← All posts · July 13, 2026 · 7 min read

Inverse copy trading: making money by copying losers backwards

Roughly 70% of Polymarket's 1.7 million trading addresses have lost money, according to an analysis by DeFi Oasis. Sooner or later everyone who reads a stat like that has the same idea.

If most traders lose, don't hunt for the rare winner. Find a reliable loser and do the opposite.

That's inverse copy trading — counter trading, fading, whatever you want to call it. It's a real strategy with a real logic behind it. It's also harder than the one-line version makes it sound, and the gap between the idea and the results is worth understanding before you flip the switch.

This isn't a fringe idea

Polymarket's own newsletter covered it in an issue on the copy-trading arms race. The quote: "We hear about people counter-trading accounts with negative P&L historically. They want to punish the big losers and do the exact opposite." Better filtering for negative P&L — tools to find accounts to fade — was described as one of the most requested features.

So the demand exists, and enough people are doing it that the platform ecosystem is building UI for it. The question isn't whether inverse copying is a thing. It's whether it makes money after costs.

Why fading a loser can work at all

A wallet that's down over ten trades might just be unlucky. A wallet that's down over four hundred trades is doing something wrong on purpose, even if the owner doesn't know it.

The classic patterns: buying whichever side just moved (chasing), paying any price for the favorite, doubling size after losses, panic-selling bottoms. These aren't random errors — they're systematic biases, which means they point in a predictable direction. And anything predictable is, in principle, a signal you can flip.

That's the key distinction. A coin-flipper can't be faded profitably; after fees you'd just be a slightly poorer coin-flipper. Only bias can be faded. Which is why the first job is telling bias apart from noise — the same skill as reading any wallet, and the reason win rate alone tells you almost nothing.

Why it's harder than it sounds

Part of a loser's losses never transfers to you. This is the piece almost everyone misses. A chronic loser bleeds money three ways: bad picks, crossed spreads, and taker fees — on crypto markets the fee peaks at $1.75 per 100 shares at 50/50 odds, exactly where impulsive traders love to trade. When you invert them, you only get paid for the bad picks. The spread they donated went to market makers, not to you. And you pay your own spread and taker fee on every inverted trade. A wallet can be genuinely, consistently down while the directional part of its wrongness — the only part you can harvest — is smaller than your round-trip costs.

The mechanics cost more than you'd guess. They buy "Yes" at $0.62; you buy "No." But "No" isn't sitting there at $0.38 — you pay the ask on the other side, maybe $0.395 with the spread. A couple of cents of friction per trade, against an edge that might be a couple of cents per trade.

A loser's variance is your variance. Bad traders aren't wrong every time — they're wrong on balance. Their occasional big win is your big loss, and losing wallets tend to be streaky in ugly ways. A doubling-down loser shows a P&L curve of steady small wins followed by a crater; invert that and you eat steady small losses while waiting for the crater to pay you. Mathematically fine, psychologically brutal. Most people quit the fade three days before it works.

The wallet might not be what it looks like. Some "losers" are one leg of a hedge or an arbitrage — the wallet loses here and wins somewhere you can't see. Both-sides bots produce exactly this signature, which is why spotting bot wallets matters just as much when you're fading as when you're copying. And the same newsletter that described counter trading also described winners running decoy accounts on purpose. An account that looks stupid may be performing stupidity for an audience.

Losers retire. Wallets that blow up stop trading — the money's gone. Your carefully chosen fade target has a shelf life, and the ones that keep going long-term are often the hedged or botted accounts you didn't want anyway.

How invert mode actually works

Mechanically it's simple. In btc5min, invert (counter) mode is a per-wallet rule: when the wallet buys an outcome, the bot buys the opposite outcome in the same market instead. All your other rules still apply on top — minimum trade size, category filters, your own take-profit.

Two settings do most of the work:

• Minimum trade size, set high. You want to fade conviction, not $2 boredom trades. A loser's small trades are noise; their big trades are where the bias shows.
• Category filter, set narrow. Fade the wallet only where it demonstrably loses. Plenty of traders are sharp on sports and donate everything back on crypto — fading their whole feed throws away the half that's right.

I've watched a wallet lose eleven straight 5-minute rounds buying whichever side had just moved. Inverting that looked like free money — until the same wallet made most of it back in two violent rounds the same night. The fade was still positive over the month, but barely, and only for someone who sized small enough to survive the two bad rounds.

Who's actually worth fading

A short checklist, in order of importance:

• Months of history and hundreds of trades — small samples lie in both directions.
• Directional trades, not both-sides patterns.
• Consistent losses concentrated in one category, from realized trades rather than open positions.
• Still active. A great fade target from March that stopped trading in May is a museum piece.
• Human-scale activity. Thousands of predictions means bot; see above.

Most wallets fail this list. That's the honest catch: reliable, active, directional, human losers with size are nearly as rare as reliable winners — which is why inverse copying is a niche tool, not a shortcut around the normal profitability math.

The verdict

Inverse copy trading is real, Polymarket's own ecosystem is building tools for it, and the logic holds when the target is genuinely biased. But the edge is thinner than the loser's losses suggest, the costs come out of your side twice, and the best-looking targets are often bots, hedgers, or decoys.

If you try it, try it the boring way: pick one fade target that passes the checklist, run it in practice mode for a week or two, and compare the paper P&L against what fees and spread took. The idea survives that test less often than it should — and when it does, you'll know exactly why.

btc5min is an independent tool and is not affiliated with, endorsed by, or operated by Polymarket. Nothing here is financial advice. Prediction markets are risky, copying another trader does not guarantee profit, and you can lose the money you trade with.