The short answer
Around 30 closed trades before any number is worth reading.
100 or more before you size a position against the expectancy.
Below 30 you are reading the stretch of price you tested, not the rule you wrote.
Why 12 trades tells you nothing
Take a rule that wins half the time. A coin. Run it 12 times. Eight wins happens about one time in five by chance. That is Tuesday.
Twelve live trades is about six weeks for most people. Six good weeks feels like proof. It is not.
What changes at 30
Thirty is a floor, not a target. Luck narrows enough that a genuinely bad rule starts to show.
You can say this does not look catastrophic. You cannot yet say this has an edge of 0.3R.
What changes at 100
Expectancy starts to mean something. You have probably met the worst losing streak.
A real +0.3R rule still prints runs of six or seven losses. If you have only seen 30 trades you may not have met that run yet. You will meet it live, with money, and quit at the wrong moment.
The trap
More trades from a shorter period is not the same as more trades.
Four hundred trades from one quarter is one market condition. A trend rule tested only in a trend looks like a god. Then it sits still for eight months.
You want trades across conditions: trending, ranging, volatile, quiet. That usually means years. That is the argument for running history rather than waiting.
How to check yours
Count the closed trades. Then ask how many distinct conditions they span. If the answer is one, the number in front of you is smaller than it looks.