in time, the less we have to worry right now).

But there's yet another factor, one that's a bit tougher to quantify: How do we assess a catastrophic risk based on sheer probability? And more to the point, given that we will soon have the means to do something about asteroids, under what circumstances should we not act?

The Torino Impact Hazard Scale

Thankfully, NASA's NEO program is in place to help us evaluate these sorts of threats and alert us to when we might need to respond. This group keeps a regularly updated list of all current threats spanning the next one hundred years.

To make their assessments, NASA uses a system called the Torino Impact Hazard Scale. Each asteroid is assigned a number from zero to 10, where zero indicates a negligibly small chance of collision with Earth (or that the object will burn-up on entry), and a ten represents an inevitable impact from a large asteroid that poses a major threat to our survival. When NASA scientists assign a numerical value to an asteroid, they look at more than just sheer probability. They also consider the kinetic energy of the possible collision (expressed in megatons of TNT), and they only evaluate threats that are less than 100 years away.

Back in 2005, NASA had to change the description of Level 1 asteroids from "events meriting careful monitoring" to "normal". They felt that the press was exaggerating the risks posed by a Level 1, so they changed the language. It's worth noting, however, that a "normal" Level 1 object could hit us — but the odds are ridiculously low. So NASA believes these impacts are so unlikely, we shouldn't worry about them.

It isn't until an object gets assigned a Level 3 that other astronomers get alerted to a potential risk. A threat of this sort has a 1% or greater chance of a collision that's capable of localized destruction. Level 3s will demand the attention of public officials if the encounter is less than a decade away. Level 4s are essentially the same, but the potential damage caused by the NEO is greater.

Once we get to Level 5 the category switches from "needs attention" to "threatening." Impacts from these NEOs are still not a certainty, but they would give us pause for thought. A Level 5 would inflict regional damage, while Level 6 and 7 are global in scale. At this point, NASA would start to alert the government that contingency planning may be warranted depending on the timescales involved.

At Level 8 and above, things are starting to look very grim. These threats have been assessed as "certain collisions". A Level 8 is a once per every 50 years event, likely on a local scale, whereas a Level 9 is a once per 10,000 to 100,000 year event, that would cause unprecedented multi-regional devastation. And a Level 10 is the Big One, the kind of impact that happens every 100,000 years or so – a collision that would likely mean the end of life on Earth as we know it.

To date, no object has ever been rated above level 4. Of the thousands of asteroids detected, there are only two that still have a rating above zero on the Torino scale — and they're both at Level 1. Those are 2011 AG5 (year range 2040-2047, impact probability of 1 in 500) and 2007 VK184 (year range 2048-2057, impact probability of 1 in 3,030). 99942 Apophis is no longer considered a threat, and has been downgraded to a zero on the scale.

Palermo Technical Impact Hazard Scale

The Torino scale is primarily meant to help the public and the media to get an uncomplicated and quick sense of what's going on. The specialists, however, use the Palermo Technical Impact Scale to quantify the threats in more detail.

The Palermo Scale looks at all NEOs in the Earth's vicinity and prioritizes them according to the degree that scientists feel they deserve our attention (i.e. the call for frequent observations and analysis). And like the Torino Scale, it looks at an object's potential impact energy and estimated date of collision. Comets or asteroids that pose no threat are given a negative value, while the more dangerous ones are assigned a number greater than zero.

The problem of probability

As already noted, AG5 has a 1 in 500 chance of hitting the Earth in 2040, a threat NASA describes as "slight". If it were to hit us, it would unleash 100 megatons of energy and devastate a region 100 miles wide. So that means it'd likely be promoted to Level 8 status, if we determine that an impact is inevitable.

So, what does a 1 in 500 chance actually mean? We can look at this figure a number of different ways.

AG5 has a 0.2% chance of hitting the Earth, or a 99.8% chance of missing. Are you a cup-half-full kind of person, or cup-half-empty?

Or consider this: In 2011, computer simulations gave the St. Louis Cardinals 1 in 500 odds to make the playoffs. And not only did they make it to the post-season, they won the World Series as well.

In other words, it's very far from impossible that the Earth won't get smuckered by this thing in 2040. A case can be made that, given the potential damage that AG5 could wreak, we should very seriously think about doing something about it. Yes, there's a chance that it could land in an uninhabited part of the world, but if it were to hit a populated area, the damage would be unspeakable.

To act or not to act?

A question that emerges at this point is, given the extreme damage that can be caused by NEOs, and assuming that we will soon have the technology to steer an NEO away, at what point do we dismiss a risk and choose not to act? Is it 1 in 1,000? How about 1 in 10,000? If we're talking about the potential for total human extinction, can the odds ever be slim enough?

Perhaps a good rule of thumb to follow would be that, if it's logistically and economically viable to remove a significant threat, we should do it — regardless of the odds.