🔍 Read the full analysis: The Week Four Trading Test: 13,000 Runs Of A Viral Strategy on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get smart everyday buys delivered free with Prime

• Fast, free delivery on millions of items
• Prime Video, Amazon Music and more included
• Member-only deals all year

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

Thorsten Meyer AI says a two-day simulated test of a viral Polymarket trading strategy lost money in both of the versions it evaluated, across about 13,000 windows on BTC, ETH, SOL and XRP. The strategy’s central premise—a profit from buying both sides of a five-minute market at two cents each—produced a double-fill in just 3 of 9,486 attempts, according to the report.

The strategy was prompted by a YouTube video showing a trader turning $1 into $50 on a Polymarket five-minute BTC Up/Down market. The source report says the video included other large returns and that the transactions could be checked on PolygonScan. The video’s creator described the event as rare and said he did not yet have results showing whether it could be repeated.

In the reconstructed paired strategy, the tester placed two-cent bids on both Up and Down. Each winning share pays $1 at settlement, so if both 50-share bids filled for a total cost of $2, the winning side would return $50, leaving a net gain of $48 regardless of the market’s direction. But that result depends on both bids filling. The report says this happened three times among 9,486 paired attempts, or 0.032%.

The more common outcome was a fill on the losing side alone: 1,297 attempts, or 13.7%, according to the report. Neither side filled in 8,162 attempts, while the winning side alone filled in 22. The paired strategy recorded paper net P&L of −$280. A second version posted bids only on the apparent winner after the market closed. It recorded eight fills from 3,482 posts; the report says four fills settled and all four lost. Its reported figures do not establish how the strategy would perform with real orders or in other market conditions.

Rare Double-Fills Drove the Losses

The test addresses a key difference between a striking trade and a repeatable strategy: a large return on one transaction does not show how often the setup occurs, or whether losses outweigh wins over many attempts. Here, the reported double-fill—the event needed for the paired approach to lock in its intended profit—was uncommon, while one-sided fills on the losing leg were much more frequent.

The result matters to readers because the video’s apparent opportunity relied on stale orders remaining available around market close. According to the report, a trader can benefit if another participant’s outdated liquidity is still resting when the market moves. But the test suggests that, in the markets and period examined, attempts to capture those orders more often left the simulated strategy holding a losing position than secured both sides.

The findings are a paper-trading result, not evidence of a guaranteed outcome for all traders or future market periods. They also do not verify the circumstances behind every transaction in the original video. They do, however, show why the frequency of fills and the costs of unsuccessful attempts matter alongside the headline return.

From Viral Trade to Repeat Test

Polymarket’s five-minute crypto markets resolve as binary Up or Down outcomes. Under the strategy described in the report, the winning side redeems for $1 per share and the other side for nothing. The proposed paired trade aimed to buy both sides cheaply during a period when orders might not have been updated after the outcome became apparent.

The YouTube video supplied the initial example, but its creator reportedly acknowledged that the trade was rare and that repeatability had not been established. Thorsten Meyer AI’s May 21 report describes its work as Part 4 of an ongoing series and says the strategy was rebuilt in two forms inside Polybot. The test covered four coins over two days and used a simulated fill model in which a bid counted as filled only when a real taker sold into it.

The report describes 24 parallel experiments and roughly 13,000 windows overall. Its paired-strategy outcome table covers 9,486 pairs; the separate post-close strategy recorded 3,482 posts. These are different counts for different test versions, not one shared denominator.

““This is going to take time because this is a rare event… probably I didn’t have anything for you in this video about the results.””

— The YouTuber, as quoted in the Thorsten Meyer AI report

Limits of the Two-Day Simulation

The reported test used simulated money; no real funds were involved, according to the source. It therefore does not establish the results a trader would get using live capital, nor does it quantify real-world trading costs, execution constraints or how the strategy might behave over a longer period. The report’s paper-fill method required a real taker to sell into a simulated bid, but paper fills are not the same as completed live trades.

The post-close version also depended on the tester’s price reading immediately after a five-minute window ended. The report attributes its four settled losses to a mismatch between that reading and Polymarket’s official resolution, saying the final settlement price was not yet locked in when the test identified a winner. The article does not provide enough detail to independently assess the price feed, exact timing, full experiment logs or whether different rules would change the outcome.

The original video’s transactions are described as verifiable on PolygonScan, but the test does not show that the same specific trade circumstances were reproduced. The conclusion is limited to the two implementations, four coins, two-day period and reported simulation methodology.

Further Tests Would Need Live Validation

The source identifies this report as one installment in an ongoing series but does not give a date for another test or specify a next milestone. Any follow-up would need to clarify whether the researchers plan to extend the sample, publish more detailed experiment data or examine changes to the strategy.

For now, the reported conclusion is confined to the completed simulation: both tested forms were net-negative under the conditions described. The article says its findings should not be treated as financial advice and warns that using real funds could result in losses. No live-trading result or broader market-wide conclusion is confirmed by the source material.

Key Questions

What was the viral Polymarket strategy?

It attempted to buy both Up and Down shares in a five-minute crypto market at two cents each. If both sides filled, the winning shares would pay $1 apiece at settlement, creating the profit described in the report.

How often did both sides fill in the test?

The report says both bids filled in 3 of 9,486 paired attempts, a reported rate of 0.032%.

Did the simulated strategies make money?

No. The paired strategy recorded paper net P&L of −$280. In the post-close version, four fills settled and all four lost, according to the report.

Did the test use real money?

No. Thorsten Meyer AI says the strategies were run with simulated money and that no real funds were involved.

Does the test prove the strategy can never work?

No. It reports results for two implementations tested over two days on four coins. The findings do not establish what would happen in every market, over a longer period or with live trading.

Source: ThorstenMeyerAI.com