作者：Odaily 星球日报
Original title:I Analyzed 112,000 Polymarket Wallets. Here's What Separates the Top 1% from Everyone ElseAuthor: darkzodchi(@zodchiii)
Compiled by Odaily Planet Daily@OdailyChina); Translator | Asher (@Asher_ 0210)
rightMore than 112,000 Polymarket wallets and six months of on-chain data.After systematic review and analysis, a rather intuitive yet surprising result emerged.Approximately 87.3% of users ultimately lose money in their transactions on the platform..
This statistical analysis covered multiple key dimensions, including every on-chain transaction record, trading volume, win rate, profit and loss, types of markets participated in, entry time, and position size. The entire data processing process lasted three weeks, and the final conclusions were...Contrary to many people's intuition.
Many believe that top players in prediction markets often possess a significant advantage, such as access to insider information or the use of complex, little-known computational models. However, data suggests otherwise.The top 1% of players consistently and consistently do a few things right and repeat them repeatedly. Meanwhile, the other 99% of users often do the exact opposite, and then wonder why their funds are continuously draining away.
Polymarket's rankings are actually quite misleading.
If you open the Polymarket leaderboard now and sort it by profit (PnL), you'll find some anomalies. For example, the top-ranked wallet only has 22 positions; the fourth-ranked wallet only has 8 trades; and the eighth-ranked wallet only has 1 bet, yet it still manages to rank in the top ten all-time.
These addresses are unlikely to be genuine traders. In many cases, it's simply a case of a whale betting over $5 million on a single event and winning by chance; or it could be people with an informational advantage participating, or even both.In either case, data from only a few trades offers virtually no learnable trading patterns. The result is more like a massive "coin toss" than a replicable strategy.
Therefore, the first step in the analysis is to filter out this noisy data and retain only the samples that are truly statistically significant. The screening criteria include the following aspects:
There are at least 100 settled positions to ensure that the sample size is statistically significant;
Accounts must have been actively trading for at least four months, excluding those that won by luck in a single instance.
Participate in at least two different markets to avoid betting on a single event;
Total transaction volume exceeded $10,000, ensuring that participants actually invested their funds.
Under these conditions, after filtering the initial 112,000 wallets, only about 8,400 wallet addresses remained with sufficient data value. These 8,400 addresses are the truly meaningful dataset for research, rather than the "hero accounts" on the leaderboard that made millions of dollars with only a few transactions. These addresses share the characteristics of continuous transactions and stable data, making it easier to observe real behavioral patterns from them.
Interestingly, once the screening was complete, the truly most consistently performing traders were completely different from those on the leaderboard. They were unremarkable, and most people had never even heard of them.Their profits are typically between $50,000 and $500,000, rather than millions of dollars.
But what's truly noteworthy isn't how much money they made, but rather the trading process and methods behind it. Because what can truly be replicated is never the result, but the process.
Three common misconceptions that need to be dispelled
Myth 1: Top traders have a win rate between 80% and 90%.
This is not the case. The data is based on a filtered sample, not on the whale accounts on the leaderboard that made a fortune from a single bet.The winning rate of truly profitable wallets is mostly between 55% and 67%.In other words, even top traders make mistakes in a significant portion of their trades. For example, one address may have completed over 900 settled positions, accumulating a profit of $2.6 million, but its win rate is only 63%. In other words, more than a third of its bets were wrong, yet it still made a huge profit in the prediction market.
An obsession with winning percentage is often the most common pitfall for novice accounts.Many beginners like to buy contracts at $0.90 because it seems "safe." The probability of a "yes" outcome is already 90%, making the result seem almost certain, so they buy at $0.90, only to earn $0.10 if the event actually occurs. However, a single wrong prediction can result in a direct loss of $0.90, making the risk-reward ratio 9 to 1. If this pattern repeats many times, the account funds will quickly be depleted. In the dataset, this scenario has been repeatedly observed on hundreds of addresses.
Myth 2: The best traders can trade in any market.
The truth is quite the opposite.The best-performing wallets typically participate in a maximum of three market categories, and most even focus on only one or two areas.Some addresses only make predictions about cryptocurrency-related events; some only participate in weather-related markets; and there is even one address that almost exclusively trades questions like "whether Bitcoin will reach a certain price before Friday".
In prediction markets, excessive diversification often leads to a decline in the quality of judgments. Broad participants tend to perform poorly, while highly focused participants are more likely to consistently profit.
Myth 3: Speed determines everything
This statement holds true only in very rare cases. For example, some crypto markets with 15-minute settlements do require rapid reactions. However, in the vast majority of markets, top traders do not win by speed. Their more common approach is...Gradually build up positions over days or even weeks.They weren't in a rush to compete with others on click speed; instead, they patiently waited for a significant price deviation. When the price deviation was large enough, even if the market took two weeks to correct, the overall mathematical expectation would still be in their favor.
Five trading patterns worth learning
Pattern 1: Trading in the opposite direction during periods of extreme emotion
This is in the entire dataset.Most obvious, most stableThis is a profit signal. Among the 8,400 wallets screened, this behavior is almost the primary indicator for judging whether an account is profitable in the long term.
When a contract's price was driven up to 88% by market sentiment, many top wallets actually started selling YES; conversely, when the price dropped to around 12%, they began gradually buying in. Of course, this wasn't blindly contrarian trading, nor was it an attempt to oppose the market for its own sake. They only entered the market on a large scale when they judged that market sentiment was clearly overreacting.
The effectiveness of this strategy is related to a classic phenomenon known as the "hot/cold bias." This phenomenon was discovered as early as the 1940s in horse racing betting research and occurs in almost all markets where humans participate in betting. Simply put, people tend to overestimate outcomes that "almost certainly happen" while underestimating low-probability events.
Further analysis revealed that the top 50 most profitable wallets typically had an entry price that deviated from the market consensus probability by 6% to 11%.I won't bet when the odds are 50/50; instead, I'll patiently wait until the odds are clearly in my favor before entering the market.This trading method may seem boring, but it is stable and highly profitable in the long run.
Mode 2: Position management method is very similar to the Kelly Criterion.
Comparing the position sizes of the top 200 most profitable wallets with the "implicit advantages" they faced at the time reveals a very clear correlation. In other words,They don't bet randomly; the size of their bets varies almost proportionally to the perceived advantage they possess.In other words, when they believe they have a significant advantage, they will significantly increase their position size; when the advantage is small, they will only place a smaller position; and if there is no significant advantage, they will simply not trade.
It's hard to say whether these traders have actually read the Kelly Criterion or simply developed this intuition through long-term losses and practical experience. But mathematically speaking, their behavior closely resembles that of the Kelly Criterion.
The Kelly Criterion is usually written as: f* = (p × b − q) / bWhere: p represents the probability that the trader believes the event will actually occur; q = 1 − p; b represents the odds-to-reward ratio (potential gain ÷ risk cost).
For a simple example, suppose a trader believes an event has a 60% probability of occurring, and the market price is $0.45. The return ratio is: b = (1 / 0.45) − 1 ≈ 1.22. Substituting this into the formula, we get: f* = (0.60 × 1.22 − 0.40) / 1.22 ≈ 0.272. In other words, the complete Kelly strategy recommends allocating 27% of the capital to this trade.
However, this approach is extremely risky in actual trading, with very high volatility, and could potentially drag an account into a huge drawdown in a short period of time.Data shows that truly profitable wallets typically use a more conservative version, roughly one-quarter of the Kelly Criterion. In other words, if the full Kelly Criterion recommends betting 27%, they usually only bet around 7%.
In the most promising trading opportunities, the position size may be increased to 12% to 15%; for opportunities with moderate confidence, the position size is usually only 2% to 5%; and in markets where there is no clear advantage, they often choose not to participate at all.In contrast, losing accounts typically fall into two extremes. Either they bet 80% of their capital in a single trade, relying entirely on luck; or they spread their $10 across forty or fifty markets, believing they are "diversifying risk." In reality, however, this is more like constantly paying transaction fees to make the account appear busy.
Model 3: Highly Focused Professional Trading
After categorizing the 112,000 wallets according to the market categories they participate in, very clear differences emerge. These categories include crypto markets, political events, sports events, weather, geopolitics, entertainment, and science, among others. The analysis concludes that:
Wallets that participate in only 1 to 2 categories have an average PnL of approximately +4200 USD;
The average PnL for participating in 3 to 4 categories of wallets is approximately -$380;
Wallets participating in 5 or more categories have an average PnL of approximately -$2,100.
This relationship exhibits an almost linear trend.The more market categories you participate in, the higher the probability of incurring losses.
Different types of prediction markets rely on completely different information systems.Crypto markets are often influenced by factors such as exchange fund flows, whale addresses, and funding rates; political markets rely on poll data, grassroots news, and congressional schedules; while weather markets depend more on NOAA weather models, atmospheric data, and satellite observations.
The two cases are particularly representative.Case 1:Wallet A only trades in Bitcoin prediction markets with 15-minute settlements and never participates in other types of markets, such as "whether BTC will be higher than a certain price in the next 15 minutes." This address completed 502 predictions with a 98% win rate, accumulating profits of approximately $54,000. Its advantage is actually very simple: it continuously monitors the depth of the Binance order book and quickly trades when the Polymarket price lags by 10 to 30 seconds. In other words, it reuses this information difference of just a few seconds hundreds of times.
Case 2:Wallet B only participates in weather-related markets. Its trading strategy is straightforward: it reads NOAA's daily publicly released temperature forecasts and compares them with Polymarket's market pricing. If the market price deviates significantly from these supercomputer predictions, which have been optimized over decades, it immediately enters a trade. In the New York temperature forecast market alone, this address boasts a 94% accuracy rate.
It's important to emphasize that these individuals are not geniuses. The real key lies in their ability to identify a niche they understand better than the average polymarket participant and then consistently leverage that advantage.There were no frequent changes in strategy, nor was there any FOMO (Fear of Missing Out) due to market trends. Instead, the same logic was consistently applied, focusing on the same strengths.
Pattern 4: Trading price fluctuations, not event outcomes.
Most Polymarket users trade in a very simple way: they buy a contract and hold it until the event settlement, resulting in either a profit or a loss—a typical binary outcome.But top-tier wallets do things completely differently.Often, they buy at $0.40 and sell immediately when news or market sentiment pushes the price up to $0.65. They don't care whether the event actually happens; as long as the price has reflected the new information, they complete the trade and exit.
In the dataset, some of the best-performing addresses didn't even have any settled positions. They never held contracts until final settlement, instead continuously engaging in price mismatch swing trading. Statistical data shows...Top-tier wallets typically hold their holdings for only 18 to 72 hours on average.Wallets in the bottom 50% of profitability tend to hold their positions for an average period until settlement, sometimes even for weeks or months.
This doesn't mean that holding until settlement is always a mistake. Sometimes, when the judgment is very certain, holding long-term is indeed a better strategy. However, overall data shows that top wallets use their funds more proactively and flexibly than most people realize. They...Not someone who passively places bets, but a true trader..
Pattern 5: Always avoid breaking news
Our intuition tells us that the most astute funds should enter the market immediately upon the occurrence of unexpected events, such as military conflicts, election results, or major news events like the resignation of company executives. However, data shows that...Top-tier wallets often proactively avoid the period immediately following a news breakout.They typically wait for sentiment-driven funds to flood the market, causing prices to fluctuate wildly in a short period, before starting to trade once market sentiment has gradually stabilized.
Looking at the entire dataset, a very clear pattern emerges:The best trading opportunities often occur before the market takes notice of an event, or after market sentiment has overreacted.When everyone is discussing the same thing, it's often the worst time to enter the market. At this point, market prices are usually highly efficient, and the advantage you can gain is minimal.
Five operational suggestions
Choose a track and focus on it long-term.
Whether it's crypto, politics, weather, or sports, it's all fine, but you must choose the area you are most familiar with. For at least the next three months, trade only this type of market. No exceptions, and don't participate in other trending events on a whim. Even "casually betting on the election" can easily disrupt your original judgment system.
Record every prediction
Before each trade, write down several key data points, including your assessed probability, the current market price, the expected advantage, and the planned position size. Review these data after accumulating more than 50 trades. For example, if some predictions are marked as having a 70% probability, is the actual hit rate truly close to 70%? If there's a significant deviation, it indicates a bias in the probability assessment, and this must be recalibrated before increasing the position size.
Position management should be as close as possible to one-quarter of the Kelly Criterion.
First, calculate the theoretical position size given by the Kelly Criterion, then divide by 4 to get the actual position size. This number usually seems small, but it is crucial for controlling risk. Over-leveraging often results in only one outcome—account liquidation.
Trade only when the advantage is obvious enough.
If the expected advantage is less than 8% to 10%, abandon the trade immediately. Even if the opportunity seems tempting, learn to wait. The best-performing wallets in the data typically only make 2 to 3 trades per week per market category. Trade quality is far more important than trade quantity.
Keep records and review
Create a complete trading table to record every trade, its outcome, and any problems that arise. Wallets that consistently improve over the long term almost always systematically review their mistakes; while accounts that stagnate or even continuously lose money often simply repeat the same errors and attribute the results to bad luck.
