Betting with the people who are winning: Prediction Radar and a wave of Instagram copy-trading accounts

Prediction Radar is a read-only Polymarket wallet tracker: see which public wallets have been winning recently, follow them, and your phone buzzes when they trade. It launched on iOS in early September, right before and during the weeks when prediction-market interest peaked around the US midterms. A batch of Instagram "copy-trade" accounts pushed one shared discount code and drove it into the top 50 of the US Finance Top Grossing chart. Anyone can build these features, and the store already has twenty-plus similar apps. What it got right is distribution and timing. Below: how it ran, and how much opportunity is left for you.

A Reel shot in a car

On Nov 4, the day after the midterms, Instagram account nathan_copies (Nate Brown) posted a car-filmed Reel with a four-step tutorial in the caption:

Step one: in Prediction Radar, browse the "Top 50" list and find the traders with the best performance. Step two: look at their past trades to see what markets they focus on. Some do sports, some politics, crypto, or weather. Step three: find someone worth following, tap "Track," and you get notified whenever they open a new position. Step four: go to Polymarket or Kalshi and bet alongside them (invite code on the homepage). It takes only a few minutes a day.

This isn't a one-off. On Oct 17, joey_markets ("Joey education") posted a Reel saying "The Kalshi strategy that was going viral just changed, and it's even better! Use code RADAR1 for 8 off your first month of Prediction Radar", tagged #makemoneyonline2026 #unitedstatessidehustle. Another post read "Followed a top sports trader today, he placed nearly $50k in 3 bets". After mid-October, a batch of accounts with similar names (xxx_markets, xxx_copies, xxx.predicts) started posting the same kind of content, all with the same RADAR1 code, and many posts also carried #kalshipartner and exchange referral lines like "Trade $25 on Kalshi, get $55".

What it is

A tool for watching how others bet. Polymarket is an on-chain prediction market where every wallet's trades are public. Prediction Radar turns those public records into leaderboards: PnL, ROI, win rate, volume, positions. You build a watchlist, and followed wallets trigger a push when they move. The store page is explicit: no wallet connection, no trading, no custody of funds, "for informational analysis only, not investment advice."

iOS ratings are nearly 4 times the Play count. The main battlefield is iPhone, Play came later. The same iOS developer also makes World Monitor (real-time OSINT) and Globus (live Earth cameras), a small studio focused on real-time data dashboards. The Play listing uses a different company name, most likely the same team (inferred).

Why now

Prediction markets became mainstream topics over 2026 years. During the week of the Nov 3 midterms, Kalshi and Polymarket election markets both reported record volume (CBC 11-02). Kalshi itself added 22.8k iOS US ratings that week. A wave of newcomers arrived for the first time, and their first question was "what should I bet on?" The easiest answer was "bet with whoever's winning."

Prediction Radar's curve matches this: Instagram posts with its code began appearing densely around Oct 13, iOS ratings accelerated from Oct 31, and Play installs jumped on election day.

Where users come from

Instagram Reels account network. This is the main channel and has the most evidence. The format is uniform: a selfie or talking-head clip in a car, saying "the trader I follow won this much today," with RADAR1 in the caption. Single posts get anywhere from single digits to hundreds of likes, while same-name TikTok accounts are nearly empty (joey_markets has only 3 followers on TikTok), suggesting these people are Instagram-only. It looks like paid UGC or an affiliate program rather than organic word of mouth. Whether the official team runs these accounts or outsources them can't be determined from outside (inference). Instagram has also auto-generated a topic page for "How To Use Prediction Radar For Polymarket Trades," so people are searching for it on the platform.

These people were already earning exchange referral bonuses for new sign-ups. On one video, viewers open a Kalshi account and he earns one referral. They download Prediction Radar and subscribe, and he earns another. For them, promoting a copy-trade app is an easy add-on.

In-app invite codes. The iOS update notes say "User referral codes & free unlocks," so inviting friends unlocks features for free.

Store search. The subtitle "Track the top 1%" targets terms like "polymarket tracker." In the App Store's "You Might Also Like" section, it's linked to PolyEdge and PolyScout.

Other channels are mostly empty: no one has mentioned it on X in the past two months, no press coverage, and no ads found in the Meta ad library (not finding them doesn't mean none ran).

Where the money comes from

Entirely subscription-based, and the price is high: $14.99 a week or $44.99 a month. The free tier only tracks 3 people and shows some charts. Searching for people and viewing the full rankings requires paying. The 8%-off first-month code suggests monthly is the main push.

The price works because buyers aren't paying for a tool, they're paying for a shot at winning. Following one right trader can cover a year's subscription in a single trade. That's why it can pay generous referral cuts to the account network and still reach the top 50 of the Finance Top Grossing chart. Real revenue isn't public. The estimate of tens of thousands of dollars a month in the US is inferred from chart position.

The cost shows up in the reviews too. The US App Store one-star reviews fall into two groups. One says they were charged: "Got a 7-day free trial in Oct 5, cancelled on the 6nd, and still got charged $14.99" (10-09). The other says the copy trading itself doesn't pay: "Traders with one star and five stars are the same. After subscribing I copied 8 trades and only 1 worked out. Don't buy." (10-21) "Kind of a scam" (10-26). Most five-star reviews are just a word or two. People who buy a "follow the winners" app come back to complain when they lose. That comes with this category.

Who it's competing with

(App Store, Google Play, 11-06)

Plenty of people are doing this: web tools, open-source scripts, two dozen small iOS apps. But Prediction Radar is the only one that's a decent mobile app and has creators pushing it. Among App Store finance apps with poly, whale or prediction in the name, its rating count is second only to the exchanges Kalshi and Polymarket and to dub. The biggest risk is Polymarket itself: it already has leaderboards and profiles, so adding "follow alerts" is a small step.

If it were you

Bottom line: worth a try, but don't copy it wholesale. The "trade-signal layer plus creator referral codes" playbook has been proven and isn't saturated yet. The problem is it got its start on one election, similar apps are everywhere, and Polymarket could take the core feature away at any time.

Who it suits: one or two people who can build data-driven mobile apps and connect to on-chain or public trading data, and who know or are willing to reach prediction-market and trading creators with small followings and can negotiate a revenue share.

Not for: people who want to rely only on app store search and organic traffic. Two dozen-plus similar apps got to a 0 rating this way. Also not for anyone who can't stand users losing money and then leaving one-star reviews calling it a scam.

How to do it:

• Don't build another Polymarket leaderboard. The opening is what Polymarket does badly: reviews say "one-star and five-star traders are the same," which points to a ranking methodology problem. Rank by market type (sports, politics, weather) separately, filter out wallets that got lucky on one or two bets, and show whether following a wallet actually made money. That's where you can differentiate.
• Swap the data source. Any place with public trading records and a crowd of new people who want to copy trades can use this same shell: other on-chain prediction markets, on-chain sports betting, copy-trading communities with public positions. Kalshi is a centralized exchange with no public wallets, so you can't track individuals there. That's also why these videos say "trade on Kalshi" but rely on Polymarket data.
• The paywall follows its lead: track 2–3 people for free, with a high-priced weekly subscription. But spell out the trial and billing rules clearly, or you'll collect one-star reviews over it.

How to get users: go straight to the Instagram and TikTok creators with small accounts who already promote Kalshi and Polymarket referrals. Offer them exclusive discount codes and a subscription revenue share. They already tell the "copy-trade to make money" story and are missing an app they can open on camera. Start with 10 accounts, see which codes convert, then add more.

When to move: election hype fades, and the next big windows are sports seasons and the run-up to 2028 elections. Lock in products and creator deals before the next wave of "copy trading" short videos, and negotiate monthly rather than annual terms.

What would mean you read it wrong:

• Over the past month since the election, Prediction Radar's iOS rating growth dropped to single digits per day and it fell out of the top 100 in Finance Top Grossing. That means the demand was driven by the event alone, so don't chase it.
• Polymarket's official app adds "follow-wallet alerts": the room for this kind of information tool is mostly gone, leaving only what the official app won't do (cross-platform, category rankings, copy-trade performance stats).
• Conversely, if name-borrowing apps like Whalecast also reach the grossing charts through creator referral codes, the playbook matters more than the product, and that playbook isn't saturated yet, so it's worth moving on now.

---

Data and sources: App Store product pages, reviews and US charts (10-28 to 11-06); Google Play product pages and country charts (10-29 to 11-06); Instagram posts from nathan_copies (11-04), joey_markets (10-17, 10-22) and others; CBC election coverage (11-02); X and Meta ad library searches (no results). Where the text says "inferred" or "estimated", it's inference or estimation.