The Drake equation
How many of them are there?
Seven numbers multiplied together. Three of them are now measured; the other four are arguments with error bars dozens of orders of magnitude wide. Move them and watch what that does.
New stars per year in the Milky Way. Measured at roughly one and a half to three.
Kepler answered this one: essentially all of them.
How many worlds per star could in principle support life. An observational question now, and a contested one.
The crux. Estimates in the literature span dozens of orders of magnitude, and we have exactly one example.
Life to minds. Took four billion years here, once, and nobody knows whether that was fast or slow.
Minds to something visible across light years — radio, laser, waste heat, anything.
How long a civilisation stays loud. Ours has managed about a century and is getting quieter.
Detectable civilisations in the galaxy
10.5
About 10.5 civilisations broadcasting in the Milky Way right now, on these assumptions.
Spread evenly through the disc, the nearest would be about 9,069 light years away.
Draws with nobody but us
97.7%
Every factor sampled across its whole declared range instead of pinned to a guess, twenty thousand times. Almost all of them come back empty — which is a fact about how wide those ranges are, not about the sky.
N between 10⁻³² and 10⁻², middle 10⁻¹⁷
Why the silence may need no explanation →What the equation is for
Frank Drake wrote it in 1961 as the agenda for a three-day meeting in Green Bank, West Virginia. It was never a prediction; it was a way of naming which questions stood between us and an answer, so that eleven people in a room could argue about them one at a time. The number it produces has always been less interesting than the list of things you have to decide before you can produce one.
Three of the seven have since been answered. We know roughly how fast the Milky Way makes stars. We know, thanks to Kepler, that planets are not rare — essentially every star has them. And the number of potentially habitable worlds per system is now an observational question being actively narrowed rather than a philosophical one. Those three are marked on the sliders.
The other four are not measurements. How often chemistry becomes life, how often life becomes a mind, how often a mind becomes detectable, and how long it stays that way: for each of those, the honest error bars span dozens of orders of magnitude, and we have exactly one example of every one of them. Multiply four such numbers together and the product inherits all of it.
What the second panel does
Every Drake calculator on the web asks you for seven point estimates and multiplies them. The problem is not that the estimates are bad; it is that the uncertainty is thrown away in the act of picking them. A factor you believe is somewhere between one in ten and one in ten thousand trillion is not well summarised by whichever of those you happened to type.
So the second panel does what Anders Sandberg, Eric Drexler and Toby Ord argued for in 2018: sample each factor across the whole range the literature spans, twenty thousand times, and look at the distribution instead of the product. The result is not a different average. It is a different shape — one with a great deal of its mass below one civilisation, which is to say below us.
This page's version is cruder than the paper's, and the difference is worth stating. Sandberg, Drexler and Ord modelled the rate of abiogenesis per star-year and integrated over the galaxy's whole history and its hundred billion stars; their headline is roughly a one-in-two chance that we are alone in the Milky Way, not the ninety-eight percent this toy reports. Drake's N is a harsher question: it counts civilisations broadcasting at the same time as us, and one and a half new stars a year times a ten-thousand-year lifetime only ever considers a few million of them. Simultaneity filters much harder than existence.
Both numbers on this page are claims about the ranges underneath them, not about the sky. The ranges are declared in the source, one line per factor, and the distribution is seeded so the same page shows everyone the same twenty thousand draws. Change a range and both numbers change — which is the entire lesson, and the reason Drake wrote the equation in the first place. What the two calculations agree on is narrower and more useful than either headline: 'nobody else is out there' sits comfortably inside the uncertainty, and nothing here measures whether it is true.
Questions this page gets
- What does N actually count?
- Civilisations in the Milky Way that are detectable right now, at this moment of cosmic time. Not civilisations that ever existed, and not civilisations that exist somewhere in the universe — the equation is about a coincidence in time as much as a count in space.
- Why are the sliders logarithmic?
- Because the factors are. A linear slider over a range from one in a billion billion to one would spend almost its entire travel above one in a thousand, which is the region nobody is arguing about. Halfway along the abiogenesis slider is one in a thousand trillion, not one in two.
- Which numbers are actually known?
- Three: the star formation rate, the fraction of stars with planets, and — increasingly — the number of potentially habitable worlds per system. The four after them have one example each, and that example is us, which is exactly the observer-selection problem the rest of this site is about.
- Why does the second panel say we are almost certainly alone?
- It says that of twenty thousand draws from these ranges, ninety-eight percent contain fewer than one currently detectable civilisation. That is a statement about the ranges, and the abiogenesis one does nearly all the work: thirty orders of magnitude of honest ignorance about how life starts swamps every other factor. Narrow it and the share collapses. The 2018 paper, with a more careful model, puts the figure nearer one in two.
- Can I share my assumptions?
- Yes. All seven slider positions are in the address bar, so a link shows whoever opens it the same seven numbers you set, and they can move the one they disagree with.