The author previously shared a self-built AI-assisted investment workstation that included several tools created through coding: a cross-market asset dashboard, an investment map, a personal content operations desk, and a Polymarket betting monitor that has recently been used frequently. Over more than half a month, the author tested the Polymarket dashboard with around $1,600 in principal and recorded a return of more than 30%. The real-time figures shown by the dashboard were broadly in line with the final actual net profit, with a difference of about 6U, mainly from small discrepancies such as open orders and liquidity rewards.
The author stresses that the point is not to say Polymarket is easy money, nor to present the experience as an arbitrage tutorial. After completing this round of testing, the author became more convinced that Polymarket is not a suitable place for anyone rushing in with a simple arbitrage mindset. The dashboard was started around May 21. The original need was basic: avoid opening more than ten betting pages again and again to check yes/no price changes, and stop using Excel to manually record trades, floating profit and loss, settlement dates and event categories.
From manual records to a structured betting dashboard
Before building the tool, the author had been using Excel to track every buy and sell decision, as well as unrealized gains and losses, settlement nodes and event types. The problem with that method, according to the author, is that many Polymarket positions can easily get out of control when the tracking system is weak. A small initial position can turn into repeated adding when odds move. If an event suddenly changes and the spreadsheet is not updated in time, the window for stop-loss or adding exposure can be missed. The author describes the entire process as too fragmented, and without a system, orders are easily driven by emotion.
The dashboard was therefore designed to place every bet back into one unified framework and turn a subjective feeling into information that can be visualized and compared across markets. After several iterations, it was divided into two tabs: a position dashboard and an opportunity monitor. The position dashboard became the core of the system. It can fetch real-time PM data and recalculate key metrics. For each market, it records the event name with a direct link to the trading page, the T1/T2/T3 classification, the current yes/no price, return, annualized return, abnormal price movement, observation nodes and a countdown to the bet’s expiration.
The abnormal movement alert can be customized. For example, if the price changes by more than 20% within 24 hours, the dashboard can show a pop-up alert as long as the webpage is open. The author highlights two designs that were especially useful. First, after finding a suitable PM interface, the user can paste in the webpage link for a betting event, and the system automatically reads the yes/no options, the corresponding prices and the categories of different options under the same event. This significantly reduces manual input. Second, the Tier classification of a given bet is automatically rearranged according to the number of days remaining.
One example involved Anthropic’s release of Mython. Before the release, the watchlist already showed a clear price movement. The author judged it as a high-probability deterministic event, and entering at that time delivered roughly 10 percentage points of return. In the author’s view, this kind of opportunity is difficult to capture steadily without a watchlist.
The binary market trap behind attractive expected returns
The author’s main reflection after the test is that binary markets such as PM contain a structural trap. They are unfriendly to players who prefer a single concentrated position, but more suitable for participants who are used to opening many small positions across a broad set of markets. The author illustrates this with a simple expected value calculation. If the yes price c of a betting event is 0.80, the market is roughly assigning an 80% probability to the event. If the author’s own estimate of the real probability q is 0.90, the expected return can be roughly calculated as EV = q / c - 1 = 0.90/0.80 - 1 = 12.5%.
That number looks attractive, but the author emphasizes that PM is not a bond. The 12.5% expected return hides sharp tail risk: if the judgment is wrong, the loss is not 12.5%, but 100%. For that reason, the dashboard does not look only at expected return. It also tracks two additional points. The first is the gap between the author’s probability estimate and the market price, namely q - c, which the author treats as the core of whether an edge truly exists. The author also sets an automatic take-profit reminder target at the midpoint between the entry price and 100. The second point is the impact on the total account if a single position goes to zero.
This is also the basis of the T1, T2 and T3 structure. T1 refers to high-conviction positions. The author’s comfort zone is mainly East Asia and certain geopolitical topics, especially cases where an information gap between Eastern and Western sources can be repeatedly checked before entry. T2 refers to relatively stable opportunities where the author feels that the current implied probability is obviously higher than the actual yes or no pricing. T3 is pure speculation, usually involving very high odds. These positions should not be held often; the author uses them more for counter-move trades and short-term price reversion.
Even T1 has hidden costs. A T1 bet can show a static return of 18%, but if it settles 180 days later, the annualized IRR can be only 3–4%. In that case, the principal is locked up for a long time, and later high-IRR opportunities can be missed. The author therefore further divides T1 positions into time buckets. Short-term T1-A can be allocated more heavily, while long-term T1-C needs restraint. Too many long-duration, low-IRR positions create a hidden loss of capital efficiency. For T2, even when an edge exists, room must be left for wrong judgment. The single-bet limit is 8–10%, so even a total loss on that position keeps the total account loss within 10% and does not prevent participation in later opportunities. T3 is observed with the smallest position size, not as a path to large profit, but as a way to track high-odds events and build a feel for that type of market.
High conviction is not the same as a large position
The author’s position limit is ultimately a way to make the cost of being wrong bearable. One important and counterintuitive point is that high conviction is not equal to high position size. Even if an event is believed to have a 95% chance of occurring, as long as there is still a 5% chance of going to zero, the position has to be limited. In an extreme example, if someone makes 10 independent bets that each appear to have a 95% win rate, each individual bet sounds stable. But as long as they are independent, the probability of at least one being wrong is around 1 - 0.95^10 ≈ 40%.
That is only the independent-event case. In reality, many PM markets are not independent. They are often correlated. The author gives three examples: whether U.S.-Iran talks reach an agreement, whether the Strait of Hormuz reopens, and whether the Middle East situation escalates during the month. These look like three separate markets, but the underlying variable is almost the same: the direction of Middle East geopolitical policy. If that direction is judged wrongly, all three positions can bleed at the same time. For the author, the biggest help from the dashboard was not improving the win rate, but limiting large mistakes. In plain terms, the core value of the dashboard is not profit statistics, but risk control.
Why the author does not treat PM as an arbitrage venue
After more than half a month of intensive testing, the author’s strongest feeling is that opportunities do exist on Polymarket, but it is not the arbitrage field many people imagine. On-chain arbitrage often has clear rules and lockable price dislocations. Polymarket is different. It tests the user’s logical understanding of how narratives and direction change across different betting events. For East Asia-related political and economic dynamics, Chinese-language users can indeed have some information-gap advantage, and the author thinks that angle can be explored. But that does not guarantee a win.
Polymarket settles according to market rules and specified data sources, not according to the user’s understanding of reality. The author also notes that UMA manipulation problems have been seen repeatedly. A fact that looks settled in a Chinese-language context is not necessarily defined the same way in English rules. The rules of each betting market often contain textual traps. Based on the author’s actual experience, PM does not have many arbitrage opportunities. The main basis is still information gaps and diversified position sizing. Even high-conviction trades can meet black swans, and when that happens, the principal is fully lost.
The author quotes a friend: "In investment, even if there is only a 1% probability of going to zero, one should not rely on luck." The author’s current understanding of PM is therefore more conservative. First, it should not be treated as a stable income tool. Even for high-conviction positions, especially after several consecutive wins, users should not assume they have found an ATM. The most dangerous part of a binary market is that it can make a participant believe, after a winning streak, that every judgment is correct; the final position then becomes large enough to return all earlier profits. Second, a high win rate is not the same as a good trade. If a 90% win-rate event is already priced at 0.95, it can be negative expected value. Conversely, if a 40% win-rate event is priced at only 0.20, it can be positive expected value.
Third, tail risk should not be ignored. Many people see 10% or 20% returns and feel the position is stable, but if a wrong outcome means the position goes to zero, it is not a low-risk return in the traditional sense. From this angle, the author even says there is no so-called low-risk opportunity on PM; every one is high risk. Fourth, false diversification should be avoided. Buying several different markets does not necessarily count as diversification. The three Middle East-related examples mentioned earlier look like separate markets, but their underlying variable is nearly identical.
The author now prefers to treat PM as a training ground for judgment. It fits with the political, economic, technology and financial information the author already reads daily, turning judgments that used to remain at the level of "I think" into something that can receive feedback. The author believes these abilities are useful outside PM as well.
In addition to the PM betting dashboard, the author also used Codex to create a private-market valuation monitoring dashboard. It mainly tracks valuation changes in unlisted unicorn companies such as Anthropic, OpenAI, Stripe and Kraken, and observes the relationship between those changes and corresponding PM bets. Polymarket is essentially an expectations market. Sometimes signals in the private market have already changed while PM prices have not moved; at other times PM prices move first while real-world data has not yet followed. The mismatch between the two is described as an interesting observation framework. However, the author again emphasizes that this is not risk-free arbitrage, because private-market valuations are not fully transparent and different data sources can differ.
The central message of the article is not that the author made 30% with a dashboard and others can do the same. The more useful lesson, in the author’s view, is that a tool can help turn feelings into a framework and a framework into discipline. Many times, making money does not mean a secret method has been found. It only means that the judgment in this specific round happened to be right.