Guide

How to tell if a Polymarket whale is actually good

Q: Are big Polymarket traders usually right?

Not necessarily. Trade size measures conviction or capital, not accuracy. Plenty of large wallets are poorly calibrated, are hedging, or are market-making. The only way to know whether a whale is good is to measure its resolved outcomes over a meaningful sample — win rate, calibration, and profit factor — not the size of its latest bet.

Q: What is the best way to measure a Polymarket wallet's skill?

Combine four things: win rate over resolved trades, a Brier score (how well its entry prices matched real outcomes — lower is better), profit factor, and sample size. A 70% win rate on 12 trades is noise; the same rate on 300 trades is signal. Skill also tends to be category-specific, so look at where a wallet is good, not just whether it is.

Q: What is smart money on Polymarket?

Smart money is net positioning from wallets with a strong historical track record (for example, above a 60% win rate), as opposed to the crowd. Stratifying flow by each wallet's measured accuracy tells you where the consistently-right wallets are leaning — a very different signal from raw volume, which counts every participant equally.

Q: How do I get Polymarket wallet accuracy by API?

Idmon exposes GET /api/whales/accuracy for per-wallet accuracy metrics (win rate, Brier score, profit factor, hold time, smart/dumb classification) and GET /api/smart-money for net smart-versus-dumb money flow per market, both under https://api.idmon.io with a Bearer key. See the docs at idmon.io/docs.

21 June 2026 · 7 min read

Every whale tracker shows you the same thing: who placed a big trade. Almost none answer the question that actually matters — is this wallet any good? A large bet measures conviction or capital, not accuracy. Here's how to judge a Polymarket wallet's real skill, and how to get it by API instead of rebuilding months of history yourself.

"Big" is not "good"

The core mistake in most whale tracking is treating size as a proxy for skill. It isn't. A wallet moving $50,000 might be a sharp with an edge — or a market maker, a hedger, an emotional punter, or someone deliberately pushing a price. At the moment the order lands, all of them look identical. The only thing that separates them is what happened after: did their positions resolve in their favour, repeatedly, across a meaningful number of markets?

That's an outcome question, and outcomes take time to observe. Which is exactly why most tools skip it — showing live trades is easy; scoring whether those trades were right requires resolved markets, ground-truth settlement, and per-trade attribution.

The metrics that actually measure skill

If you want to judge a wallet honestly, these are the numbers worth computing — and the traps in each:

Metric	What it tells you	The trap
Win rate	Share of resolved trades that paid off.	Meaningless without sample size, and inflated by betting heavy favourites.
Brier score	How well entry prices matched real outcomes — calibration. Lower is better.	The honest one. A wallet that buys YES at 0.55 and is right 55% of the time is well-calibrated, even at a "low" win rate.
Profit factor	Gross winnings divided by gross losses.	One huge win can mask a leaky strategy; read it alongside trade count.
Sample size	How much to trust everything above.	70% on 12 trades is noise. 70% on 300 is signal. Always check `n`.
Breadth & hold time	Specialist vs generalist; scalper vs conviction holder.	Edge is usually category-specific — a wallet good at politics may be random in crypto.

The single most under-used metric here is the Brier score. Win rate rewards a wallet for piling into 0.95 favourites; calibration rewards a wallet for being right as often as its prices implied. The second is what skill actually looks like.

Smart money vs the crowd

Once you can score wallets, a second signal opens up: stratifying flow by accuracy. Instead of "how much volume is on YES?", you can ask "how much of the net YES exposure comes from wallets that have historically been right?" That's the difference between smart money (say, wallets above a 60% win rate) and the crowd. Raw volume counts everyone equally; smart-money flow weights the wallets with a track record. When the two diverge — the crowd piling into one side while the proven wallets lean the other way — that gap is often more informative than either number alone.

Why this is hard to do yourself

The raw trades are public on Polygon. Turning them into trustworthy accuracy scores is the hard part:

You need resolved outcomes, anchored to truth. Scoring a wallet means knowing how each of its markets actually settled — ideally against on-chain settlement, not a possibly-wrong scrape. Get the labels wrong and every accuracy number downstream is fiction.
You need months of history. A wallet's track record only becomes meaningful across hundreds of resolved trades. That's a long ingestion-and-labelling job before you can rank anyone.
You need per-trade attribution. Partial fills, scaling in and out, and fees all have to be matched to the eventual outcome to get profit factor and calibration right.
You need to resist the obvious-but-wrong shortcut. Following named wallets blindly tends to disappoint, because edge decays and is narrower than it looks. The honest unit is usually a wallet's edge within a category, not the wallet overall.

Getting wallet accuracy by API

Idmon maintains this layer on top of the raw Polymarket data — resolved outcomes anchored to on-chain settlement, per-wallet histories, and the calibration maths — and exposes it as two endpoints:

GET /api/whales/accuracy — per-wallet win rate, Brier calibration, profit factor, average hold time, breadth, and a smart/dumb classification. Sort by winRate, brier (best-calibrated first), or profitFactor.
GET /api/smart-money — net smart-versus-dumb-money flow per market, ranked by the gap, or pass conditionId for one market's time series.

# best-calibrated wallets, not just the biggest
curl -H "Authorization: Bearer YOUR_KEY" \
  "https://api.idmon.io/api/whales/accuracy?limit=20&sort=brier"

# where the proven wallets are leaning right now
curl -H "Authorization: Bearer YOUR_KEY" \
  "https://api.idmon.io/api/smart-money?limit=20"

Each wallet row carries the metrics above, so you can apply your own bar — for example, "win rate above 0.60 and at least 200 resolved trades and a Brier score under 0.20" — rather than trusting a single headline number.

Try it free, no signup. Idmon's /public/* endpoints serve much of the data at 60 requests/min with no key. See the docs for what's available without authentication. For per-wallet accuracy, smart-money flow, real-time push, and a commercial licence, request a key — you'll be live within a business day.

What this is not

Accuracy is evidence, not a guarantee. A strong track record raises the odds a wallet is skilled; it does not promise the next trade wins. Sample size and calibration tell you how much to trust it.
This is not financial advice. Wallet accuracy is one input among many, not a trade trigger.
Idmon scores pseudonymous wallet addresses from public on-chain data. It does not attempt de-anonymisation and does not surface personal data.

FAQ

Are big Polymarket traders usually right?

Not necessarily. Size measures conviction or capital, not accuracy. Plenty of large wallets are poorly calibrated, hedging, or market-making. The only way to know if a whale is good is to measure its resolved outcomes over a meaningful sample — win rate, calibration, and profit factor — not the size of its latest bet.

What's the best way to measure a Polymarket wallet's skill?

Combine win rate, a Brier score (how well entry prices matched outcomes — lower is better), profit factor, and sample size. A 70% win rate on 12 trades is noise; the same on 300 trades is signal. Skill also tends to be category-specific, so look at where a wallet is good, not just whether it is.

What is smart money on Polymarket?

Net positioning from wallets with a strong historical track record (for example above a 60% win rate), as opposed to the crowd. Stratifying flow by measured accuracy tells you where the consistently-right wallets are leaning — a different signal from raw volume, which counts every participant equally.

How do I get Polymarket wallet accuracy by API?

Idmon exposes GET /api/whales/accuracy for per-wallet accuracy metrics and GET /api/smart-money for net smart-versus-dumb money flow, both under https://api.idmon.io with a Bearer key. See the docs for the full response shape.

Does a high win rate mean a wallet will keep winning?

No. Past accuracy is evidence, not a guarantee — especially on small samples or in changing conditions. Calibration and sample size tell you how much to trust a track record, but none of it is financial advice.