After earning $200 passively, I stepped into these three "weather deep pits" in the prediction market.

Byline: Changan I Biteye Content Team

Abstract: A roundup of 4 hands-on strategies and 2 typical counterexamples—showing you the underlying logic behind weather market “weather games.”

In the previous post, the weather system used five methods to predict the highest temperature, but the model won’t always give a single solution forever: ECMWF comes up with 14°C, WC comes up with 13°C, and real-time corrections give 13.5°C—so which tier do you bet on?

Even with a more refined weather system, predictions will always remain a range.

What’s more, the weather market has risks beyond prediction: the data source may not match, the rules may quietly change, and the price action may flip in the final hour.

So prediction must be paired with a trading strategy. After two weeks of live trading, there are winners—and also people who stepped into traps. I’ll share both.

1. Besides prediction, there are these three traps

The previous weather system solved the prediction problem, but once you enter the market, you realize that the reasons for losing money sometimes have nothing to do with prediction at all.

Mismatched data sources, rules quietly changing, price action flipping in the last hour—only after you’ve stepped into these traps do you understand that the risks in the weather market are not just on the prediction layer. There are three layers in total:

1. Data source issue: WU and METAR don’t match

The weather market rules are generally written to be based on WU data.

WU is Weather Underground, an American weather platform. Its data comes directly from observation records reported by weather stations around the world. For the weather market, WU is reading data from the local airport weather station.

Every half hour, the airport weather station sends out a standardized meteorological bulletin called METAR—a format commonly used across global civil aviation. It contains information like temperature, wind speed, cloud cover, visibility, and more, serving as an important basis for air traffic control. In theory, the airport temperature shown by WU should come from that METAR message.

In other cities, this is indeed the case: WU readings and METAR match closely, with errors so small they can be ignored. Many traders develop a habit because of this: directly watch METAR and treat it as a real-time preview of the settlement temperature.

But in Shenzhen Bao’an Airport, WU data often doesn’t match METAR, and there have even been cases where it differed by 2 degrees.

This deviation wouldn’t matter in other cities at all, but in Shenzhen it can directly turn a trade that should have been correct into a losing one.

2. Rules frequently change

Maybe it’s precisely because WU and METAR have been mismatched for a long time that on March 29, Polymarket replaced the data source for the Shenzhen market. The settlement data source changed from WU to NOAA.

On the Rules page, there’s an update log entry: the time is March 28, and it’s only one sentence: “This market’s resolution source has been updated.”

NOAA uses data from weather.gov, which corresponds closely to METAR, and often differs from WU readings.

There’s an ilovebigbiscuit address that bet No in the 27°C tier, with an average price of 99.8¢. They lost $7,883. Most likely, WU’s displayed temperature didn’t reach 27°C, so they thought it was safe. Then the NOAA reading was different, and they sold the close-to-the-end positions to end up losing it all right before settlement.

Later, the Shenzhen weather market switched back to WU. In a single market, in just a few days, they changed the data source twice.

So develop a habit: every time you enter a new weather market, the first thing isn’t to look at the temperature—it’s to open the Rules in the bottom-left corner and confirm which data source will be used for settlement. If you skip that step, then no matter how much analysis you do afterward, it may all be wasted.

3. Whipsaw price action

Over these past two weeks, the Shanghai weather market repeatedly showed the same kind of move: the temperature in a certain tier started leading from the morning, with its probability steadily holding ahead of the other tiers. As settlement looked close, in the final hour it suddenly reversed. Another tier surged from near 10% straight to 100%.

It’s like in the day shown in the chart: 20°C was the main tier all morning, and by 2:00 PM it rose to nearly 90%, making it look like the outcome was already decided. But after 3:00 PM, 21°C flipped from near 0% straight to 100%, and the final settlement was at 21°C. People who bet on 20°C were correct until before the last hour—but they were all wrong at settlement.

Shanghai’s spring weather is inherently unstable. Temperature trends in the afternoon are heavily influenced by cloud cover and wind speed, so the judgment formed in the morning may become completely invalid by the afternoon.

2. Four practical playbooks observed over two weeks

Betting on a single temperature is too hard to get right, so most players buy multiple adjacent tiers at the same time. As long as the combined total cost of these tiers is no more than $1, there’s potential profit. But even with multi-temperature coverage, the outcome can vary wildly depending on when to buy, how to buy, and which market to buy. Here are several trading strategies observed over these two weeks:

Strategy 1: Buy low-priced chips a few days early

Another way of thinking is completely different—it leverages uncertainty in the weather itself.

Polymarket weather markets open for trading four days in advance. The earlier the market opens, the more dispersed the pricing across tiers becomes. For many temperature probabilities, they haven’t yet been priced in fully, and you can still find orders sitting below 5¢.

Some players specialize in exploiting this timing gap. Today is April 1. They go buy the weather market for April 4, sweeping all the tiers priced below 5¢. As long as it’s cheap enough, they buy in. The logic is simple: the forecast is still three days away. The weather may change at any time. A temperature that looks impossible today may become a popular tier in two days, and those 5¢ chips could rise to 30¢, 50¢, or even higher.

The core of this strategy is betting that the weather will change. You hold the position all the way to that day. As long as you cover enough tiers and keep the total cost within $1, then at settlement the tier that resolves will recoup $1 while the rest go to zero—so overall you don’t lose money. But if one tier spikes significantly in the middle, you can also sell early to lock in profit.

Strategy 2: Use meteorological factors to capture undervalued temperatures

Popular tiers are usually already priced in well. Buying them costs more, and the odds are compressed very low. But there’s a group of traders who do the opposite: they look specifically for undervalued, unpopular cold temperature tiers.

These traders observe real-time meteorological factors. For example, if it’s currently 1 PM, they’ll look at the wind direction and wind speed over the next one or two hours. Southerly winds often bring warm, moist air with potential for warming. If factors like wind speed, cloud cover, and air pressure all point in the same direction, they’ll bet on the temperature tier that the market hasn’t reacted to yet.

Because it’s an unpopular tier, the price is low and the entry cost is small. Even if they’re wrong, the loss is limited. But if they’re right, the upside on low-priced chips can be very substantial.

WU data updates every half hour. If the latest data shows that the meteorological factors aren’t developing as expected—for example, if the wind direction shifts or warming stalls—then they sell immediately to cut the loss.

This strategy requires a relatively high level of meteorological knowledge. You need to truly understand the mechanisms of how factors like wind direction and cloud cover affect temperature. It’s not something you can judge just by glancing at a forecast. It’s best suited to traders with professional backgrounds or those who have spent some time deeply trading this market.

Strategy 3: The end-of-day (tail) strategy

In the Shanghai weather market, there’s a pattern: after 3 PM, temperature basically stops rising. The highest temperature usually appears before then.

The tail strategy uses this window. WU data updates every half hour. After warming stops, you watch the WU data on each half-hour update. You enter at the exact moment the data just refreshes. At that point, Polymarket’s price hasn’t had time to react yet—so usually there are only a few percentage points of profit to capture.

There are two directions for the trade: buy Yes on the current temperature, or buy No on the next higher temperature. These two actions are essentially the same. Since warming has stopped, only these two possibilities remain. Which one you choose depends on which side has the more favorable price.

The biggest risk of this strategy is changes during the warming period. The time when warming stops isn’t fixed every day. Cloud cover, wind speed, and air mass all affect the day’s warming rhythm. If you think warming has ended but later the temperature rises again, then your tail judgment will be completely invalid.

Strategy 4: New market limit-order strategy

Polymarket’s new weather markets have an obvious feature: there are no market makers, so the spread is large. The difference between the bid and ask can be tens of percentage points—for example, the bid is 20¢ and the ask is 60¢.

That spread is the opportunity.

The specific execution is to place limit orders on the bid side below the popular temperature tiers, waiting for those buy orders to get filled. Because the spread is large, even if you place orders to pick up several popular tiers, your total cost can still be controlled below $1.

But this kind of market has one must-notice characteristic: extremely low liquidity. Trading just a few hundred dollars can push down the probability of the currently most popular temperature enough to make it look like it’s about to lose.

So the execution principle for this strategy is simple: after your limit orders fill, hold continuously until settlement—don’t operate frequently. Short-term price fluctuations don’t have much reference value in low-liquidity markets.

Strategy 5: Entering positions one day early (counterexample)

WU publishes a forecast for the next day’s highest temperature one day in advance. The most intuitive way to execute is to reference that forecast and buy the three tiers around the forecast temperature one day early.

That’s exactly what some players did. On March 27, they bought 15°C, 16°C, and 17°C. On March 28, they bought 19°C, 20°C, and 22°C. They put a few hundred dollars into each tier, keeping total cost within $1.

But weather forecasts aren’t fixed. WU’s predicted temperature data adjusts in real time as the weather changes. Today it forecasts tomorrow at 22°C, but tomorrow morning it may already be changed to 19°C. Building the position one day early locks in last night’s forecast—but by settlement, the actual temperature has already drifted.

In the end, everything settled to zero. March 27 settled at 19°C, and March 28 settled at 21°C. The covered range differed from the actual settlement by about two or three degrees.

The idea behind multi-temperature coverage is correct, but the entry time was too early. When the forecast hadn’t stabilized yet, they entered positions—essentially using yesterday’s information to bet on today’s outcome.

Strategy 6: The “buying No” win-rate trap (counterexample)

In the comments, someone said buying Yes is too hard to guess—if you can’t pick any of them, you might as well buy No because the win rate is higher. But can it really be that way?

Weather markets usually have 11 temperature tiers. Buying Yes means guessing 1 out of 11. Buying No means guessing 10 out of 11. So on paper, No has a natural win-rate advantage, and it sounds reasonable.

However, No for popular tiers is usually listed at 80¢ or higher, and only a few tiers have prices like that. If you buy all No positions above 80¢, assume there are 4 popular tiers:

Cost: 4 × 80¢ = $3.20

Settlement is likely to fall into one of those 4 popular tiers. That means 3 of the No positions win and 1 No position loses:

Win: 3 × 20¢ = 60¢

Lose: 1 × 80¢ = 80¢

Net result: lose 20¢

Yes, the win rate is higher—but the odds fully wipe out the win-rate advantage. Each time you win, you make only a little. Each time you guess wrong, the loss can cover several rounds of profit. The price of No already incorporates the win rate. Buying No doesn’t give you any extra edge.

3. Prediction and strategy—both are indispensable

Your prediction system is your “eyes,” and your trading strategy is your “armor.” Together, that’s what makes you unstoppable in weather market trading.

The weather market is still early-stage: rules are unstable, data sources can change, and price action can continuously reverse. But precisely because of that, there’s still an information gap—and opportunities still exist.

If you’re also trading weather markets, feel free to share your trading approach and the traps you’ve stepped into in the comments.