Original Title:I Analyzed 112,000 Polymarket Wallets. Here's What Separates the Top 1% from Everyone Else, Author: darkzodchi（@zodchiii）

Compiled by｜Odaily Planet Daily（@OdailyChina）；Translator｜Asher（@Asher_0210）

After a systematic review and analysis of over 112,000 Polymarket wallets and 6 months of on-chain data, a rather intuitive yet surprising result emerged. About 87.3% of users end up losing money on the platform.

This statistical analysis covered multiple key dimensions, including each transaction record on-chain, transaction volume, win rate, profitability, types of markets participated in, entry timing, and position size. The entire data organization process took 3 weeks, and the conclusions drawn are not in line with what many people might intuitively believe.

Many believe that top players in prediction markets often have some obvious advantages, such as insider information, or employ obscure and complex calculation models. However, the data shows otherwise. The top 1% of players consistently focus on doing a few things right for an extended period, while the other 99% often do quite the opposite, and then wonder why their funds continually dwindle.

The Polymarket leaderboard can be misleading

If you were to open the Polymarket leaderboard now and sort it by profit (PnL), you would find some irregularities. For instance, the top wallet has only 22 positions; the fourth-ranked wallet has just 8 transactions; while the eighth wallet has only placed a single bet and still makes it into the historical top ten.

These addresses are not genuinely considered traders. In many cases, they are simply whales placing single bets exceeding $5 million on a one-off event and happening to get it right; or it may involve those with information advantages, or even a combination of both. Regardless of the case, the data from just a few transactions offers nearly no replicable trading rules. The outcome resembles a large-scale “coin toss” rather than a set of strategies that can be copied.

Thus, the first step in the analysis is to filter out these noisy data and retain only statistically meaningful samples. The screening criteria include the following:

• At least 100 settled positions to ensure a statistically significant sample size;
• Active trading period of no less than 4 months to exclude accounts that succeeded solely by luck;
• Participation in at least 2 different markets to avoid betting on a single event;
• Total trading volume exceeding $10,000 to ensure that participants actually invested funds.

Under such conditions, of the initially analyzed 112,000 wallets, only about 8,400 wallets remain with sufficient data value after screening. These 8,400 addresses represent a truly meaningful data set for research, rather than those “hero accounts” that have conducted just a few transactions yet made millions. The common characteristic among these addresses is ongoing trading and stable data, making it easier to observe genuine behavior patterns.

Interestingly, after screening, the truly stable traders look completely different from the profiles on the leaderboard. They are not flashy, and most people have likely never heard of their names. Their profits usually range between $50,000 and $500,000, not the millions others might achieve.

However, what deserves attention is not how much they earn, but their trading processes and methods behind them. Because what can truly be replicated is not the outcome, but the process.

Three common misconceptions to break

Misconception 1: Top traders have win rates between 80% and 90%

This is not the case. Based on the screened data sample, rather than on the leaderboard that showcases whale accounts that profited from a single bet, the actual long-term profitable wallets generally have win rates between 55% and 67%. This means that even top traders will make incorrect judgments in a considerable portion of their trades. For example, one address has completed over 900 settled positions, accumulated profits of $2.6 million, but has a win rate of only 63%. In other words, over one-third of their bets are wrong, yet they still make substantial gains in the prediction market.

The obsession with win rates is often the easiest trap for novice accounts to fall into. Many novices prefer to buy contracts at $0.90, because it seems “safe.” The probability of YES is already 90%, suggesting that the outcome is nearly certain, so they buy at $0.90, ultimately only making a profit of $0.10 if the event occurs. However, if there’s a misjudgment, they might lose the entire $0.90, with a risk-reward ratio reaching 9 to 1. This cycle of behavior, repeated enough times, quickly depletes their account funds. In the dataset, such situations have recurred across hundreds of addresses.

Misconception 2: The strongest traders participate in every market

The reality is quite the opposite. The best-performing wallets typically engage in no more than three market categories, with most focusing on one or two fields. Some addresses only predict cryptocurrency-related events; some only participate in weather-related markets; there's even one address that almost exclusively trades on whether “Bitcoin will reach a certain price by Friday.”

In prediction markets, excessive diversification often means a decline in the quality of judgments. General participants often perform averagely, while highly focused participants are more likely to achieve consistent profitability.

Misconception 3: Speed is everything

This statement holds true only in rare cases. For instance, some crypto markets that settle in 15 minutes do require quick responses. However, in the vast majority of markets, top traders do not win solely based on speed. More frequently, their practice is to gradually build positions over several days or even weeks. They are not rushing to compete with others in clicking speed but patiently wait for prices to show a clear deviation. When the price has diverged sufficiently, even if the market takes two weeks to correct, the overall mathematical expectation remains favorable to them.

Five trading patterns worth learning

Pattern 1: Trade in the opposite direction during extreme emotions

In the entire dataset, this is the most obvious and stable profit signal. Among the 8,400 screened wallets, this behavior is almost the primary indicator for determining whether an account is profitable in the long run.

When a certain contract is pushed up to 88% by market sentiment, many top wallets start to sell YES; while when the price drops to about 12%, they begin to buy in gradually. Of course, this is not blind contrarian trading, nor are they opposing the market just for the sake of it. They only enter the market on a large scale when they judge that market sentiment is clearly overreacting.

The effectiveness of this strategy relates to a classic phenomenon known as the “hot-cold bias.” This phenomenon was discovered as early as the 1940s in horse racing betting studies and occurs in nearly all markets where human betting is involved. Simply put, people often overestimate the likelihood of outcomes that seem “almost certain” while underestimating low-probability events.

Further statistics reveal that the top 50 wallets with the highest profits typically have entry prices that deviate from the market consensus probability by 6% to 11%. They do not engage in betting on a 50/50 basis, but instead wait patiently for odds to be significantly favorable to them before entering. This trading approach may seem a bit dull, but it is stable and highly profitable in the long-term data.

Pattern 2: Position management closely resembles the Kelly Criterion

When comparing the position sizes of the top 200 profitable wallets with the “implied advantages” they faced at that time, a very clear correlation can be seen. In other words, they do not randomly bet; the betting size changes proportionally to what they believe their advantage is, meaning when they believe the advantage is substantial, the position significantly increases; when the advantage is small, they only bet a small position; if there is no clear advantage, they simply do not trade.

Whether these traders have actually read the Kelly Criterion (Kelly Criterion), or simply developed this intuition through long-term losses and practical experience, is hard to determine. However, mathematically, their behavior is very close to the Kelly Criterion.

The Kelly Criterion is typically written as: f* = (p × b − q) / b, where: p represents the trader's perceived probability of the event occurring; q = 1 − p; and b represents the odds-to-risk ratio (potential reward ÷ risk cost).

As a simple example, suppose a trader judges that an event has a 60% chance of occurring, while the market price is $0.45. The reward ratio would be: b = (1 / 0.45) − 1 ≈ 1.22, and substituting into the formula gives: f* = (0.60 × 1.22 − 0.40) / 1.22 ≈ 0.272. This means that the complete Kelly strategy suggests putting 27% of funds on this trade.

However, this approach carries extremely high risk in actual trading, and the volatility can be very large, likely plunging the account into significant drawdowns in a short time. From the data, truly profitable wallets typically adopt a more conservative version, approximately close to a quarter of the Kelly Criterion. In other words, if the full Kelly formula suggests betting 27%, they usually only bet around 7%.

In the most certain trading opportunities, the position might be raised to 12% to 15%; opportunities with moderate confidence typically allocate only 2% to 5% of the position; while markets without a clear advantage, they often choose not to participate at all. In contrast, losing accounts usually fall into two extremes. Either they bet 80% of their funds in a single trade, completely relying on luck; or they spread $10 thinly across forty to fifty markets, thinking they are “diversifying risk.” But in reality, this looks more like constantly paying transaction fees, making the account appear busy.

Pattern 3: Extremely focused specialized trading

When categorizing the 112,000 wallets by the types of markets they participated in, noticeable differences emerge. These categories include crypto markets, political events, sports, weather, geopolitics, entertainment, and science among others. The analysis concludes that:

• Wallets participating in 1 to 2 categories have an average PnL of about +$4,200;
• Wallets engaging in 3 to 4 categories have an average PnL of about -$380;
• Wallets involved in 5 or more categories have an average PnL of about -$2,100.

This relationship demonstrates a clear linear trend. The more categories one participates in, the higher the probability of loss.

Prediction markets across different categories rely on entirely different information systems. Crypto markets are often influenced by exchange fund flows, whale addresses, funding rates, etc.; political markets depend on poll data, grassroots news, congressional schedules, etc.; while weather markets rely more on NOAA meteorological models, atmospheric data, and satellite observations.

Two case studies are especially representative. Case 1: Wallet A only trades Bitcoin in 15-minute settlement prediction markets and never participates in other types of markets, such as “Will BTC exceed a certain price in the next 15 minutes.” This address completed 502 predictions with a win rate of 98%, accumulating about $54,000 in profits. Its advantage is quite simple, just continuously monitoring the depth of the Binance order book and trading rapidly when Polymarket prices lag by 10 to 30 seconds. In other words, relying solely on a few seconds of information discrepancy has been exploited hundreds of times.

Case 2: Wallet B only participates in weather-related markets. Their trading strategy is also straightforward; they read temperature forecast data publicly released by NOAA each day, then compare it with Polymarket's market pricing. If there's a significant deviation from the prices predicted by these supercomputers, which have been optimized over decades, they enter the market for trading. Just in the New York temperature prediction market, this address has an accuracy rate of 94%.

It is important to emphasize that these individuals are not geniuses. The real key is that they have found a niche where they understand far more than the average Polymarket participant, and they continually exploit this advantage. They do not frequently change strategies, nor do they experience FOMO driven by market trends. They merely execute the same logic over and over around the same advantage.

Pattern 4: Trading price fluctuations, not event outcomes

Most Polymarket users trade in a very basic manner, holding onto contracts until events settle, either winning or losing, resulting in typical binary outcomes. But the approach of top wallets is completely different. Many times they buy at $0.40, and when news or market sentiment drives the price up to $0.65, they sell immediately. They do not care whether the event ultimately happens; as long as the price reflects new information, they complete the transaction and exit.

In the dataset, some of the best-performing addresses even have no settled positions at all. They never hold contracts until final settlement, opting instead for continuous swing trading based on price mismatches. Statistics show that the average holding time of top wallets is usually only 18 to 72 hours, while wallets in the bottom 50% of profitability often hold onto positions until settlement, sometimes even for weeks or months.

This does not mean that holding until settlement is always wrong. Sometimes when the judgment is very certain, long-term holding does indeed become the better strategy. However, from the overall data perspective, top wallets use funds in a much more proactive and flexible manner than most people imagine. They are not passive bettors, but real traders.

Pattern 5: Always avoid breaking news

Intuitively, we often believe that the most keenly aware funds should enter the market immediately when emergencies arise, such as military conflicts, election results, executive resignations, and other significant news. However, the data shows that top wallets often actively avoid the immediate period right after news breaks. They typically wait for emotional funds to flood into the market first, causing prices to fluctuate dramatically in a short time, and only commence trading once market sentiment stabilizes.

Across the entire dataset, a very clear pattern is that the best trading opportunities often arise before the market acknowledges the event or after the market has overreacted. When everyone is discussing the same issue, that often represents the worst time to enter the market. At that point, market prices are usually highly efficient, and the advantages that can be gained are minimal.

Five recommendations for operation

Select a niche and focus on it long-term

Whether it's crypto, politics, weather, or sports, it can be any, but you must choose the field you are most familiar with. Next, trade only in this category of markets for at least three months. Do not make exceptions, nor participate in other trending events out of impulse. Even a simple “bet on an election” can easily disrupt your original judgment system.

Record every prediction

Before each trade, write down several key data points, including your estimated true probability, current market price, expected advantage, and planned position size. After accumulating more than 50 trades, review them. For example, if certain predictions are marked as having a 70% probability, check if the actual hit rate is truly close to 70%. If there is a significant deviation, it indicates that there are biases in probability judgment that must be calibrated before increasing positions.

Position management should approach a quarter of the Kelly Criterion

First, calculate the theoretical position according to the Kelly Criterion, then divide it by 4 to get the actual position. This number may seem small, but it is the key to controlling risk. The result of over-leveraging usually leads to one outcome—account liquidation.

Only trade when the advantage is clear enough

If the expected advantage is below 8% to 10%, forgo it altogether. Even if the opportunity seems tempting, learn to wait. The wallets that perform the best in the data typically make only 2 to 3 trades per week in each market category. Trade quality is far more important than trading quantity.

Persist in recording and reviewing

Establish a complete trading spreadsheet to record every trade, the results, and the issues encountered. Wallets that have long-term performance improvement generally conduct systematic reviews of their mistakes; while those that remain stagnant or continuously lose often just repeatedly make the same errors, attributing outcomes to bad luck.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。