The Plan

They stopped minting them

When I heard they were not going to mint the penny any more, I started watching for them. And they were already going. None in my change at the store. None in my pocket at the end of the day. Nothing in the tray by the door that used to fill up without anyone putting anything in it.

They had not gone anywhere, of course. A hundred and fourteen billion are still out there, still legal tender, sitting in jars and cup holders and the bottoms of bags. They just stopped moving. And the Treasury, of all people, has been asking us to spend them.

So I wondered how many I could get moving again before the only ones left are in collectors' albums. Not gathered up in a bucket somewhere. Moving. Hand to hand, one person at a time.

That is the experiment. And for that, I need some help.

Worth stating plainly, because it matters to anyone who knows coins: the penny has not been demonetised and has not changed in value. Production for circulation stopped. Minting, stamping. Every cent already made stays legal tender and always will.

Four steps, none of them clever

1. One. Single. Penny. Not a dollar, not a subscription, not a product. A penny is the smallest thing a person can be asked for, and that is the reason this could actually work. You can send more. The ask, and the only requirement to take part, is that one penny.
2. Ask two people. By name. Send a text, share the link, one at a time, sent by you and not from this site. No spam, no AI bot, no Nigerian prince. Two is the floor and not the ceiling.This site has never sent a message to a single person and it never will, because a machine asking on your behalf is worth nothing to anybody. A person asking a person is worth something, and it turns out a penny is a fair price for a real human interaction.
3. Let people with audiences race. There is a challenge for anyone with a following. Take a slot on the board and find out whether your audience actually moves when asked. You receive nothing for it but bragging rights, which is the point: what you get is a number nobody else will give you, and the traffic this page sends back to you. Grow your subs. Tell them you are making a millionaire.
4. Publish the result either way. The branching factor is on the front page and it updates as it happens. If it crashes and burns, the egg is on my face and on my site, for everyone to see.

That is the whole plan. The rest of this page is the arithmetic underneath it, including the parts that are unflattering.

How this is counted

The interesting number here is not the money. It is k, the average number of new participants each participant actually brings. The arithmetic of one-tells-two is famous and nobody has measured the coefficient in the wild. Everything below says exactly how it is computed, what it excludes, and where it is wrong.

Current value: not yet defined from 0 verified participants. At this sample size it means almost nothing. It is published anyway.

The definition

Every participant is a node in a tree. Each has exactly one parent: whoever invited them. k is total verified joins divided by verified participants, excluding the root.

k = (verified participants − 1) / verified participants

For a chain where each person brings on average k < 1 new people, from S seeds, the expected eventual total is S / (1 − k). Below 1 the chain always dies; the only question is how big it gets first. Above 1 it can run away, and even then any individual line still goes extinct with probability greater than zero.

What counts

• A verified participant: someone who passed a bot check and either confirmed an email code or had a payment matched to them.
• Unverified joins exist in the tree so their children are not orphaned, but they are excluded from k and from the public count.
• One person, one node, per campaign. A second invite opened by an existing participant records an open, not a join.
• Self-invites are rejected at the database level.

Where this is biased, honestly

• Immature cohorts drag it down. Someone who joined an hour ago has not had time to invite anyone. Every live k is therefore an underestimate, and it is worst early.
• Attribution is imperfect. Someone who hears about this from a friend but arrives by typing the address is counted as a cold join, not as that friend's child. Real-world propagation is undercounted.
• Ratio, not generational. This is a simple ratio over the whole population, not a per-generation estimate. It is stable and honest but it hides whether propagation is accelerating or decaying.
• Cycle time is not in it. A k of 1.2 that takes a week per generation and a k of 1.2 that takes a day are wildly different outcomes and this number cannot tell them apart. Generations deep is the partial answer.
• Seeds are not children. People recruited directly, offline, before launch inflate the numerator without anyone having propagated anything.

Activation and amplification

On the challenge board, two further numbers:

activation    = direct verified joins / claimed followers
amplification = total verified in subtree / direct verified joins

Follower counts are self-declared and labelled as such. Treat activation as indicative and amplification as the more reliable of the two, because amplification has no self-reported term in it.

Theoretical is not a forecast

The exponential table on the front page is arithmetic: 2³⁰ is 1,073,741,824, which is more people than exist. It describes what would happen if every single person did the thing, once, on schedule. Nothing has ever done that. It is printed because it is the reason the experiment is interesting, not because it is a projection.

Money

Two different things are being measured and they are not the same. A free drop measures whether someone will act when asked. Money measures how much they meant it. The gap between those two is the more interesting result, which is why both are asked for and why they are reported separately rather than blended into one flattering figure.

So the pennies are free and always will be, and real ones are encouraged — not because the free ones do not count, but because a coefficient without an intensity measure only tells you half of what happened. It is also, plainly, a better spectacle. A number that moves is a better story than a number that does not, and there is no point pretending otherwise on a page about honesty.

Giving more buys nothing. Not rank, not position, not priority, not a better place in the tree. A person who sends 2,500 pennies and a person who drops one symbolic penny occupy exactly the same square on the wall. That is deliberate and it is not going to change.

Pennies dropped are free and symbolic and are never rendered as dollars. Money received is counted by hand from payment records, because the payment rails used here send no automated confirmation. Refunds and disputes are subtracted. The two numbers are never added together.

Corrections

If you find an error in any of this, say so and it gets fixed and noted. Raw counts are available on request.