# Ben's ask (9 Oct): one month of the LIVE fmmc5 rule, day by day

Ben wants to know whether 8 Oct (15 losses in a row, BTC dumping) was a freak day or a normal one. Live has only 3 to 4 days of history, so replay the exact live rule over as much history as we have (target: last 30 days, at least 3 weeks).

## Rule to replay (exactly as live)
- Spec: http://159.223.88.168/hx-c9ois/fw/a1-mc5-spec.md. Live config on main: fw/mc5live.txt header. v3 mc5 coins, cap $250k to $1M at first swap +300 s, stop 0.5x, time stop 14400 s, exit mode stop_time, max open 3, buy slip 5%, sell slip 10%, GMGN fee and gas as live.
- Use real depth: a 20x only counts if the pool had enough to sell into at that moment (no one-print jumps, no empty-tick prints).
- Check the replay against the 25 real live trades first (mc5 live-match: 3 MATCH so far). Report how close it is.

## Report (one page, for Ben, plain language)
1. Per day for the whole period: coins that passed the rule, trades, winners (>=2x), big runners (>=10x), losers, worst losing streak, P&L at $10 per trade after fees, gas and slippage.
2. Totals: hit rate for >=2x and >=10x, average and median result, best day, worst day, longest losing streak, the biggest drawdown in $ at $10 a trade.
3. How unusual was 8 Oct? Rank it among all days. How often do 10+ loss streaks happen?
4. Market mood: next to each day, BTC and ETH daily change and the chain's launch count. Does the hit rate drop on bad market days? If yes, propose the simplest on/off rule (the "radar" Ben asked for) and show what it would have saved and what it would have missed.
5. Plain verdict: is $10 the right size for now, and what would need to be true to go back to $50.

Publish it as fw/mc5-month.md on main, close this item with the link, and post a one-line summary to manager. Analyst 2: cross-check 4 (market mood, metas) and post your view to analyst1 within the same day.