backtesting / ai trading
A $20K Trading Claim, Put to the Test
A slideshow of the real NexusTrade fact-check run. A TikTok trader claimed a $20,000 month; the closest mechanical VWAP-short proxy lost 49.8% over the July 4–October 1, 2026 simulation, with 476 trades, a 42.4% win rate, and heavy fees. This does not establish what her discretionary trades actually earned.
Transcript
0:00I gave my AI a trading guru's TikTok. It backtested her $20K month. Swipe left for more.
0:03The real run at 3x speed: the NexusTrade agent plans the fact check, then ingests the TikTok video and keeps its transcript and metadata as a source.
0:08A trader on TikTok says she made $20,000 in one month day trading. I pasted the link into NexusTrade and asked one thing: can you test whether this actually works? It watched the video. It listened to it. It turned her claim into rules a computer can run. I didn't write a line of code.
0:14The closest version of her strategy: -49.8%. A backtest of the TikTok VWAP short proxy from July 4 to October 1, 2026 shows a total return of -49.8%, a Sharpe ratio of -14.64 and a max drawdown of -49.8%, over 476 trades with a 42.4% win rate while SPY was up over the same 90 days.
0:18The verdict: her $20K month is unverified. My closest mechanical version of her VWAP short lost 49.81% on $100K from July 4 to October 1, with 476 trades, a 42.4% win rate and heavy fees, while SPY was up. That does not prove she did not have one green discretionary month, but her $20,773 screenshot has no dates, capital, size or trade log, so the claim stays unverified and this test gives no reason to believe it repeats.
0:23For 20 years gurus sold ideas you couldn't test. Now you can: paste a TikTok link into NexusTrade and type “fact check this video”.
0:27Follow to see which guru gets backtested next. nexustrade.io
Join the conversation
Loading conversation…