| Account | Executor | Market | Open | Today | Balance | P&L today | Last signal |
|---|---|---|---|---|---|---|---|
| nas2m Vantage demo | LIVE beat 4m ago (every 5m) | OPEN next macro 03:50–04:10 NY in 1h 7m | flat | 0/4 | 97,466.42 | +0.00 | 13h ago |
| xau2m FTMO demo | LIVE beat 76s ago (every 5m) | OPEN next macro 08:50–09:10 NY in 6h 7m | long 0.05 XAUUSD | 2/4 | 101,742.53 | +35.36 -21.39 banked + costs · +56.75 still open | 4h ago |
⭐ Kept because the trades are evidence about that configuration. ⚠ They are not pooled with the live figures — comparing two different strategies as one sample would say nothing true about either. Gold 2-minute · Nasdaq 3-minute
⭐ What running this on live accounts has made visible that a backtest cannot show — the division of labour between the strategy and the executor, what we measured, the gaps, and the design questions that belong to the author rather than to us. ⛔ Including the four changes we expected to help, measured, and dropped.
The relay runs on a free tier capped at 1,000 writes a day, and ⛔ nothing would announce running out — writes would simply start failing while the relay kept answering and this page kept rendering the last good record. ⭐ So the figure is derived rather than counted: a counter would itself cost a write every time it moved.
| Source | Writes/day | Scales with |
|---|---|---|
| Heartbeats | 864 | ⚠ executors, not instruments — another instrument on an existing account costs nothing here |
| Trade reports | 18 | trades × ~5 legs each |
| Events | 0 | plans, and ⚠ a refusal costs two |
| Projected total | 882 (88%) | at today's rate, 28% of the day elapsed |
⭐ One more executor would add 288 writes a day before it placed a single trade — that, not more trading, is what would breach this. The durable record already lives in a database; this budget only governs the relay's own working store.