Prediction Market Alerts: Browser Rules That Watch Sharp Money on Polymarket and Kalshi

Why do you keep missing the smart money on Polymarket and Kalshi?

Prediction markets reprice in seconds. A Polymarket contract can drift from 41 cents to 53 cents in a single afternoon because a handful of sharp traders sized in ahead of the crowd. By the time you notice the chart has moved, the edge that caused it is already priced in. Prices here are probabilities: a share pays $1 if the outcome happens, so a move from 41 to 53 cents is the market telling you its estimate jumped twelve points. The repricing is the signal, and the signal is perishable.

The raw information is public. Polymarket settles on-chain, so every fill is attached to a persistent wallet you can follow across markets. Kalshi publishes anonymous order flow you can watch in aggregate. The problem is not access, it is volume and speed. No trader can eyeball hundreds of markets around the clock, remember which wallets have been right before, and catch the exact moment three of them converge on the same side. Manual refreshing does not scale, and staring at a Live Feed all day is not a strategy.

Alerts solve part of that monitoring problem by turning it into a rules engine. You define the conditions that matter, and while you have the site open, the page watches for them and notifies you instead of making you scan. That shifts your job from surveillance to decision-making, which is the only part where your judgment actually adds value. It is a narrower promise than it sounds, and the next section is about exactly how narrow.

How do these alerts actually work, and where do they deliver?

Browser notifications are the only delivery channel. There is no push server, no mail, and no chat-app relay — not because those are switched off, but because none of them exist in this product and none are planned. That means rules are evaluated by the WhaleTracks page itself: every fifteen seconds, while a tab is open and in the foreground. A backgrounded tab pauses evaluation, and a closed tab stops it entirely. If a rule needs to be watching an afternoon, the tab is part of the setup, and the page says so above the rule builder rather than letting you find out by missing something.

Setup is two steps. Grant the browser notification permission on the Delivery card, then add rules. Rules live in your own browser's storage, so they are per-device and private to you, and they survive reloads. If you save rules without enabling notifications the page tells you plainly that nothing can reach you, because a rule that fires into a void is worse than no rule.

There are six rule types and no others. A wallet trades: a named address opens a position above a dollar floor you set. A wallet exits: that same address sells or reduces. Any sharp enters: any wallet above a Sharp Score you choose enters above a size you choose. Insider flag: the Insider Radar signal crosses a strength you set. Sharps converge: a market where at least N tracked wallets hold the same outcome. Kalshi block: a block or a large print on the Kalshi tape above a dollar figure. Every one of those reads from feeds the rest of the product already computes — there is no separate alerting data source.

Two behaviours keep the notifications tolerable rather than punishing. The engine primes on its first evaluation after a page load, recording what is already on the feed without notifying, so a reload never dumps history on you. And each fire carries a stable identity, so the same trade never notifies twice however long it lingers on the feed. Delivery is capped at five notifications per tick so a busy window cannot bury your desktop, and everything that fired, capped or not, is written to the triggered log on the page.

How do you track a Polymarket whale with a rule?

Start by choosing whose money is worth watching. Not every large wallet is smart money, and size alone is a weak filter. Use Sharp Score to rank wallets by demonstrated skill rather than balance, and use the Watchlist to assemble a shortlist of the traders whose behavior you actually want to follow. A polymarket whale tracker that pings you on every large fill from every big account will bury you in noise. A tracker pointed at ten wallets with a track record gives you a feed you can act on.

Once you have a shortlist, build a wallet rule for each address and attach a dollar floor so you only hear about meaningful entries rather than dust. The address field wants the full thing — 0x followed by forty hex characters — and the form will tell you if you paste a fragment instead of failing silently. Because Polymarket is on-chain, a wallet identity persists across every market it touches, so a rule on a proven address catches it whether it moves into an election market, a sports market, or a macro market.

One boundary shapes what a wallet rule can do. Rules are matched against the live sharp feed — the recent tape of the wallets WhaleTracks tracks — not against an independent scan of the whole chain for your address. If the wallet you named is not among those whose tape the feed carries, the rule will stay quiet no matter how much that wallet trades. Pointing rules at wallets that already appear on the Sharp Leaderboard is the way to avoid that.

For Kalshi the mechanic is different and you should treat it that way. Kalshi signals are anonymous flow, not identified people, so you cannot follow a named trader the way you can on-chain. The Kalshi rule fires on blocks and large prints hitting the tape, which tells you that informed size appears to be moving without any claim about who is moving it. Write the rule to that reality and read the notification the same way.

Which alert rules should you build first?

Begin with the sharp-entry rule, which is the page's own suggested starting point: any wallet with a score of eighty or better entering with at least ten thousand dollars. It requires no shortlist, it fires often enough to teach you the rhythm of the tape, and it is rare enough not to become wallpaper. Run it for a few sessions before you add anything, because a baseline is the context that makes every fancier rule mean something.

Next, add a consensus rule. Independent agreement among skilled wallets is far harder to explain away than a single large trade, so 'three or more sharps converge on a market' is the highest-value pattern the engine surfaces. Pair it with an insider-flag rule, which catches young wallets taking large concentrated positions and is at its most interesting on thin or freshly listed markets. Then, once you have wallets you genuinely follow, add named wallet rules and their matching exit rules so you see the full arc of a position rather than just its opening.

Do not run these rules in isolation. Cross-reference a consensus alert against Divergence and Arbitrage to see whether the same conviction shows up as a mispricing between Polymarket and Kalshi, open the Market Analyzer for a single-market readout, and keep the Live Feed handy to read the tape around any alert before you act. An alert is a prompt to look, not a conclusion. The traders who get the most out of this build a small stack of complementary rules and prune ruthlessly, killing any rule that fires often but rarely leads them to a trade they would keep.

Is copy trading prediction markets profitable?

The honest answer is that copying can be a real edge and it can also quietly lose you money, and which one you get depends almost entirely on how you use the information. Treat these alerts as intelligence, not as instructions to blindly mirror. A wallet alert tells you a skilled trader took a position. It does not tell you their sizing relative to their bankroll, their time horizon, their hedges in other markets, or whether they hold an edge that only works at the price they got and not the price you would pay after the market has already moved on the news.

Adverse selection is the core risk in copy trading prediction markets. You see the entry after it has printed, which means you are frequently buying at a worse price than the trader you are copying. The sharpest wallets sometimes exit before a slower follower has even entered, which is exactly why an exit rule matters as much as an entry rule. Use the alert to start your own analysis: check the market, form your own view of the probability, size the position with the calculator at /bankroll on your own risk settings, and only trade when your independent read agrees with the flow.

Be clear-eyed about performance claims, including your own. Any backtest or historical Sharp Score is a hypothetical simulation, and past performance does not guarantee future results. A wallet that was sharp last quarter can go cold, market conditions shift, and a pattern that worked in one regime can stop working without warning. No alert, ranking, or convergence signal promises profit. The value of the tool is that it puts you in front of informed activity faster and with less noise than manual monitoring, so your judgment operates on better inputs. The judgment is still yours to make and yours to be wrong about.

What are the limits of these alerts?

The most important limit is the delivery model, and it is worth restating rather than burying: browser notifications from an open, foregrounded tab, evaluated every fifteen seconds, with nothing watching when the tab is closed. No mobile push, no background service, no second channel. Any workflow that assumes a notification will find you away from the screen is assuming a feature this product does not have.

Even inside that window there is latency. You are reacting to a trade that already happened, the feed refreshes on its own cadence, and the price that made the print interesting may be gone before you open the market. Rules also only see what the underlying feeds see: wallets outside the tracked sharp tape, and any activity that never prints as a trade, are simply not visible to the engine.

There is also noise and the risk of false positives. A large trade can be a hedge, a roll, or a mistake rather than a conviction signal, and thin markets can produce insider-style flags that turn out to be nothing. Wallets can be split or rotated. Convergence among high-scoring wallets is the sturdiest signal precisely because it is expensive to fake, but no single alert should be treated as proof. Tune your thresholds, prune rules that cry wolf, and let repeated confirmation, not one ping, drive your conviction.

Finally, remember what each platform is. Polymarket and Kalshi are real exchanges where you trade shares that pay $1 if the outcome happens, and the signals there reflect money at risk. Manifold is play-money, so treat any Manifold activity as sentiment and crowd-interest data, never as smart money. Across all of them, alerts surface information, they do not make decisions. Whether there is a trade worth taking, and at what size, is a judgment you own, and no notification will make that call correctly on your behalf.