Measure a provider before you fund it: hit rate, expectancy, drawdown contribution and parse reliability.
| Metric | Read it as |
|---|---|
| Hit rate | Share of closed trades that ended positive. Meaningless without the average win/loss beside it. |
| Expectancy | Average result per trade. Negative expectancy with a 70% hit rate is common and fatal. |
| Max drawdown | The worst peak-to-trough this source put your account through. |
| Parse rate | Share of messages that produced a usable signal. A low rate means format problems, not bad trading. |
| Rejected | Signals blocked by the completeness gate or your filters, with the reason. |
No orders, full parse logs. You learn the format problems for free.
Real sizing, real slippage, no money at risk.
Ask whether the average win covers the average loss.
Cut risk per trade to a quarter of your normal size for the first weeks.
Send the raw signal, the timestamp and the account name — that is usually enough to answer on the first reply.