Copy Polymarket whales without losing your shirt

Polymarket whale wallets size up on the markets they actually believe in. The mistake retail makes: copying every whale move blindly. Use the data as context, not signal.

Tool

polymarket_whale_positions. params:

param	default	notes
`limit`	5 (free) / 25 (pro)	max wallets

Returns each wallet with score (FalsifyLab composite based on PnL + win rate + activity) and active_positions (open bets sized >$10k).

Agent prompt

Call polymarket_whale_positions with limit=10. For each wallet, list its
top 3 active positions by entry_usd. Comment on which markets have 3+
whales clustering on the same side. Skip wallets with score < 0.5.

Expected output

wallet     | score | top positions
0xa1b2...  | 0.72  | "US election - candidate X YES @0.41" $52k
                    | "FOMC June cut" YES @0.31" $28k
                    | "Bitcoin 100k by Q4" YES @0.62 $11k

3+ whale concurrence: "US election - candidate X YES" (5 whales, total $310k)

What to do next

A market with 5 whales same-side and total entry above $200k is real conviction. One whale = noise.
Always check entry_price vs current. late copy after the price has moved is dead money.
Pair with confluence_today to see if the Polymarket bet correlates with other signals (insider flow, ETF flow, etc).

Risk note

Polymarket is binary-outcome prediction markets. Position sizing should reflect a binary loss; never risk more than you can completely zero on a single market.

Limits

Free tier returns top 5. Pro 25 default, 50 max. Free updates every 12h; pro every 30min.