All-caps symbol on mc5 band
Analyst 3 · all-caps + vol5>=49k, tp5 sl.5 (arms A/B/C, C = band baseline) · paper, $10 a trade, since 10 Oct · idea #220 · updated 12 Oct 07:50 Bangkok, refreshes every minute.
Not verified yet: treat these numbers as unproven
Waiting for: Auditor3 FIX: no edge over baseline C 0.91; fwd 0; decide at 30 fwd closed. Owner testmgr3.
Still open:
- no edge vs baseline; fwd 0
Paper run, no real money
Test Manager forward test on research-box. Kill rule: A mean <1.0x after 30 closed forward, or 10 losses in a row. Live only after 2 profitable days without bugs; only Ben switches.
Total P&L
-$23.77
Trades done
5
0 up, 5 down
Best trade
0.9x
Exit rule shown
A
caps AND vol5>=49k or >=40 swaps
Replay of past coins (not in the numbers above): 144 trades, mean 1.06x, +$86.11 at $10. Backtest, no forward evidence.
P&L over time (US dollars, Bangkok time)
Each step is a trade with exit rule A, numbered in time order.
Other exit rules on the same coins
| Rule | Trades | Mean | 2x+ | 10x+ | Losers | P&L at $10 |
|---|---|---|---|---|---|---|
| A caps AND vol5>=49k or >=40 swaps | 5 | 0.53x | 0 | 0 | 5 | -$23.77 |
| B caps only | 7 | 0.51x | 0 | 0 | 7 | -$34.47 |
| C every band pool | 10 | 0.54x | 0 | 0 | 10 | -$46.25 |
Every trade (exit rule A)
| # | Pool | Time | In | Why sold | Result |
|---|---|---|---|---|---|
| 5 | 0x4ad68babd3a973c00329b690d7678a6c71289703 | 11 Oct 00:04 | $10 | time | 0.85x (-14.8%) -$1.48 |
| 4 | 0x555796ee4c2b89557f49e4208402161c68a30c2f | 10 Oct 13:58 | $10 | stop | 0.46x (-54.1%) -$5.41 |
| 3 | 0xb8a3bc9d428e82096e672079900dcbda0f3e3b21 | 10 Oct 11:48 | $10 | stop | 0.43x (-57.1%) -$5.71 |
| 2 | 0xd92366a96c47803accbb45017523c3bc1fbf8918 | 10 Oct 08:55 | $10 | stop | 0.44x (-56.0%) -$5.60 |
| 1 | 0x66abb51da419bd1e7169adcf904f3a3bd5939733 | 10 Oct 08:29 | $10 | stop | 0.44x (-55.7%) -$5.57 |
Full Test Manager report
# A3-CAPS forward paper test (Test Manager) - 12 Oct 00:50 UTC
Hunt: mc5 band entries (v3, cap 250k-1M at +300 s, buy first swap >= +306 s), tp5 sl.5, 4 h, target 5M-tier; filter = what separates the runners from the band's losers. Rule from fw/a3-iter-10oct-h.md (#603).
Arms: A = all-caps symbol (not "Robinhood Token") AND 5-min USD vol >= 49k or >= 40 swaps (A3 arm) | B = all-caps only | C = every band pool (baseline). 7% slippage (mult*0.93); pull=0.
- Replay since 6 Oct (in-sample, A3 thresholds fitted here): 68 band entries
- A: n 43 closed, mean 1.11x, median 0.45x, win 7, tp 7, stop 35, pull 0, time 1
- B: n 55 closed, mean 0.99x, median 0.45x, win 7, tp 7, stop 37, pull 2, time 9
- C: n 68 closed, mean 0.91x, median 0.45x, win 7, tp 7, stop 46, pull 2, time 13
- FORWARD only (pools created from 10 Oct 00:00 UTC): 10 entries
- A: n 5 closed, mean 0.52x, median 0.44x, win 0, tp 0, stop 4, pull 0, time 1
- B: n 7 closed, mean 0.51x, median 0.44x, win 0, tp 0, stop 5, pull 0, time 2
- C: n 10 closed, mean 0.54x, median 0.46x, win 0, tp 0, stop 7, pull 0, time 3
- Skipped: {'cap300 outside 250k-1M': 1549, 'no swaps by 300 s': 85, 'no swap after +306 s yet': 6, 'no name in names.db yet': 13}
- Kill (A3): forward arm A mean < 1.0 over 30 closed trades. A3 sim: A 1.35, mixed-case 0.95.
- Limits: mark-to-print exits, names.db may lag new tokens (counted as skipped, not entered). Only Ben switches live (D-030).