“Show Me the Size”: Reading Confidence From Bets on Polymarket

▶ Your Real Opinion Is the Money You Put Behind It

Traders say they “like” a market. They’ll tweet hot takes and share charts. But when it’s time to click Buy or Sell, there’s only one honest signal of confidence: how much they’re willing to risk.

Prices in prediction markets tell us what the crowd thinks. Bet size tells us how sure a specific trader is.

At Polyburg, we’ve operationalized that belief into a practical edge. We do two things:

1. ▪Identify the most prolific smart wallets with a high winning score.
2. ▪Score every trade’s confidence by comparing its size to that wallet’s own historical bet-size distribution.

If a smart wallet’s new position lands in its top percentiles, that’s high confidence. If it’s small, it’s more likely a hedge, a flyer, or a liquidity probe. The punchline: knowing who is acting and how big they’re going lets you separate conviction from noise—and trade accordingly.

▶ Why Size Matters: The Market’s Language of Conviction

Prediction markets price contracts between 0 and 100 cents—implied probabilities from 0% to 100%. A 60-cent contract is the market saying “60%.” That price mechanism is why prediction markets are so useful: they aggregate diverse beliefs into one number.

But the market’s other language is stake size. Larger bets typically reflect higher subjective probability and/or a willingness to tolerate risk for expected value. And there’s real money behind this: platforms like Polymarket have seen massive growth—billions wagered on the 2024 U.S. elections alone—raising the stakes and the informational value of size.

Two data-backed ideas are worth carrying with you:

• ▪Prices tend to be well-calibrated. Analyses of regulated prediction data (e.g., Kalshi) show that last-traded prices align closely with eventual outcomes, within tight confidence bands. Price is a meaningful baseline.
• ▪Size is a personal signal layered on that price. On markets with no house edge (betting exchanges and many prediction platforms), bigger tickets often reflect rational risk assessment: taking larger swings when the edge is better, and staying small when it isn’t.

There are caveats. Herding and crowd psychology can create overstability or slow-to-update prices (think Brexit or the 2016 U.S. election). Large bets can sometimes echo groupthink rather than independent insight. That’s exactly why we don’t treat raw size as gospel—we normalize it per wallet and filter through a winning score to find who deserves attention.

▶ The Polyburg View: Confidence = Risk Taken, In Context

I hold a simple conviction: the only reliable signal of how confident someone is comes from how much they’re willing to risk. Everything else is noise until there’s capital attached.

But “how much” must be measured in context:

• ▪A $10,000 trade from a wallet that routinely swings six figures is small.
• ▪A $2,000 trade from a wallet that typically risks $100 is enormous.

So we built a two-layer process:

1. ▪Identify smart wallets via a winning score. We’re interested in consistency, not just one-off windfalls. We weigh realized PnL, calibration (buying at prices that resolve favorably), and discipline across time.
2. ▪Score trade-level confidence by wallet-specific history. For each wallet, we track the distribution of its stake sizes and compute percentiles. Top-decile trades (90th percentile and above) signal high conviction. Middle-range stakes suggest moderate conviction. Small-size nibbles are usually hedges, scouts, or speculative flyers.

In other words, confidence isn’t the absolute dollar amount—it’s the size relative to that wallet’s norm.

▶ A Quick Tour of the Research—And What It Means for Traders

• ▪Price as probability: Prediction contracts trade from 0 to 100 cents, reflecting implied probabilities. Calibration studies show the crowd’s price tends to map well to reality. That’s your baseline model.
• ▪Scale and liquidity: Over $2.7 billion wagered on a single election cycle signals deep liquidity. With more money comes more informative sizing—whales can actually express conviction without moving price 10 points.
• ▪Geopolitical flow examples: Markets like “Will China invade Taiwan?” sitting around 8% probability, with millions traded, show how macro news flows shift both prices and sizes. After major events (e.g., developments in Russia-Ukraine), watch for unusually large bets from proven wallets—they’re often early to new information.
• ▪No house edge dynamics: On exchanges and many prediction platforms, users optimize expected value. Confident bettors scale up when the edge per dollar is higher—size and conviction move together.
• ▪Behavioral risks: Herding and information cascades happen. Big tickets aren’t always brilliant; sometimes they’re momentum. That’s why we care about who is betting big, not just that someone is.

For your trading: treat price as the starting probability and bet size from smart wallets as a confidence overlay. You’ll find better entries, avoid weak signals, and identify true conviction faster.

▶ From Signal to System: How We Quantify Confidence

Here’s the playbook we use to translate “size” into something you can trade.

→ 1) Build the roster of smart wallets

• ▪Winning score: Combine realized returns, price calibration, and drawdown discipline.
• ▪Stability filters: Penalize PnL that comes from a single moonshot; reward repeated, smaller edges.
• ▪Cross-market competency: Bonus for wallets that win across multiple categories (politics, macro, crypto, geopolitics).

→ 2) Normalize bet size by wallet

• ▪Percentile rank: For each wallet, compute the percentile of the new trade’s size relative to its historical stakes. A 95th percentile buy is high conviction for that wallet.
• ▪Time-adjustment: Compare within relevant time windows (recent months) so percentiles track changing bankroll or risk appetite.
• ▪Market-adjustment: Optionally adjust for market liquidity—placing $50k into a thin market can be more “expensive” than $50k into a liquid one.

→ 3) Detect hedges versus conviction

• ▪Size asymmetry: Small, out-of-the-money positions alongside a large core bet often indicate hedging.
• ▪Net exposure: Look across correlated markets (e.g., multiple election states) to see if new size increases or offsets directional risk.
• ▪Scaling pattern: Conviction trades tend to add on favorable moves and defend during dips. Hedgers are one-and-done or add only when correlation spikes.

→ 4) Price interaction

• ▪Impact and timing: Big buys that push price through nearby liquidity are stronger signals than passive fills far from mid.
• ▪Contrarian size: Large tickets against prevailing price moves are especially informative when placed by high-scoring wallets.
• ▪Clustering: If several proven wallets all size up within a short window, that’s often a pre-news or pre-pricing signal.

→ 5) Assign a confidence label

• ▪High: Top 10% by wallet percentile and aligned with increasing net exposure.
• ▪Medium: 50–90% percentile or high percentile but clearly offsetting other risk.
• ▪Low: Below 50% or offsetting/lotto-style structures without supporting size elsewhere.

The output you want is a clean, tradable confidence score per trade, per wallet.

▶ Practical Applications for Polymarket Traders

Use confidence-informed signals to make better decisions before you commit capital.

→ 1) Entry timing: follow size, not noise

• ▪If a top wallet takes a 95th percentile swing at 44 cents and price lifts to 48, consider chasing less aggressively but aligning directionally. Enter partial size; look for pullbacks to the mid-40s to complete.
• ▪If three independent smart wallets size up within 30 minutes on the same side, treat it as a signal upgrade. The probability has likely moved even if the screen hasn’t.

→ 2) Sizing your own positions

• ▪Map your conviction to size explicitly. If you believe the market is mispriced by 10–15 percentage points and you have a live informational edge, size larger (within your risk plan).
• ▪Use fractional Kelly to avoid blowups. Estimate edge (p − price) and odds, then use half- or quarter-Kelly as a practical upper bound. Confidence comes from the math; survival comes from fractionality.

→ 3) Avoiding false positives

• ▪Small tickets at long odds (e.g., 5–10 cents) from a wallet that’s otherwise disciplined? Often a hedge. Don’t overreact.
• ▪Large tickets following a one-way price grind with no new information and from a wallet with weak winning score? That’s likely herding. Downgrade the signal.

→ 4) Cross-market alignment

• ▪If a wallet sizes up across correlated markets (e.g., multiple Senate races, or a macro outcome plus its sectoral impacts), treat the aggregate size as the conviction. Correlated baskets express belief more fully than a single line.
• ▪Conversely, if they add size in one market and trim in a correlated one, that can be a structure adjustment, not fresh conviction.

→ 5) Exit strategy based on confidence decay

• ▪If the original conviction wallet reduces or fully unwinds, treat that as a downgrade—even if price hasn’t moved much. Scale out or tighten stops.
• ▪If new, equally credible wallets add even more size in your favor, consider pyramiding with discipline.

→ 6) Event-driven tilts

• ▪Pre-catalyst: When you expect a news drop (debate, CPI print, court decision), anchor on which smart wallets moved size ahead of time. Their percentile sizes can guide your pre-positioning.
• ▪Post-catalyst: If price jumps but smart wallets double down with top-percentile adds, there’s often more room to run. If they fade, the move may be overextended.

▶ Case Studies (Patterns You’ll See Repeatedly)

→ 1) The contrarian, high-percentile buy

• ▪Market: “Candidate X to win” at 38 cents.
• ▪Action: A top-3 wallet by winning score buys a 98th percentile stake, pushing price to 41 cents.
• ▪Takeaway: That’s conviction. Consider joining with partial size and adding on confirmation. Momentum from informed flow often persists.

→ 2) The hedge that looks spicy but isn’t

• ▪Market: “China to invade Taiwan by 2027” at ~8 cents.
• ▪Action: A macro fund-style wallet adds a small bet here but simultaneously increases exposure in “Taiwan equities up” analog markets (or trims in correlated defense-spending markets).
• ▪Takeaway: Net exposure suggests a hedge. Treat the Taiwan yes-bet as low-confidence; don’t over-adjust your base probabilities.

→ 3) The liquidity probe

• ▪Market: “Inflation above X%” binary near 52 cents.
• ▪Action: Repeated small clips from a mid-tier wallet test the book, followed by a single large sweep from a different high-scoring wallet.
• ▪Takeaway: The probe is noise; the sweep is the signal. The person who moved size told you what they believe.

→ 4) The basket conviction

• ▪Market: Multiple state races in the same election cycle.
• ▪Action: A high-score wallet increases size across four states that share demographic features and polling shifts.
• ▪Takeaway: Aggregate stake—measured across correlated lines—confirms a thesis. Treat the basket as a stronger belief than any single race.

▶ Turning Confidence Into a Repeatable Edge

Here’s a simple, repeatable workflow you can implement today:

• ▪

  Track the right wallets

  • ▪Build a watchlist of top winning-score wallets.
  • ▪Check for consistency across categories (not just “one lucky cycle”).
• ▪

  Monitor percentile-size alerts

  • ▪Flag 90th, 95th, and 99th percentile trades by these wallets.
  • ▪Note whether the trade increases or offsets net exposure across correlated markets.
• ▪

  Filter by timing and price impact

  • ▪High-confidence trades that also shift price/consume liquidity are strongest.
  • ▪Size added during thin market hours can exaggerate impact; keep that in mind.
• ▪

  Confirm with independent information

  • ▪Headlines, polls, on-chain news, and order book dynamics can corroborate the signal.
  • ▪If no fresh information and size clusters late, be cautious: could be herding.
• ▪

  Execute with discipline

  • ▪Start with partial size on the signal; add as information confirms.
  • ▪Use fractional Kelly caps and define exits when the originating wallet unwinds.
• ▪

  Log and learn

  • ▪Keep a diary linking each trade you followed to the initiating wallet’s percentile size, timing, and outcome.
  • ▪Patterns emerge fast: some wallets are monsters in geopolitics, others shine in macro prints.

▶ How Polyburg Scores Confidence (And Why It Works)

Our confidence engine is built for clarity and action:

• ▪

  Wallet-level normalization

  • ▪We compute a rolling distribution of each wallet’s stake sizes and map new trades to percentiles.
  • ▪This controls for bankroll differences and risk preference shifts over time.
• ▪

  Trade context scoring

  • ▪We incorporate price impact, depth consumed, and whether the trade net-increases exposure.
  • ▪Contrarian large trades get a stronger boost than momentum chases.
• ▪

  Correlation-aware aggregation

  • ▪We group related markets (e.g., state clusters, macro-linked outcomes) and sum size for a better conviction read.
  • ▪We discount size that is clearly hedging existing positions.
• ▪

  Winning score integration

  • ▪Confidence only becomes a trade signal if the wallet’s long-run record warrants it.
  • ▪We prioritize calibration and risk-adjusted outcomes over raw PnL.

The output: a concise, interpretable confidence tag on every relevant trade, ranked and ready for you to act on.

▶ Risk, Reality, and the Psychology of Size

A few grounded truths to internalize:

• ▪Not all big bets are smart. Herding exists. Brexit and the 2016 U.S. election are reminders that crowds can be confident and wrong.
• ▪Not all small bets are noise. A sharp wallet may test before committing. Look for the follow-through; that’s when the signal fires.
• ▪Size reflects risk tolerance as much as belief. A well-capitalized trader using fractional Kelly will still scale bigger than a small bankroll even at equal conviction. That’s why normalization by wallet history is essential.
• ▪Liquidity conditions matter. A $100k ticket in a deep election market is different from $100k in a niche policy market. Adjust your interpretation to market depth.

Remember: prices reflect aggregate probability; sizes reflect individual conviction. Use both.

▶ What About Profitability Stats?

On exchanges without a house edge, user profitability tends to cluster: a minority of users capture a disproportionate share of profits. Some studies peg profitable user rates around 40% on certain exchanges, but the distribution is heavy-tailed—a handful of disciplined wallets do most of the winning. Those are the wallets that increase size when they have edge and stay small otherwise. In other words, the exact behavior our framework is built to identify and follow.

▶ Putting It All Together: A Trader’s Checklist

Before you hit Buy/Sell:

• ▪

  Who is trading?

  • ▪Is the wallet on your smart list? What’s their winning score?
• ▪

  How big is the trade relative to their history?

  • ▪Percentile rank 90+? If yes, treat as high-confidence.
• ▪

  What’s the net exposure story?

  • ▪Are they adding risk across correlated markets or offsetting something?
• ▪

  How did the price respond?

  • ▪Did they consume liquidity and shift price, or did they get passive fills?
• ▪

  Is there corroborating information?

  • ▪News, polling, on-chain data, or order book signals?
• ▪

  What’s your plan?

  • ▪Entry price, target, adds on confirmation, exit if the originator unwinds.

If you can’t answer these, wait. The right trade will survive your patience.

▶ Conclusion: Respect the Price, Trade the Size

Prediction markets are efficient because prices summarize the crowd’s belief. But the crowd doesn’t place trades—individual wallets do. When a proven wallet risks more than usual, they’re telling you something: I believe this more than I usually do.

That’s the edge: treat bet size as a calibrated confidence signal, not in absolute dollars but in wallet-relative terms, filtered through a winning track record and market context. It won’t make you right every time, but it will systematically tilt you toward better-informed trades and away from noise.

Ready to put size to work?

• ▪Build a watchlist of high winning-score wallets in Polyburg.
• ▪Turn on alerts for 90th/95th/99th percentile trades.
• ▪Focus on net exposure across correlated markets.
• ▪Execute with fractional Kelly and disciplined scaling.
• ▪Review outcomes and refine your playbook.

Trade smarter. Follow confidence, not commentary. And let the size show you what people truly believe.