🔍 Read the full analysis: Week Four’s 13,000-Test AI Trial Of A Viral Trading Strategy on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Thorsten Meyer AI reports that a two-day paper-trading test of a viral Polymarket strategy ran across four crypto markets and about 13,000 windows. Both tested versions lost money; the key paired-order strategy filled both sides just three times in 9,486 attempts.
Thorsten Meyer AI says a two-day simulated test of a viral Polymarket trading strategy lost money across two versions, four cryptocurrencies and about 13,000 market windows. In the main test, the supposed guaranteed-profit setup filled both sides only three times in 9,486 attempts, while one-sided fills more often left the simulated bot holding the losing position.
The test followed a YouTuber’s video showing a trade in Polymarket’s five-minute Bitcoin Up/Down market that reportedly turned $1 into $50, with other examples described as 48-fold and 100-fold returns. The source says the transactions were verifiable on PolygonScan. The video’s creator acknowledged that the examples did not establish whether the trade could be repeated, saying, “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.”
Thorsten Meyer AI reconstructed the strategy in two forms using its Polybot system. The paired version placed simulated bids for both Up and Down at two cents per share. If 50 shares on each side filled at that price, the source says, one side would pay $50 at settlement, leaving a $48 net gain before any other costs. That outcome depends on both orders filling; if only the losing side fills, the position can lose its cost.
The source reports three double-fills among 9,486 paired attempts, 22 winner-only fills, 1,297 loser-only fills and 8,162 attempts with no fill. The reported paper result was a net loss of $280. A second version posted only on the side the system judged had already won: of 3,482 posts, eight filled, and the four fills with completed settlements all lost. The source says BTC and XRP had no fills in that version.
Week Four · Simulated strategy report
Week Four’s 13,000-Test AI Trial Of A Viral Trading Strategy
A two-day paper-trading test across four crypto markets found losses in both reconstructed approaches. The paired orders that depended on filling both sides did so just three times.
Simulated money only. No real funds were used at any point.
From viral trade to repeatable test
A YouTube video showed a reported Polymarket five-minute Bitcoin trade turning $1 into $50, with other examples described as 48-fold and 100-fold returns. On-chain records can verify that a transaction happened; they do not show how often the setup repeats or whether both sides can be acquired profitably.
“This is going to take time because this is a rare event…”
The YouTuber, as quoted in the reportWhy both sides mattered
In the described binary market, one side pays $1 per share and the other pays nothing. Buying 50 shares on both sides at $0.02 each costs $2; if both fill, the winning side’s $50 payout leaves $48 before other costs.
The intended double-fill was rare
The bot placed simulated bids for both Up and Down at two cents per share. The strategy’s payoff depended on both orders filling. One-sided fills could leave it holding the losing position.
Bars show counts relative to 8,162, the largest category. Categories sum to 9,484; the report states 9,486 paired attempts.
Reading the winner after close
The second version posted only on the side its price feed judged had already won. The report says the feed could read the market before the final settlement price was locked, creating mismatches with Polymarket’s official resolution.
| Strategy | Posts / attempts | Fills | Settled outcomes | Reported result |
|---|---|---|---|---|
| Paired Up + Down bids at $0.02 | 9,486 attempts | 3 double-fills; many one-sided | Not detailed here | −$280 |
| Bid on apparent winner after close | 3,482 posts | 8 fills | 4 settled; all lost | Losses reported |
Double-fills across the paired test’s 9,486 attempts.
Settled winner-only fills lost, according to the report.
BTC and XRP had no fills in the winner-only version.
Useful evidence, with clear limits
Counts matter
A dramatic individual return does not establish repeatability. The report counts unsuccessful attempts alongside unusual wins.
Timing can decide outcomes
Fill probability, order execution and final settlement may matter as much as the direction of a price move.
Future performance
Two days of simulated trading do not establish how the strategy performs over longer periods or under different liquidity conditions.
Methods need more detail
The supplied material does not include independent replication or enough information to assess fees, slippage, order priority, latency and all timing rules.
Paper test, not live results
The reported $280 loss and fill counts are the source’s simulated results. They are not independently verified live-market outcomes.
What would strengthen the evidence?
Any follow-up would be easier to assess with a clearly stated test period, execution assumptions, cost treatment and side-by-side settlement records. The material identifies this as Week Four, but does not specify the scope or timing of a next installment.
Show how results change across a longer range of market conditions.
Publish assumptions for fees, slippage, queue priority and latency.
Document feed readings and official resolution timestamps for each position.
Make the method and data available for independent replication.
“Every strategy described here was run with simulated money. No real funds were involved at any point.”
Thorsten Meyer AI · Report statementWhy Rare Fills Matter
The test addresses the gap between a striking transaction and a repeatable trading method. A large return from a single trade can be genuine without showing that the same opportunity occurs often enough to overcome losing fills. Here, the source’s central finding is that the paired orders’ intended double-fill was rare, while one-sided fills were substantially more common.
The second test also points to a different risk: identifying a market’s likely winner immediately after a window ends may not match the platform’s eventual settlement. According to the report, all four filled positions with settled outcomes lost despite the system bidding on its price feed’s apparent winner. That result is limited to this simulation, but it illustrates why settlement timing and execution can matter as much as the trading idea itself.
The findings are not a guarantee about every trader, market condition or future order book. They are a reported result from a specific, short test. The study’s practical relevance is that it counts unsuccessful attempts as well as the unusual wins, rather than treating a viral return as proof of an ongoing edge.
As an affiliate, we earn on qualifying purchases.
Polymarket’s five-minute crypto markets resolve as binary Up/Down outcomes: one side pays $1 per share and the other pays nothing, according to the source’s explanation. The proposed paired strategy aimed to buy both sides cheaply and profit whichever direction the market moved. Its premise was not simply that prices would move favorably, but that stale orders might remain available near the close and be filled by a taker.
In the YouTube example, the large return came from a trade that the source describes as real and verifiable on-chain. But a transaction record can show that a trade occurred without showing how often similar orders fill, whether both sides can be acquired, or how the approach performs over many attempts. The follow-up report says it tested the reconstructed method on BTC, ETH, SOL and XRP against the live order book for two days.
The report describes its paper-fill model as requiring a real taker to sell into a simulated bid. It says the work covered 24 parallel experiments and roughly 13,000 windows. No real funds were used, and the source explicitly labels the results as simulated rather than live trading performance.
““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 by Thorsten Meyer AI
As an affiliate, we earn on qualifying purchases.
Limits of the Simulation
The report is a publication by Thorsten Meyer AI, and the supplied material does not include independent replication or a separate review of its data and methods. Its results cover two days of simulated trading on four markets; they do not establish how the strategy would perform over longer periods or under different liquidity conditions.
The source does not provide enough detail here to independently assess every modeling choice, including fees, slippage, order priority, latency, or the exact timing rules used in each experiment. It also does not state whether the YouTuber’s specific trade used the same execution conditions as the reconstruction. The reported $280 loss and fill counts should therefore be understood as the source’s paper-test results, not verified live-market outcomes.
For the winner-only version, the report attributes the settled losses to a mismatch between its price feed’s reading just after close and Polymarket’s official resolution. The supplied material does not give the detailed timestamps or resolution records needed to independently examine each mismatch.
As an affiliate, we earn on qualifying purchases.
Further Testing and Verification
The source presents this as Week Four of an ongoing series, but the supplied material does not name a publication date or planned scope for the next installment beyond listing this report as published May 21, 2026. It does not announce a revised strategy, a longer test, or a live-money trial.
Any follow-up would need to make its test period, execution assumptions, cost treatment and settlement comparisons clear to show whether the findings hold beyond this two-day simulation. Until further evidence is published, the confirmed development is the report’s own result: both reconstructed strategies lost money in the stated paper test, while the test was not conducted with real funds.
As an affiliate, we earn on qualifying purchases.
Key Questions
What strategy did the report test?
It tested a paired-order approach that bid on both Up and Down shares at two cents, aiming to profit if both sides filled, plus a second version that bid only on the side its price feed judged had won after the market window closed.
How often did both paired orders fill?
Thorsten Meyer AI reports that both sides filled three times in 9,486 attempts. It also records 1,297 loser-only fills, 22 winner-only fills and 8,162 attempts with no fill.
Did the test use real money?
No. The source says the strategies were tested with simulated money and that no real funds were used.
Why did the winner-only version lose?
The report says its price feed sometimes disagreed with Polymarket’s official resolution because the final settlement price was not yet locked when the system read the market. It reports that all four fills with settled outcomes lost.
Does the report prove the strategy can never work?
No. It reports losses under a specific two-day simulation across four coins. The material does not establish results for every market condition or future test, and it does not provide independent replication.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
