Post

@GoshawkTrades

This Thread may be Removed Anytime!

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

More from @GoshawkTrades

Goshawk Trades

@GoshawkTrades

Jan 3

I don't understand why people aren't using automation more for trading.

It's one of the best ways to solve execution and scale.

Here is a list of ways you can automate your trading in 2026:

First before you can automate, you need to quantify.

How do you do this?

Turn the strategy into if/and/or logic, breaking down every single part.

This way the strategy is repeatable and testable.

Without this, as Jim Simons says, it's impossible to simulate one-off decisions.

1. Crypto Exchange APIs (Binance, Hyperliquid, Bybit, etc.)

What they provide:
- REST/WebSocket endpoints for real-time trading
- Spot, futures, and derivatives access
- 24/7 market availability

Read 16 tweets

Goshawk Trades

@GoshawkTrades

Nov 28, 2025

4 backtesting techniques I used to find my 10 profitable algos:

1. Parameter sensitivity
2. Walk-forward optimization
3. Stress testing
4. Monte Carlo

Here's why each one matters:

1. Parameter Sensitivity

Tests how small changes in your parameters affect performance.

If tweaking your MA from 20 to 22 kills your returns — you're overfit.

Robust strategies should show smooth performance across parameter ranges, not cliff edges.

Quick note on "overfit":

It's when your strategy "memorizes" noise.

It's like creating a rule for every day in your backtest — works perfectly backwards, fails completely live.

The more complex your strategy, the more likely it's overfit.

That's why these 4 techniques matter.

Read 10 tweets

Goshawk Trades

@GoshawkTrades

Oct 23, 2025

I've been algorithmic trading for 5 years.

In that time, I went from 0 to 10+ live algo strategies making multi 6 figures.

Here are 12 things I wish I knew on day 1:

Context:

I've been trading for 6 and bit years. My first year was a disaster.

Then I discovered I could use coding and data for trading. Since then, it's been up year-over-year, but I've made tons of mistakes.

Here are the lessons learned (the hard way) so you don't have to:

1) Sample size matters more than you think.

A strategy with 5 trades that made 200% historically is meaningless.

You need at least 300+ trades, ideally 500+, to have statistical confidence.

More samples = more confidence it wasn't just luck / noise

Read 17 tweets

Goshawk Trades

@GoshawkTrades

Sep 23, 2025

Expected returns.

An Investor's Guide to Harvesting Market Rewards.

I've read it more than I can count, and it will change how you think about trading.

Here’s a full breakdown - so you don’t have to read all 979 pages:

1. Most returns come from risk (or mispricing).

If you think you’ve found alpha, it’s usually just a disguised risk premium.

The key is knowing which risk you’re being paid to hold.

It could be:
• Bearing a risk others avoid
• Providing liquidity
• Exploiting behavioral bias

2. Momentum works across assets.

3–12 month momentum is one of the most robust anomalies.

It’s also long volatility: it shines when crises hit, making it one of the few diversifiers against equity risk.

Read 10 tweets

Goshawk Trades

@GoshawkTrades

Aug 12, 2025

Algorithmic trading doesn't just solve execution.

It eliminates human bias, enables 24/7 trading, and
scales diversification.

If you care about consistency, scalability, and edge, read this:

Let's bust some myths:

Algorithmic trading is easy money with no effort.

Wrong.

It requires upfront work.

But that effort pays off with consistent and scalable execution.

1. Eliminates Emotional Trading Execution.

Often the destroyer of trading accounts isn't bad strategy—it's execution.

Code executes without emotion, following your rules exactly as backtested.

Read 11 tweets

Goshawk Trades

@GoshawkTrades

Jul 16, 2025

I've read Machine Trading more than I can count.

My favorite line:

"In trading, complexity doesn't pay"

Here are 6 lessons from the book you need to know:

1. Start with simple.

Ernest Chan emphasizes starting with the simplest technique, before moving to complex methods.

Most traders like to jump the gun to advanced machine learning thinking it guarantees them an edge...

It normally is the complete opposite.

2. Each parameter must fight for its place.

Chan warns that excess parameters are the easiest way to overfit a strategy.

Keep only parameters that show statistical significance and serve a clear purpose.

Whether generating returns or controlling risk.