Where Polymarket Edge Actually Lives — A Cohort ROI Analysis of 219,000 Whale Trades

TL;DR for beginners — short version, no jargon:

1. Don't bet early on crypto markets. The math is against you. Traders who buy Bitcoin/Ethereum predictions before the price moves lose about a third of their money.
2. The best profits are in politics and geopolitics — especially when you bet against the crowd. Being right only 1 out of 3 times still makes you money, because the payoff is 5×.
3. You don't need to win often. You need to win when the odds are long. The trader who bets $100 at 20¢ odds and wins 35% of the time makes more money than the trader who bets $100 at 94¢ odds and wins 92%.

Below: the full math, the data, and why this is the opposite of what social-media advice usually says.

Every few months someone posts a screenshot on X of a Polymarket trader with an 80% win rate and says "this is the edge." Then three months later the same trader is −$100,000 and nobody posts the second screenshot.

We wanted to know, empirically, which kinds of Polymarket trades actually generate positive ROI — not which kinds of traders get lucky for three days. So we did the brute-force version. We took 219,000 whale trades (≥$1,000 each) across two-plus years of Polymarket history, grouped them by three simple dimensions — entry timing, market category, bet size — and computed the net P&L per cohort on resolved positions.

The results are surprising. They are also useful, in that they tell you specifically what to stop doing.

The method

For every closed whale position, we tag it with three bucket values:

• Entry timing — the average price the whale paid for their shares: contrarian (<0.30), early (.30–.60), momentum (.60–.85), endgame (.85+)
• Category — from slug patterns: political, crypto, sports, geopolitical, weather, other
• Bet size — total USD committed to the position: small (<$5k), medium ($5–25k), whale ($25k+)

We then look at every cohort with at least 50 resolved positions and compute aggregate ROI — if you had copied the average trade in that cohort, what would you have made (or lost).

Only resolved positions count. Open positions are noise.

The winners

Here is the top of the table — cohorts ranked by ROI:

Timing	Category	Size	N	WR	ROI
contrarian	geopolitical	medium	62	35%	+81.6%
early	geopolitical	medium	95	62%	+40.9%
contrarian	other	medium	130	25%	+32.7%
momentum	political	small	89	87%	+16.7%
momentum	other	whale	165	83%	+14.6%
momentum	crypto	small	113	85%	+12.3%
momentum	sports	whale	372	75%	+11.7%
early	sports	whale	673	51%	+8.6%
endgame	geopolitical	whale	75	99%	+8.5%

Two things jump out immediately.

The biggest ROI cohort has the lowest win rate. Contrarian medium-size bets on geopolitical markets run a 35% win rate and still print +81.6% because when they hit, they hit at 5-to-1 payoff. This is the mathematical opposite of the high-WR fallacy. You do not need to be right often. You need to be right when the market is pricing 15¢ on a 35¢ true probability.

Geopolitical is where the edge is. Both the #1 cohort (contrarian medium) and the #2 cohort (early medium) are geopolitical. The same category appears three times in our top 15. This is the only category where entering early is rewarded. Everywhere else, "early" is a coin flip or worse.

The full picture — sort and filter yourself

The interactive table below is a live mirror of our cohort grid. Click column headers to sort. Use the dropdowns to filter by entry timing, category, or bet size. ROI is green when positive, red when negative — the red rows are the ones that have systematically lost money.

The losers (key takeaways)

If you only look at one thing in that widget, sort by ROI ascending:

This is the single most important line in this article: early-entry crypto is the worst cohort on Polymarket. It's not close. −33.4% ROI across 133 resolved positions means that, systematically, people entering crypto markets at 0.30–0.60 price lose about a third of their stake.

The reason is straightforward. Crypto markets are the most efficient markets on Polymarket. Bitcoin and Ethereum prices are tradeable 24/7 on liquid venues with near-zero spreads and microsecond execution. By the time a Polymarket "Will BTC hit $100k by year-end?" market reaches 0.30, hundreds of hedge funds with actual edge have already priced it. You are not getting in early. You are getting in after everyone with information.

The contrarian-on-sports cohort (−3.9%) is a similar story — sports books are efficient, and taking the other side of the consensus at low prices is almost always just paying the vig.

What this means for normal traders

The cohort table looks like a map, but it's really a list of instructions:

1. If you are trading Polymarket for profit, your priority list should be:

• (a) Contrarian or early entries on geopolitical markets at medium size. This is where your capital has the best expected value. The catch is you'll be wrong most of the time; you need the bankroll and the discipline to keep sizing the same after losses.
• (b) Momentum on political markets at small size. 87% WR at 0.60–0.85 entry is a boring strategy that actually works. Safer than (a), lower ceiling.
• (c) Endgame scalping across the board. +7–9% ROI on whale-size tickets at 0.85+ entry. Capital-intensive and capacity-limited, but almost mechanical.

2. Your avoid list:

• Early entries on crypto — the data says this loses money. Not by a little. By a third.
• Contrarian bets on random ("other") markets without a specific thesis — 25% win rate with no structural reason is just gambling with worse odds.
• Anything on the BTC/ETH/SOL hyperactive updown micro-markets — a separate backtest we ran on those specifically came back negative after 428 trades.

3. The signal we keep missing: the category of the market matters more than any individual trader you're trying to follow. A 75% WR trader on sports is barely breakeven after payoff asymmetry. A 35% WR trader on geopolitics is a goldmine. Before you copy-trade anyone, ask what category they play in.

Why geopolitical markets reward contrarian entry

Geopolitical outcomes are news-driven and catalyst-heavy. "Will Iran agree to X by date Y?" markets sit at very asymmetric probabilities for long periods — 5¢ or 10¢ on the resolution that media consensus treats as unlikely. When that outcome does land, price rockets from 0.10 to 0.95 in hours.

Compare to crypto. "Will BTC hit $100k by year-end?" moves continuously with the underlying spot price. There's no catalyst-driven price jump; there's just slow diffusion. An entry at 0.40 gives you 1.5x payoff if right, not 9x. The math of contrarian betting is completely different.

This is also why whales with domain expertise (journalists, analysts with regional sources, institutional desks with their own models) keep finding edge in geopolitical. The asymmetry is structural, not a market pricing error.

Caveats — because you shouldn't take any of this on faith

The data is two-plus years, but the composition of Polymarket changed a lot during that window. Early periods had thinner markets and different cohort compositions. We don't weight by time, so a 2024-heavy sample shape can bias results.

We filter out the updown/hourly hyperactive micro-markets because they're a different beast. If your strategy lives in those markets, this analysis doesn't apply.

"Resolved" means a settle price crossed 0.97. Some markets may have stalled mid-range and our filter drops them. Survivorship bias runs in favor of clean resolutions.

Paper-trade results ≠ live results. Slippage, latency, and liquidity constraints eat into ROI everywhere. The top cohort's +81.6% won't translate to +81.6% if you actually trade size; it'll be +40% after you cross the spread. Still worth it.

See the live data

The cohort grid is updated whenever our Neon snapshot refreshes. polyloly.com/alpha renders the full sortable table; polyloly.com/wr-leaderboard filters by category. Both are free and require no signup.

If you want to act on this rather than read about it, polyloly.com/insider-picks shows you, live, which Polymarket wallets with ≥75% historical WR and average entry ≤0.60 on non-updown markets have just opened new positions. That's the practical crystallization of the "early geopolitical" cohort insight — the list of traders who already execute the winning strategy. Trade alongside them if you trust the paper; study what they're buying if you don't.

Either way, stop betting early on crypto markets. That one's not up for debate.

---

About the author

Poly Loly — Prediction Markets Expert

Lead analyst behind Polyloly, a real-time analytics platform tracking whale positions across $1B+ in monthly Polymarket volume. Focus areas: on-chain data aggregation, insider-detection heuristics (80%+ win-rate flags on resolved markets), and market microstructure across political, sports, crypto, and esports prediction markets. Published daily trading-terminal intel, trader leaderboards, and automated alerts via @PolylolyHi.

🌐 polyloly.com · 𝕏 @PolylolyHi · ✉ hi@polyloly.com

---

This article is for informational purposes only and does not constitute investment advice. Prediction markets carry a risk of capital loss.

This article contains affiliate links. If you sign up through our links, Polyloly.com may earn a commission, which helps us produce free analytics. It does not influence our analysis.