How to Interpret Whale Signals on Polymarket On-Chain
On-chain data makes Polymarket whales visible in a way that is simply not possible on most financial markets. But visibility is not information — knowing which wallet placed which trade tells you almost nothing on its own. The information comes from patterns: size relative to baseline, timing relative to news, return relative to dormancy, and coordination across wallets. This guide organizes the four families of signal Polyloly surfaces and explains how to read them together.
1. Size signals
The most obvious whale signal is an unusually large trade. But "unusually large" only has meaning relative to a baseline — a $10,000 position is noise from a wallet that routinely places $50,000 trades and a statement from one that usually trades $1,500. Every detector on Polyloly normalizes size against the wallet's own personal history rather than a fixed threshold.
The dashboard's live whale feed filters at a $1,000 minimum, but the real signal in the "Unusual Volume" box (top of the homepage) is markets where the last 15 minutes of whale volume exceed three times the rolling 24-hour baseline. That is a size signal with timing built in, and it catches early-moving money before the price has fully adjusted.
2. Timing signals
When a whale trades matters as much as whether they trade. Entries in the last 15 minutes before a market resolves are structurally different from entries placed days ahead — the former carry little information edge beyond what the price already reflects, while the latter can precede the news-driven move that pulls price toward their entry. Our Speed Traders leaderboard ranks wallets by how much price moves in their favor in the 15 minutes after they enter, which is the quantitative way to ask whether a trader's timing edge is actually real.
For editorial traders, timing shows up differently: large entries made just before a news cycle picks up a topic, repeated across several market instances, are a signature of operators with real-world information access. The combination of early + repeated + consistent-direction is stronger than any of the three on their own.
3. Dormancy-return signals
A wallet that has been silent for weeks and then suddenly places a large trade has made three decisions simultaneously: deploy capital, pick this specific market, pick this direction. Each of those decisions is information on its own, and in combination they carry more content than the same trade from an always-on wallet. Our Awakened Whales feed tracks every such event within the last 90 days; the full methodology is in our dormant-whale study.
The exceptional case is a wake wave — three or more wallets returning on the same day after each being independently dormant for fourteen days or more. In 2.3 years of data we have seen exactly two wake waves, both aligned with spikes in geopolitical news coverage. Watching for them is lightweight; interpreting them is high-leverage.
4. Coordination signals
Multiple wallets that consistently trade the same markets, on the same side, within five-minute windows of each other, are not an accident. Our wallet-clustering analysis applies Louvain community detection to the weighted co-trading graph and surfaces groups that behave like a single actor despite appearing independent. Modularity scores in the 0.45–0.50 range — above the 0.3 threshold considered meaningful in network-science literature — are typical on a normal week.
Coordination does not imply wrongdoing; it does imply that the "N traders on this market" counter overstates actual decision-makers by some factor. When interpreting a market's consensus, the cluster view tells you how much of that consensus is actually one operator echoed across multiple wallets versus many independent participants.
Reading the signals together
The strongest signal the Polyloly dashboard produces is a convergence — a returning dormant whale, who is also a member of an active coordinated cluster, placing an outsized position in a market that has just shown a 15-minute volume spike against baseline. Any one of those alone is suggestive. All four at once is as close to smart-money conviction as on-chain data will let you get without subpoena power.
Convergences are rare. Most days only one signal fires, and often none do. That is fine — the point is not to find a signal every day but to be in position to recognize one when it appears. The dashboard is structured so that convergences surface themselves: any trade that triggers two or more detectors appears flagged on the Daily Digest with the overlapping tags visible.