This weekend, a 3.3-magnitude earthquake rattled San Francisco ever so slightly. The small quake, like so many before it, passed, and San Franciscans went back to conveniently ignoring their seismic reality. Magnitude 3.3 earthquakes are clearly no big deal, and the city survived a 6.9-magnitude earthquake in 1989 mostly fine---how how much bigger will the Big One, at 8.0, be than 1989?

Ten times! As smarty-pants among you who understand logarithms may be thinking. But...that's wrong. On the current logarithmic earthquake scale, a whole number increase, like from 7.0 to 8.0, actually means a 32-fold increase in earthquake energy. Even if you can mentally do that math---and feel smug doing it---the logarithmic scale for earthquakes is terrible for intuitively communicating risk. “It’s arbitrary,” says Lucy Jones, a seismologist with the US Geological Survey. “I’ve never particularly liked it.”

Just how arbitrary? Oh, let us count the ways.

First, there’s the matter of star light, which has no relevance to earthquakes except that Charles Richter was once an amateur astronomer. When Richter and Beno Gutenberg were developing what would become the Richter scale in 1935, they took inspiration from magnitude, the logarithmic measure of the brightness of stars.They defined earthquake magnitude as the logarithm of shaking amplitude recorded on a particular seismograph in southern California.

Now, logarithms might make sense for stars a million, billion, or gazillion miles away whose brightnesses varied widely---but back on Earth, the rationale is shakier. Understanding the severity of earthquakes is important for millions of people, and the logarithmic scale is hard to grok: 8 seems only marginally larger than 6, but on our earthquake logarithmic scale, it's a 1000-fold difference in intensity. Seismologists have to unpack it every time they use it with non-experts. “We just undo the logarithm when we try to tell people,” says Thomas Heaton, a seismologist at Caltech.

A better way to measure earthquakes does exist---at least among scientists. That would be seismic moment, equal to (take a breath) the area of rupture along a fault multiplied by the average displacement multiplied by the rigidity of the earth---which boils down to the amount of energy released in a quake. The Richter scale uses surface shaking amplitude as a proxy of energy, but seismologists can now get at energy more directly and accurately. The moment of the largest earthquake ever recorded came to 2.5 × 10^{23} joules.

A big number but meaningless without context, right? (That biggest earthquake ever was a 9.6 in Chile in 1960.) Seismologists now have a tortured formula^{1} (below) to convert seismic moment (M_{o}) to the familiar old logarithmic magnitude scale (M). That gets us the aptly named moment magnitude scale, which supplanted the Richter scale in popular use in the 1970s. The Richter scale may be obsolete now, but its logarithm definition of magnitude remains. "Through the years, seismologists have tried to be consistent,” says Heaton. “It’s been confusing ever since.”