Quant Science Profile picture
Mar 18 9 tweets 3 min read Read on X
This strategy has delivered 18% annual returns since 1926.

No black box. No complex ML model.

Just breakouts, trailing stops, and volatility sizing applied to industry portfolios.

Here's how it works (and link to the 37 page PDF): Image
─────────────────
🔹 THE ENTRIES
─────────────────
A long position triggers when an industry breaks above either:
• A Donchian Channel (20-day high, 40-day lower band)
• A Keltner Channel (20-day EMA ± 1.4× ATR, 40-day lower band)

The asymmetric lookback keeps you invested during sustained trends.
No short positions. When nothing is trending, capital sits in T-bills.
─────────────────
🔹 THE EXITS
─────────────────
Trailing stop = the higher of the 40-day Donchian or Keltner lower band.
Once it moves up, it never moves back down.

Winners run. Losers get cut fast.
─────────────────
🔹 THE SIZING
─────────────────
Volatility-targeted. Each position is sized inversely to its 14-day volatility.
More volatile industry = smaller position.
Total exposure capped at 200%.
─────────────────
🔹 THE RESULTS (1926–2024)
─────────────────
• 18.2% annualized return vs. 9.7% for passive equities
• Sharpe ratio of 1.39 vs. 0.63 for the market
• Max drawdown of 33% vs. 84% for equities
• Upside beta of 1.16, downside beta of 0.31
Get the 37 page PDF here:

🚨Still trading manually in 2026?

You’re not alone. But the window to change that is closing.download.ssrn.com/2025/10/2/4857…
In 2026, traders are splitting into two groups:

Track 1: Manual traders. Chart watching. Indicator stacking. Hoping it works.

Track 2: Algorithmic operators. Tested strategies. Automated execution. Running while they sleep.
I'm running a free live workshop that shows you the bridge from Track 1 → Track 2. Live Python demo + Q&A. 500 seats.

👉Register here (500 seats):  learn.quantscience.io/become-a-pro-q…Image
P.S. - Want Algorithmic Trading with Python tutorials every Sunday?

Register here to join our Sunday Quant Scientist Newsletter (it's free): learn.quantscience.io/quant-scientis…

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Quant Science

Quant Science Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @quantscience_

Mar 18
A Bloomberg Terminal costs $30,000 a year.

Here's how to build 90% of it for free.

Wall Street doesn't advertise this. But every function that matters has a free alternative in Python.

Here's the DIY version (with Python): Image
━━━━━━━━━━━━━━━━━
🔹 REAL-TIME PRICES & CHARTS
━━━━━━━━━━━━━━━━━
Bloomberg: GP
Python: yfinance + Plotly

yfinance pulls live and historical price data for any ticker in seconds. Plotly turns it into interactive charts. Add your own indicators, overlays, and alerts. No subscription needed.
━━━━━━━━━━━━━━━━━
🔹 FINANCIAL STATEMENTS & RATIOS
━━━━━━━━━━━━━━━━━
Bloomberg: FA
Python: yfinance + pandas

yfinance exposes full income statements, balance sheets, and cash flow statements directly. Pull any company's financials into a DataFrame, calculate your own ratios, and build custom models — all in a notebook.
Read 11 tweets
Mar 17
Most traders skip the math.

Then wonder why their backtests fail in live trading.

You don't need a math degree. But you need the right foundations.

Here's the complete math & stats roadmap for trading: Image
─────────────────
🔹 STATISTICS & PROBABILITY
─────────────────
• Sample Size & Law of Large Numbers — a 10-trade backtest proves nothing. You need hundreds before trusting your edge.
• Expected Value — the only number that actually matters: (Win% × Avg Win) - (Loss% × Avg Loss)
• Standard Deviation — becomes volatility when applied to returns. Every position sizing formula runs through this.
• Correlation — 3 momentum strategies on correlated assets isn't diversification. It's 3x the risk.
• Conditional Probability — your strategy's win rate changes by market regime. Know when it works.
─────────────────
🔹 LINEAR ALGEBRA
─────────────────
• Vectors & Matrices — your portfolio is a weighted sum of vectors. Risk is matrix multiplication.
• PCA — turns 50 correlated indicators into 5 independent signals. Less noise, more edge.
Read 9 tweets
Mar 3
Stock Prediction AI: Using Machine Learning and Deep Learning to predict stock price movements in Python.

The Python code is 100% free on GitHub.

Let's dive in (bookmark this): Image
1. The Python Machine Learning and Deep Learning Libraries:

- mxnet
- gluon
- sklearn
- xgboost Image
2. Stock Price Data (Train/Test)

The dashed vertical line represents the separation between training and test data.

GS is shown but will use 72 assets.

Daily prices for each asset. Image
Read 12 tweets
Mar 2
Stop trading in Excel like it’s 2009.

Start using Python + ChatGPT.

Here's 7 small steps: Image
1. Pick ONE strategy rule (keep it boring)

Start with something simple like moving averages or a 12-1 momentum.

Your goal is execution and learning, not “printing money.”
2. Pull clean price data

Use a real data source (even free to start). If your data is messy, your results are fiction.

Do it with this (free): yfinance github.com/ranaroussi/yfi…
Read 12 tweets
Feb 28
Build an End-to-End Python Algorithmic Trading System (complete roadmap + skills + tools)

Bookmark this. Image
1) Foundations (1–2 weeks)

Learn these:

1. Python basics → data/ML
2. Pandas (data wrangling)
3. Scikit-learn (ML)
4. SQLAlchemy (DB access) Image
2) Data & Storage (free+paid)

1. Prices: yfinance (free)
2. Fundamentals: FMP (paid)
3. DB: DuckDB (fast, file-based)

Tip: go from raw data ▶ cleaned ▶ features as separate SQL tables. Image
Read 11 tweets
Feb 26
Top 10 Algorithmic Trading Strategies (and how they work) 🧵 Image
1. Pairs Trading

Trades two correlated instruments simultaneously. It goes long on one asset and short on the other to profit from deviations from their historical relationship, expecting the correlation to eventually resume.
2. Scalping

Involves making numerous small trades to capture minimal price differences over a short time. For example, tape reading is used to analyze order flow and timing, enabling scalpers to profit from very brief price fluctuations.
Read 16 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(