Metas and narratives: where attention goes

This page explains why attention moves memecoins, and how to tell a coin's story from evidence. By the end, you can test a story, trace where its attention started, read a trending tab, and write your exit rule before the story changes.

The Trenches lessons named here sit on the Crypto road, which opens in order, starting with the free Memecoin starter.

What are narratives and metas?

Memecoins are also called attention markets: money goes where attention flows. A memecoin's price rides on attention. Owning one gives you no claim on the picture, the backstory or the website.

When buyers arrive faster than sellers, the price climbs. Each buy takes coins out of the pool, the pair of reserves that trades swap against, so the coins left cost more. Why a coin has a price, and why thin coins slip shows how.

A narrative is the story told about why a coin matters: a theme, a use or a community. A meta is a narrative that a whole group of coins shares for a while, like the meta in a game. Say many posts this week are about "robot pets" coins: that shared story is a meta. Both gather attention and can change quickly, but neither is evidence that the story is true, or a promise that the price will hold.

Easy to mix up: attention and buying. A view puts no money anywhere, and a mention is not a trade. Attention can change what people choose, but people can also wait, leave or sell. The free Memecoin starter teaches this in its first unit, in Attention, big holders and market value.

How do you test a coin's story?

Take the story of a made-up coin in Narratives as research claims. It says the coin powers thousands of AI bots, its developer ships updates every day, the team's tokens are locked, and it is the next big AI coin. That is four claims, none checked yet. Repeating a claim does not check it, and unchecked is not the same as false.

Rewrite each part as a plain claim: one sentence with a number, a source and a time window. "Thousands of bots" becomes: at least 1,000 different addresses used the project's software last week. "Ships every day" becomes updates on most days in the project's public code history. "Locked" becomes a lock you can see on the chain, the public record of every transaction.

A claim is testable when a stated result would count against it. In the lesson, the plain claim needs 1,000 addresses, and the record shows 37: that claim is contradicted. One failed claim does not sink the rest, so each part gets its own check.

The code history shows updates on 5 of the last 7 days, so the "most days" claim is supported. "The next big AI coin" names no measure and no date, so nothing today counts either way. It is untestable, which is not the same as false. Give each claim its own verdict: supported, contradicted or untestable.

Where did the attention start?

A citation is the part of a post that names where its wording came from. Following citations back, one link at a time, is a trace. It ends at the original statement, the earliest wording, which cites nothing earlier. It comes from the party the claim is about, or from whoever gathered the facts.

Say a big account posts that a coin's team signed a game deal, citing a smaller account. That account cites the team's own post, which cites nothing earlier: the original statement. The big account, however large its audience, only passed it on (Find the original claim).

The earliest post you found may not be the first, only the earliest in the places and hours you searched. A bot that reposts new coins can be the earliest find without being the origin. Places you did not search are unchecked, not empty.

Some traces cannot be finished. A screenshot from a private group cannot be opened at its source, so its claim stays untraced: not shown true, not shown false. A deleted post is a broken link, so mark the gap and leave it empty. One lesson has you write a coin's source timeline, gaps included (Where attention starts).

What does a trending tab really rank?

A trending tab is a ranked list under an app's own rule, not a safety rating or a forecast. The rule names what it counts, over which window of time, and across which pools. One app might count swaps, trades against a pool, in the last hour, across the pools it covers. If the app changes its rule or its pools, ranks move with no change in demand.

A short window favors a burst. In Trending tabs and rankings, a practice coin leads the hour with 900 swaps, 700 of them from three wallets. Over the whole day, it ranks last of the three coins. A swap count is not volume, the value that changed hands.

Some rows were never ranked. A pinned row marked "Boosted" in small type is paid placement: it records who paid, not a measurement of the coin. That small label is the disclosure, easy to miss while you scan.

Read a caller, someone who posts a coin to a group, the same way: a big group is an audience, not research. Ask who gains if readers act: a caller who bought before posting gains when the price rises, paid or not. An "#ad" line is evidence neither for the claim nor against it. Two lessons take calls apart (Who is calling it, Promotion, groups and paid attention).

When attention fades, can you still get out?

Watch three things over the same windows and keep them apart. Attention: mentions on one tracked site. Trading: volume, and unique buyers counted as addresses, not people. Liquidity: the pool's reserve, what a seller sells into.

In Attention decay, mentions fall first, and trading falls too, a day or two later. The reserve falls least, because only swaps, deposits and withdrawals move it. A reserve is not a sign that buyers are coming. So give each of the three its own written line and the action it triggers.

For a holder, what counts is remaining exit capacity: what a full sale would return from the pools that exist now. In the same lesson, a wallet screen values your tokens at the last price: 9,000 USDC, a coin built to track the dollar. Selling them all into the pool returns 7,200 USDC, because your own sale pushes the price down as it fills. Nothing on the chain says the price must bounce back (Inactive projects and changing demand).

The coin with the most posts is not a reason to buy. In A popular narrative candidate, most posts about the leader are reposts of three accounts.

Checked against five written conditions, the leader passes only one: nobody can create more of it. It fails on holders, pool size, exit and real use. The third coin by posts passes all five, so it is eligible under the check. Eligible does not mean buy.

Waiting has a cost. An exit rule is a condition for selling that you write down in advance. In one lesson, yours fires while a full sale would still get 3,000 USDC (Attention weakens during a position).

Then other holders sell into the same pool. A holder who waited an hour gets a quote, the estimate shown before you trade, of 400 USDC for the same sale.

Write your exit rule before the story changes

A decision rule says in advance what you will do when a fact you can check changes, including doing nothing. A position is the coins you hold. The part of the rule that ends it is the invalidation condition, the exit rule that fires when your reason for holding is gone. You write it before the trade.

Say you hold a coin because players spend it in a game. Then your condition counts spending addresses each week, not posts and not price.

In A changed narrative and a decision rule, the line is fewer than 100 spending addresses, two weeks in a row. Posts fall and the story drifts to a new theme, but neither is in the rule, and a new story is a new claim. The count reads 240, then 90, then 60, so the condition is met in week three. By then the earlier price is gone.

A friend with a price rule keeps holding, because the price has not fallen to their line. You both followed your rules. Judging either rule by later prices is hindsight bias: judging a past choice as if you had known what came next.

When a coin gets busy, your written limits stay the same. In A crowded attention cycle, a coin tops a trending tab and swaps hit its pool about once a second. A quote goes stale in seconds, and a failed swap still pays its network fee, the charge for processing it.

Slippage is the gap between the price you expected and the price you get. Your slippage tolerance sets how far below its estimate a swap may fill. Widening it breaks a written limit, and so does a bigger order. The lesson ends with no trade, a valid result.

Sizing and risk on thin coins: who buys your exit picks up from there.