CBaquero

CBaquero / BTCautoresearch

Public

Autoresearch: autonomous formula discovery for Bitcoin price prediction (time-based)

45
4
100% credibility
Found Mar 30, 2026 at 45 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

An open-source project where an AI agent autonomously tests and refines mathematical formulas to better predict Bitcoin prices from historical daily data using walk-forward validation.

How It Works

1
🔍 Discover Bitcoin Prediction Magic

You find this exciting project that uses AI to hunt for the best ways to predict Bitcoin's future price based on time.

2
💻 Get it ready on your computer

Download the project files and prepare everything with a simple setup so you can start playing with predictions.

3
▶️ Try the current top prediction

Run the best model so far to see how it forecasts Bitcoin prices into the future.

4
📈 See amazing results!

Check out charts and stats showing predictions improved by over 50% compared to basic trends, with graphs for months to years ahead.

5
🤖 Invite an AI assistant to experiment

Point a smart AI helper at the project and ask it to start testing new prediction ideas.

6
🔄 Watch improvements happen

The AI runs hundreds of quick tests, tweaking the formula and keeping only changes that make predictions sharper.

🎉 Unlock better Bitcoin forecasts

Celebrate having a superior model that guesses Bitcoin prices more accurately, even up to 5 years out.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 45 to 45 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is BTCautoresearch?

BTCautoresearch is a Python project for autonomous formula discovery in Bitcoin price prediction using time-based models. It sets up an AI agent to run experiments, iteratively improving prediction formulas evaluated on walk-forward out-of-sample RMSE across horizons up to five years. Users get a ready-to-run best model via simple CLI command, plus a protocol to extend it with their own AI like Claude.

Why is it gaining traction?

This autoresearch setup stands out by automating the search for better Bitcoin prediction formulas, delivering 50.5% RMSE improvement over power-law baselines through 328 AI-driven experiments. Developers dig the strict evaluation harness—multi-split, multi-horizon validation with bootstrap stats—and the easy loop for continuing autonomous research. Plots and logs make results transparent, hooking those into AI-augmented quant work.

Who should use this?

Crypto quants tweaking time-based Bitcoin models for trading signals. ML engineers experimenting with autonomous agents for formula optimization. Python devs in finance wanting a baseline for long-horizon price forecasts up to 36 months.

Verdict

Worth forking for AI-driven research experiments, but 45 stars and 1.0% credibility score signal early-stage maturity—docs are solid, but expect tweaks for production. Try the CLI model run first; extend if autonomous discovery clicks.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.