Nearly every day of every year, even in tornado-prone areas, tornadoes don't strike. Of the 43,856 days since April 25, 1893, on just 18 have tornadoes struck Moore, Oklahoma. On just five of those days — 0.01 percent of the time — did anyone die. On those five days, though, 102 people were killed and over a thousand were injured.

If you live in Moore, 99.99 percent of the time, there's no threat that you'll be killed by a tornado. How much is it worth to protect against that 0.01? The question, as it so often does, becomes: How much do we spend to save a life?

You cannot prevent people from dying during tornadoes. There will always be people who don't heed warnings. There will always be people who aren't able to make it to shelter in time. Beyond covering a city with an impenetrable dome, tornadoes will hit and people will die.

You can, however, significantly reduce how many people die. It's possible we've already had some luck in doing so: according to the Washington Post, between 1950 and 1980, ten years had storms that resulted in 90 cumulative deaths. Between 1980 and 2012, there were five. There are a lot of variables (number of storms, for example), but it's the right trend.

The first step to preventing deaths in tornadoes is better prediction. The National Oceanic and Aeronautic Administration explains the history of forecasting tornadoes, which largely center on improved radar. Much of the forecasting comes from the National Weather Service, which is good and bad. It's good, in the sense that it's centralized and has access to the full resources of the US government. It's bad in the sense that those resources are subject to change, depending on how much money Congress wants to spend on it. One side effect of the sequester is that the NOAA has to furlough forecasters, even as tornado season continues.

If funding can be found, the Capital Weather Gang notes that there's still room for substantial improvement.

“There IS the potential for major forecasting advances in the period from 1 to roughly 6 hours before the storm, if we can run models with enough resolution and can get enough information to describe the initial 3D atmosphere with lots of detail,” [University of Washington atmospheric science professor Cliff] Mass writes in his blog.

The advantage of earlier prediction, of course, is to reduce the number of people caught in locations where they don't have access to shelter.

That's the second step: Creating as many secure places in as many buildings as possible. Which, again, comes down to money. USA Today created a graphic showing the different types of shelters, ranging from an extensive, below-ground facility to an in-house safe room, of the Jodie Foster-variety. Prices for these differ substantially. There are also shelters that can be buried in the ground, offering similar protection. (They run from $3,000 to $8,000.)

In a report about Moore, Bloomberg reports that neither of the two damaged schools had any sort of safe facility.

Neither Plaza Towers nor Briarwood Elementary School had safe rooms, chambers made of steel and other materials that are bolted to the ground or installed underground, said Albert Ashwood, director of the Oklahoma Department of Emergency Management. It would have cost between $600,000 and $1 million to retrofit each with the shelters, he said in an interview.

“We have funded over 100 school safe rooms. The two hit were not among them,” Ashwood said. “We are checking as to how many were applied for in Moore and how many were applied for in Oklahoma City.”

One resident who lives near Plaza Towers School, where at least seven children died, thought that the problem might be priorities.

“If they can afford a $5 million football stadium, they can afford a safe room,” said John Lemmon, 67, who lives near Plaza Towers school. “They should have done it right after they had the last big one.”

One difference is that there's a football game every Friday. There's a deadly tornado every 24 years. Because they hit so rarely, standard procedure in a storm, those who grew up in tornado-prone areas know, is to stand in a doorway or along a reinforced hallway wall. Windows should be avoided, since many injuries result from broken glass rather than collapsed houses. At AgapeLand Learning Center, teachers did exactly what should be done: huddling with the children in a bathroom, away from windows. That the storm actually hit and destroyed the building was an exceptional bit of bad luck. That none of the kids died, an exceptional bit of good luck.

Disaster prevention is the economics of luck. A shelter at AgapeLand would have ensured the same outcome, but may have cost tens of thousands of dollars. Is the investment worth it for every facility that didn't get hit? Is the investment worth it at Briarwood Elementary, the construction of which helped children evacuate easily? Right now, we're in the period during which it seems the answer is, yes, every saved life is worth it. As time passes, that inclination will likely fade.

Particularly for penny-conscious elected officials. Which may be why the mayor of Moore is today suggesting that he'll push for a law requiring shelters in homes in the city, as CNN reports. According to a FAQ at the city's website, the city participated in a reimbursement program for the installation of safe rooms from 1999 to 2003. The last few years, that program hasn't existed in Moore.

During a press conference, Mayor Glenn Lewis noted that his call for shelters would still be open for negotiation. "We don't want to be so expensive that our homes are not affordable." Oklahoma's governor, Mary Fallin, echoed that concern.

[The state] isn't going to require people to do anything, but if someone encourages them to do that, we're certainly going to go along with that. … [We'll] be announcing something later today that will allow for safe room donations for anyone who wants to help.

We want to save every life we can. And if you can chip in, that would be great.