fomo trading, explained.

fomo is a social trading app. you log in, you follow traders, and when a trader you follow opens or closes a position, your phone buzzes. the leaderboard ranks everybody on realised profit and loss, publicly, in dollars, forever. that is the entire product, and it is a much stranger machine than it looks.

most trading apps sell you tools. fomo sells you other people. the consequence is that the price action inside it does not behave like ordinary market price action, because a meaningful share of the buying pressure on any given token is triggered by a notification rather than by an opinion. if you are trading on fomo, or thinking about copying somebody on it, that distinction is the only thing worth understanding first.

what fomo actually is

strip the social layer away and fomo is a wallet with a feed bolted to it. three parts matter:

• a custodyless smart wallet. you sign in with a normal login and an erc-4337 smart account is created for you behind the scenes. no seed phrase ceremony, no browser extension. this is why the app onboards people who would never have made it through a metamask install, and it is why the average participant is newer than the average dex trader.
• a public, permanent leaderboard. realised pnl, all-time and rolling windows, with follower counts next to it. it is not a list of calls or predictions. it is a list of filled trades, which makes it far harder to fake than crypto twitter and far more useful as a starting point.
• a notification bus. follow someone and their fills push to you. this is the part that changes the physics.

execution is routed rather than fixed to one venue: orders go through relay's router on robinhood chain by default, with bsc, ethereum, base, arbitrum, polygon and solana in the registry. for a user it is invisible. for anyone trying to reason about slippage it matters, because a route is a route and the size in front of you is real.

the leaderboard is the product

a public pnl leaderboard does something no marketing team could do on purpose: it creates a small number of people with enormous, verifiable, quantified credibility, and then it hands each of them a broadcast channel.

the naive reading of this table is "these are the best traders, copy the top one." the useful reading is that the two columns measure completely different things, and only one of them predicts what happens to a token after they touch it.

the cascade, not the alpha

here is the mechanic that makes fomo different from a brokerage. when a leader with three hundred thousand followers opens a position in a token they have never held before, two things happen in sequence. first their order fills. then, within seconds, a notification lands on hundreds of thousands of phones saying that it did.

the second event is much larger than the first.

we measured this on one wallet's first buys — the very first purchase of a token that trader had never previously held — over roughly a month. the median move was +39% one minute after their fill and +43% at fifteen minutes, followed by a steep decay over the rest of the hour.

the tempting explanation is that the trader is simply good: they find things early and the market agrees with them. the data points the other way. when we split the sample by how much of that minute's volume the trader's own order represented, the result inverted the intuition:

when the leader's order was less than 40% of the minute's volume, the token moved +53%. when it was 40% or more, it moved +31%.

if the move were the trader's own price impact, more of their volume would mean a bigger move. it is the reverse. the moves are biggest exactly when somebody else supplies the volume — and the somebody else is their followers, arriving through the notification.

if the cascade thesis holds, one practical consequence falls out of it immediately: the variable that predicts a move is the follower count, not the pnl rank. a trader with 500,000 followers and mediocre returns triggers a bigger wave than a trader with 40,000 followers and spectacular ones. that is uncomfortable, and it is what the numbers say.

copying is the native behaviour of the app

fomo does not have to encourage copying. the notification is a copy prompt. the leaderboard is a shortlist. the follow button is a subscription to somebody else's decisions. every part of the interface is pointed at the same action, so the overwhelming majority of activity in the app is people doing a slower, worse version of the same trade a leader just made.

slower and worse are the operative words. the sequence for a follower is: notification fires, phone lights up, you notice, you unlock, you open the app, you read what happened, you decide, you size it, you confirm. call it thirty seconds if you are holding the phone and paying attention. call it eight hours if you were asleep, which, given the schedule these tokens keep, you usually were.

and the price does not wait. the median first-buy move tops out at around fifteen minutes and gives most of it back within the hour. a follower who arrives at minute thirty is not early to the move. they are the move, in its second half, buying from the people who were early.

why doing it by hand does not work

three failure modes account for most of the money lost by people who are, on paper, following good traders.

the 4am problem

these tokens do not respect timezones. the alert that mattered fired while your phone was face down on a nightstand. you read it at breakfast, forty percent later, and file it under bad luck. it was not bad luck; it was a scheduling conflict with sleep, and it will recur every week for as long as you do this manually.

the catch-up trade

you see the move late and buy anyway, because the trader is still holding and the chart is still green. this is the single most expensive habit in copy trading and it is almost never described honestly. you are not copying their trade. you are taking the other side of it, at a price they helped create.

the missing exit

the entry is the exciting half and the exit is the half that determines the outcome. most people who copy manually catch some entries and almost no exits, because an exit notification arrives with no urgency attached and nothing on the screen is flashing. copying entries without exits is not a strategy, it is a way of accumulating other people's bags.

four rules that survive contact with the app

1. sort by followers before you sort by pnl. if the cascade thesis is right, audience size is the closest thing to a leading indicator the leaderboard offers. profit rank tells you about the past; follower count tells you how much force the next notification carries.
2. treat first buys differently from adds. the effect we measured lives in a token the trader has never held before. a trader adding to a position they have held for two weeks is not the same event and does not carry the same wave.
3. decide the exit rule before the entry. if the median move decays inside an hour, "i will watch it" is not a plan, it is a hope with a screen attached. either the exit is mechanical, or it is copied, or you should assume you will miss it.
4. size against your own account, never theirs. a leader opening $49,900 is expressing conviction with a risk budget that has nothing to do with yours. conviction is cheap for them. the same position at your size can be your whole month.

and the rule underneath the other four, from a trader sitting near the top of the follower column, who put it better than we could:

"Never blindly follow anyone's trades - including mine." — PoorGoat

we left him on our leaderboard, and we think you should take him seriously. copying is a way of executing a decision faster. it is not a way of avoiding having made one.

where to go next

if you want the data behind the cascade claim, including the method, the decay curve and the things that would prove it wrong, read what happens to a token when a fomo whale buys it first. if you have already decided who you want to follow and want the mechanics, read how to copy a fomo trader. if you are still choosing, which fomo traders are worth copying lays out the filters that actually correlate with something.