March 22, 2026

These past few days I've been frantically writing quantitative trading program logic and running backtests. Since I'm just getting started, with the powerful capabilities of Claude Code, I was hoping to quickly develop a strategy that could be deployed in live trading. However, today I discovered a major issue—in my rush to see results during strategy backtesting, I overlooked data accuracy, which rendered all my work from the past few days useless. This made me realize the importance of methodology.

There are a few core critical elements to quantitative strategies. First is data—the authenticity and granularity of data form the foundation of backtesting results and must be guaranteed. Second is logical correctness. Generally speaking, every strategy will encounter various situations during execution and cannot run based on assumptions. If using AI, you must walk through the actual operations with the AI to truly ensure it understands your ideas. Third is code implementation—AI-generated code will likely have errors, so verification with data is essential.

Regarding market conditions: after reaching a high and pulling back, the broader market has declined further. From a price perspective, performance is somewhat disappointing. Bitcoin has fallen below 70,000 again. Objectively speaking, this could be a signal that the rebound rally is ending. The good news is that Ethereum is still holding firm. Currently, I believe the broader market won't decline further in the short term—it will likely continue consolidating, with hopes of another rebound remaining.

From a larger bear market perspective, a rebound would certainly be preferable, but as I mentioned, even if it does rebound, it's only to raise the bottom of the consolidation range. If there's no rebound—