If you've yet to integrate AI in your trading, this session just might be the one for you 😉

That's a wrap for tonight's session! Thank you everyone for attending 🫶
Recording and the resources for tonight's session will be uploaded soon~
Good Night Tribe🌙

Hey Tribe, Eri here 🫶

Last night Charmain Tan from QD Academy showed us how she actually uses AI for quant trading. Here's the technical shape of it 👇

🧩 Three levels of AI: Chat, Skill, Agent. Chat is asking. Skill is the capability you give the AI. An Agent completes a task. Agentic AI completes the whole process toward an objective, like a target P&L.

⚙️ IBOT is a chain of specialists, not one prompt: ideation → quality gate → data engineering → backtest → optimisation → trading. Each specialist has a personality and hands off to the next, and the quality gate only returns conditional, pass or fail.

🧪 Six-stage research framework: know your data → variance exploration → temporal structure → cross-sectional patterns → falsifiable hypothesis → pre-backtest validation.

📐 A hypothesis must specify five things: signal, universe, condition, prediction, horizon. Most hypotheses die at pre-backtest validation, before a backtest is ever run.

🔬 Backtest spec: in-sample to build, out-of-sample to validate, commissions and slippage accounted for, look-ahead bias-free. Vectorise first; use bar-by-bar loops only where the logic is genuinely stateful.

📈 Optimisation: systematic grid search, visual heatmaps and overfitting guardrails. Robust shows a plateau of good performance across neighbouring parameters; fragile shows a single spike.

🧾 Integrity report: every issue detected and fixed during the scan, documented before the strategy moves forward.

📊 Performance: CAGR, max drawdown, Sharpe ratio, equity curve, win rate and trade count.

That's the session in one screen 🫶

Here is the recording for last night's session:
https://youtu.be/YKBZcyPlCM0

Here is the slide deck!

[3 red days in a row on SPY… the backtest says that's when to BUY 🟥🟥🟥➡️🟩]

📊 298 trades since 1993
✅ About 74% of them closed as winners
⏱️ Average hold: just 1.6 days
🛡️ Worst drop: -11%, while SPY itself fell -55%

Most people panic after three down days… this rule does the opposite 👀

🎥 Watch the Reel and comment STRATEGY on Instagram (not here) to get the full rules in your DMs 👉 https://www.instagram.com/reel/Dev3e2wy5px/

🔥 Want more strategies like this? Hit 🔥 below!

Backtested results. Not financial advice.

[+671% in a crash year... with just 3 simple strategies 🤯📉➡️📈]

🏆 A four-time trading world champion, +671% in 2008
🥣 Just 3 strategies: trend following, breakout, mean reversion
⏱️ Same gold bot: 1-minute bars lost money, 1-hour bars lost less
✅ On 4-hour bars, it made money

Most people keep tweaking the strategy... the backtest says change the timeframe instead 👀

🎥 Watch the Reel and comment ROBUST on Instagram (not here) to get the free Just-Right Strategy Card in your DMs 👉 https://www.instagram.com/reel/Dewa59jCihk/

🔥 Want more breakdowns like this? Hit 🔥 below!

🦆 Explained by Prof Quack, our duck who turns quant lessons into bite-size reels.

Backtested results. Not financial advice.

[A 1970s indicator that still spots dips: 202 backtested trades, 79% winners 📜📈]

📊 202 trades on SPY since 1993
✅ Roughly 4 out of every 5 closed as winners
⏱️ Average hold: just 3.7 days, in the market only about 9% of the time
🛡️ Worst drop: -8%, while SPY itself fell -55%

An indicator older than most traders… and the backtest says it still works 👀

🎥 Watch the Reel and comment STRATEGY on Instagram (not here) to get the full rules in your DMs 👉 https://www.instagram.com/reel/DeycC6PSV4-/

🔥 Want more strategies like this? Hit 🔥 below!

Backtested results. Not financial advice.

[$2 trillion has to trade on the same days every month. Quants mark it on their calendar 📅🐋]

🏦 Pension funds must hold 60% stocks and 40% bonds, no matter what

📅 So every month end, they're forced to buy or sell, on schedule

🔍 Researchers told 16 of them: "We can see your trade coming."

🤷 Their answer: "We know. We're not changing."

Most traders chase the news. This trade is on the calendar 😏

🎥 Watch the Reel and comment WHALE on Instagram (not here) to get the Whale Bot Prompt in your DMs 👉 https://www.instagram.com/reel/DezArHmAHb6/

🔥 Want more trades the big funds can't hide? Hit 🔥 below!

🦆 Prof Quack, our resident quant duck, digs up the research so you don't have to.

Backtested results. Not financial advice.

[That "genius" stock-picking edge? It might just be two hidden factors in disguise 🎭]

📊 US stocks sorted into 25 portfolios by size and value, 1963 to 1991
🐷 Cheap "value" stocks earned an extra 0.40% a month on average
🎯 With market, size and value combined, 21 of 25 portfolios had R² above 0.9
😮 With the market alone? Just 2 of 25

A mouse, a piggy bank, and the 1993 study that changed how quants spot real alpha 👀

🎥 Watch the Reel and comment PAPER on Instagram (not here) to get the full paper in your DMs 👉 https://www.instagram.com/reel/De1AK1bS-iA/

🔥 Want more research breakdowns like this? Hit 🔥 below!

Educational content. Not financial advice.

[Buy low, sell high... but when is low LOW ENOUGH? 📉➡️📈]

📉 Fast RSI falls 3 days in a row and closes below 30
📈 Only while SPY is still above its 200-day average
🎯 91 trades over 32 years, about 81% winners
⏱️ Median hold: about 6 days

Everyone says "buy the dip"... this one tells you exactly when 👀

🎥 Watch the Reel and comment DIP on Instagram (not here) to get the AI Prompt that builds the bot in your DMs 👉 https://www.instagram.com/reel/De1kufQir18/

🔥 Want more strategies like this? Hit 🔥 below!

🦆 Prof Quack, our resident quant duck, digs up the research so you don't have to.

Backtested results. Not financial advice.