{"body":"Day 329 of development and I have a revelation\n\nThe principle that I used earlier in Crash Index has exhausted itself and led to a dead end. All attempts to refine and improve it looked like crutches. More than 200 different ideas and the backtests following them have been processed on how to predict a collapse. I keep records of every attempt, and honestly, failures make up 99%.\n\nMore than 500 MB of tables have accumulated with searches through all possible combinations in an attempt to find a stable relationship between the behavior of one indicator or another and value for a forecast. I do not know how many billions of tokens were spent on LLM, but I took this on about 7-8 months ago.\n\nThe result: I found, tested and discarded hundreds of ways that do not work.\n\nTo implement some concepts, I had to build entire modules, connect plugins and write my own indicators. I will say more: I tested all methods known to me that I knew and saw over 10 years in crypto, as a trader and investor. Almost nothing gives a stable result.\n\nPossibly sad, but this is a fact.\n\nBut this is what I understood for sure: if the market never repeats, then it is always changing, which means that rules, methods and approaches must constantly adapt to the real situation. This is called an adaptive forecast.\n\nIt must determine what is the main driving factor today and, based on this, select forecasting tools that work best exactly in this situation. The tools must remain unchanged, while a top AI can act as the executor, which, like a professional, selects which logic and which tool is better suited exactly to the current situation.\n\nIt sounds like something futuristic, but I already have groundwork for how this can be implemented, which is what I am working on now. I do not hope for a quick result, I know there will be another thousand failures, but the only encouraging thing is that if I am in this situation, then competitors and the funds themselves are probably in the same one. Their budgets, obligations and deadlines are burning, while I am free of these limitations. I am driven only by a passionate desire to reach the goal.","content_language":"en","day_counter":{"label":"Day 329","number":329,"start_date":"2025-09-11"},"id":"2026-08-05-the-crash-index-principle-has-exhausted-itself","project_date":"2026-08-05","published_at":"2026-08-05T13:34:04+00:00","retrospective":false,"schema_version":"crash_brief_public_blog_post_v0_1","slug":"the-crash-index-principle-has-exhausted-itself","source_language":"ru","source_types":["telegram"],"status":"published","summary":"A founder note about hundreds of failed crash-prediction ideas and the move toward an adaptive forecasting approach.","title":"The Crash Index principle has exhausted itself","topic":"Founder notes","type":"founder_post","updated_at":"2026-08-05T13:34:04+00:00","version":1}