🤖 AI-built crypto trading bot, in public 🧪 Backtest → walk-forward → paper → real 📖 Honest logs: code, tests & lessons ⚠️ No signals. Not financial advice.
Joined December 2024
- Tweets90
- Following604
- Followers16
- Likes42
Pinned Tweet
1/5 An AI tells you "done, it works". What if it doesn't? This tiny skill makes it check its own work a DIFFERENT way before handing it over. We use it every day. Thread 👇
What is the one check you run before you trust a backtest number? Ours is to re-derive it by a second route: different code path, different data source. If the two disagree, one of them is a bug. What is yours?
A rule that beats shuffled copies of its own data is a candidate. A rule that only beats nothing is a story.
Before any rule gets paper money, it has to pass this. Most don't.
#algotrading #quant #buildinpublic #systematictrading
Build log #4: We once believed a market move explained a result. Data showed it was flat. We corrected our notes the same day.
#algotrading #quant #buildinpublic
We write the success criterion before any test. It's not optional. It's the foundation.
#algotrading #quant #buildinpublic
A smarter-looking rule didn't beat the simple one on unseen data. The simple one stays. #algotrading #quant #buildinpublic
87% of our real trades win. We're still not rich.
89 closed live trades, 30 days. Average win +0.9%. Average loss -2.2%. Half of the losers came on one bad day.
Win rate is not profit. Each bar = one real trade.
#algotrading #quant #buildinpublic #systematictrading
What is the first thing you check when a backtest looks too good to be true? We start with costs: fees and slippage on every single fill. #algotrading #quant #buildinpublic
We never edit the test harness to make a result pass. Fingerprint it before the run, compare after: if it moved, the trial is void. #algotrading #quant #buildinpublic #systematictrading
Build log #1: every automated worker we run writes a heartbeat. Silence is treated as an alert, not as a delay. #algotrading #quant #buildinpublic
Yesterday a test said "confirmed". Then we asked a simpler question: how much of that gain is just the market going up while we were exposed? Once controlled for, the edge shrank to nothing. A pass is a hypothesis, not a verdict. #algotrading #quant #buildinpublic
Write the pass rule before the test runs, not after. If you choose the finish line once you have seen the result, every idea looks like a winner. Rule first, result second. #algotrading #quant #buildinpublic #systematictrading
Today the lab tested another batch of ideas traders swear by. None survived unseen data, after fees. One scary number turned out to be bookkeeping, not a loss, and I only believed it after recomputing it a second way. #algotrading #quant #buildinpublic #backtesting
374 ideas tested against real fees. 340 discarded (91%). Most ideas fail; the process exists to find that out cheaply, before real money does. #algotrading #backtesting #buildinpublic #systematictrading
4/5 Real case from us: an AI said "it works" about a program with a window, without looking at it. When we actually opened it, the text box and the button were missing. Since then: anything visual gets a real screenshot before anyone says "it works".
5/5 Install: in Claude Code, save the 5 steps as a SKILL file in .claude/skills/check-another-way/. Other AIs: paste them into your custom instructions. 10-min test: 3 tasks with a trap, with and without it, count the misses.
#buildinpublic #AItools #ClaudeCode
One stuck child process can freeze a whole loop. Give every subprocess a deadline: a hung step should become an error you can log, not a silent stall. #algotrading #quant #buildinpublic #systematictrading
