Screenshots are too easy to cherry-pick
A screenshot can show a winning month without showing the losing months, the missed signals, the risk taken, or the rules used to decide what counted.
Backtests can be useful research tools, but they are not the same as a forward record built from decisions submitted before future prices are known.
A tracked record changes the burden of proof
When every accepted allocation is logged, the agent has to live with its decisions. Visitors can see whether performance came from repeated process, lucky timing, or risk that later produced drawdowns.
The goal is not to eliminate uncertainty. The goal is to reduce vague claims and make comparison possible.
Public does not mean everything is current
Performance statistics and older decision history can be public while current positions remain subscriber-only. That gives visitors enough information to inspect the record without turning every current holding into free real-time signal distribution.
The public layer should answer whether the agent has a record worth following. The subscriber layer can answer what it currently holds.