Exit Alerts: A Browser Rule That Tells You When a Tracked Wallet Sells
Point a rule at a wallet, keep a tab open, and the sells it prints on the sharp tape reach you as a browser notification instead of as a price move you did not understand.
Why do prediction market tailers get left holding the bag?
Following smart money into a trade is the easy half. You watch a sharp trader load up on YES at 38 cents, you size in behind them, and the position moves your way. The hard half is the exit. Entries are loud and social. People screenshot them, talk about them, and copy them within minutes. Exits are quiet. The same trader who advertised the entry rarely announces that they just sold, and by the time the price has drifted back you are the one still marked to a position the original conviction has already left.
This is the single biggest fear for anyone copy trading prediction markets. The thesis you borrowed was never really yours. You inherited someone else's read, which means you also inherited their timing, and timing is where most of the edge in prediction markets actually lives. A sharp who buys at 38 and sells at 61 made a clean trade. A tailer who buys at 41, holds through the round trip because nobody told them the sharp was gone, and watches it settle back to 44 turned the same idea into a loss.
On Polymarket the raw material for solving this is public: sells settle on-chain like everything else, so a wallet reducing a position leaves a record anyone can read. The problem is that nobody reads a chain by hand across dozens of markets, and the exit is precisely the event you are least likely to be watching for. On Kalshi the raw material does not exist at the trader level at all — the exchange publishes no wallet identities, so there is nothing to attach an exit rule to and this guide is about Polymarket wallets throughout.
An exit rule closes as much of that gap as an honest tool can. You name a wallet, and when a sell from that address shows up on the sharp tape, a browser notification tells you so, on the same clock as the entries.
What is an exit rule and how does it work?
On the Alerts page at /alerts, 'A wallet exits' is one of six rule types. You paste the wallet address you want watched and give it a label, and the rule is saved to your browser. From then on, while a WhaleTracks tab is open and visible, the alert engine reads the live sharp feed every fifteen seconds and fires when a row on that feed belongs to your address and is a sell. The notification names the trader and the market; clicking it opens the market.
Two mechanics are worth knowing because they shape what you actually experience. The engine primes itself on its first evaluation after a page load, recording what is already on the tape without notifying, so re-opening the site does not dump a stack of stale exits on you. And every fire carries a stable identity, so the same sell never notifies twice no matter how many ticks it stays on the feed. Triggered alerts are also written to a log on the page — the last hundred and fifty — so a notification you missed while looking at another window is still there to read.
Delivery is a browser notification and nothing else. There is no push server behind Alerts, no phone-native push, no mail, and no chat relay of any kind — those channels are not disabled, they do not exist in this product. That has a hard practical consequence: rules evaluate only while a WhaleTracks tab is open and in the foreground, and a backgrounded or closed tab is a paused rule. If you want an exit rule watching an afternoon, the tab has to be part of the setup.
It helps to remember what a sell means here. A share pays out one dollar if the outcome happens and zero if it does not, so a large holder reducing a position is a direct statement about how much longer they expect their read to hold. That is the information the rule is trying to get in front of you while it is still worth something.
What counts as an exit here, and what the rule cannot see?
The rule fires on one specific thing: a sell print by that address on the live sharp feed. That feed is built from the recent trade tape of the wallets WhaleTracks tracks, so the honest description is that you are watching sells as they appear on that tape, not auditing a wallet's net position against the whole chain. If a wallet you follow is not among the tracked sharps whose tape the feed is built from, its sells will not appear and your rule will sit quiet.
It also means several genuine ways of getting out are invisible to it. Redeeming a resolved position, transferring shares to another address, or merging a matching YES and NO pair back into collateral all reduce exposure without printing a sell on the tape, and the rule will not fire on any of them. So will an unwind executed entirely off the window of tape the feed carries. The rule catches the common case — a wallet selling into the book — and states plainly that it is not a complete reconstruction of anyone's net position.
Kalshi is a harder boundary, not a softer one. That exchange exposes no trader identities at all, so there is no address to name and no per-trader exit to detect. The Kalshi rule type on the Alerts page is a different thing entirely: it fires on large or block prints hitting the Kalshi tape, anonymous and unattributed, and it says nothing about anyone leaving a position. Treating an anonymous Kalshi print as a named whale's exit would be reading in precision that does not exist.
Knowing the shape of what a tool misses is what lets you use the part that works. An exit alert is a fast, accurate report of one observable event. It is not a guarantee that a wallet still holds what it held an hour ago.
How do you actually trade with exit alerts?
Start by deciding whose exits matter to you. An alert is only as good as the trader behind it, so build your list from accounts that have earned it. Use Sharp Score to rank the traders you are considering, star the ones whose reads have held up so they sit on your Watchlist with a note explaining why, and write an exit rule for each. A tight list of genuinely sharp traders produces far more useful alerts than following fifty wallets and drowning in noise.
When an exit alert fires, treat it as a prompt to look, not a reflex to sell. The useful questions are concrete. Is this one trader stepping back, or are several rules firing on the same market inside an hour? Did anything in the underlying event actually change, or is this pure position management? Open the market and read the current price against where the sharp entered. The alert gives you the trigger and the raw facts; you supply the judgment about whether your own reason for being in the trade still stands.
Pair the exit signal with the rest of your read before you act. If Sharp Consensus shows the skilled lean on that market decaying at the same time, that agreement strengthens the case to follow them out. If Divergence and Arbitrage shows the price has run far from the same event's price on the other venue, that is a separate reason to look. If your thesis is intact and one lightly weighted account trimmed, you may reasonably stay. The point of copy trading prediction markets done well is that the alert is intelligence, not an instruction.
Keep the rule set small enough to stay signal. Exit rules have no size floor, so a wallet shaving a small piece off a large stake notifies exactly like a full unwind — the fix is fewer, better-chosen wallets rather than a cleverer threshold. Pair each exit rule with an entry rule on the same address so you see the full arc of a position, and delete any rule that fires often without ever changing what you did.
Is copy trading prediction markets profitable?
Honestly, sometimes, and never automatically. Copying entries alone is a reliable way to underperform the people you are copying, because you get their idea without their timing, their cost basis, or their exit. Adding disciplined exit intelligence narrows that gap, since the most common way tailers turn a good idea into a bad trade is holding a position the smart money has already left. Exit rules are aimed squarely at that failure, but closing one gap is not a guarantee. Past performance does not guarantee future results, and any backtest or simulation you see is hypothetical, not a promise of what your account will do.
The mindset that tends to work is intelligence, not blind copying. Sharp traders are wrong regularly. They size differently than you and they hedge positions you cannot see. Use the signals to inform your own decisions about entries, exits, and sizing. A trader who leaves a position may be taking profit, cutting a loser, rotating capital, or reacting to information you also have access to. The alert tells you they sold. It does not tell you they are right, and it never removes your responsibility for the trade.
It also helps to be clear about what these instruments are. Polymarket and Kalshi are exchanges where you trade shares in outcomes, and each share settles at one dollar if the outcome happens or zero if it does not. Your job as a trader is to buy exposure below what it is worth and reduce it before that edge decays. Exit rules contribute to the second half of that job by making one observable form of departure visible while it still matters. They do not create edge on their own, and no tool can promise profit.
Measured expectations are the right ones. Traders who benefit most from exit rules already have a process. They select their sources carefully, they combine the signal with an independent read on the event, and they accept that some alerts will be false starts. Used that way, as one input among several, exit intelligence can meaningfully improve how you manage risk. Used as a green light to mirror every move a whale makes, it will disappoint, because that is not what any signal in prediction markets can deliver.
What are the limits of exit alerts?
The first limit is the open tab. Rules are evaluated by the page itself, every fifteen seconds, and only while a WhaleTracks tab is open and in the foreground. Close it, background it, or lock the phone and nothing is watching until you come back. There is no server-side watcher and no push channel to fall back on, so an exit rule is a companion to a trading session rather than a standing sentry over your positions.
The second is coverage, described above and worth repeating: the rule sees sells that appear on the tracked sharp tape, not every way a position can be reduced, and not wallets outside that tape. Combined with the fifteen-second cadence and normal upstream latency, that means a fast unwind can be finished before your notification lands. Build your expectations around reacting to informed flow, not front-running it.
The third is interpretation, and it is permanent. An exit alert tells you a position is being sold. It cannot tell you why. Locking in a win, cutting a loss, freeing collateral for something better, and plain rebalancing all look identical, and any product claiming to know the motive is guessing. That gap is yours to fill with your own read of the market.
Finally, keep the honest framing front and center. Alerts, Sharp Score, the Watchlist, Sharp Consensus, and Divergence and Arbitrage are prediction market tools that sharpen your own decisions. They do not replace them, they do not promise profit, and nothing in a backtest or hypothetical simulation guarantees a future result. The realistic promise is narrow and worth having anyway: while you are at the screen with a tab open, a sell from a wallet you care about will not slip past you unnoticed.
WhaleTracks is informational analytics, not financial advice. Past performance does not guarantee future results.