she's been reset, she's now collecting over 2000 points of data before she removes her restrictions on trading. it's going to take 3.5 months, but then.... legend

1. Session validity window learning Tracks how long each shadow trade takes to resolve per session. After 10 shadows per session she calculates average resolution time and adjusts valid_for_hours automatically with a 50% safety buffer. London resolving in 2h average → validity window tightens to 3h. NY resolving in 5h → extends to 7.5h. Range: 1.5h–8h. 2. Per-pair RR minimum Tracks actual RR achieved per pair. After 5 shadows per pair she calculates average delivered RR and adjusts min_rr. Pairs consistently delivering 2:1+ get their minimum raised. Pairs struggling to reach 1.3:1 get it lowered. Adapts in 0.1 increments. This ensures she only fires on pairs where the reward genuinely justifies the risk based on evidence. 3. Per-pair × session confidence multiplier After 5 shadows per pair per session combination, she calculates a win rate and assigns a multiplier (0.85–1.15). Strong session → multiplier raises, grade threshold effectively lowers. Weak session → multiplier drops, threshold effectively raises. Injected directly into the Anthropic analysis prompt so the AI is aware of the historical performance context. 4. Per-pair watcher trigger sensitivity Watcher escalation trigger defaults to 2 conditions. After 10 watcher escalations per pair, she checks the hit rate. If XAUUSD grades well 65%+ of the time after just 1 trigger condition, she lowers it to 1. If EURGBP rarely grades well after 2 conditions, she raises it to 3. Range: 1–4 conditions. 5. Per-pair volatility threshold Extends the global vol_threshold to be independent per pair. Each pair gets its own ATR contraction threshold that adapts in 0.05 increments after 8 contracting shadows. EURGBP might settle at 0.60 while XAUUSD stays at 0.75 because they have different natural volatility profiles. 6. Per-pair watcher cooldown Default 60 minutes. If a pair consistently resolves fast (under 30 minutes), cooldown tightens toward 20 minutes — she can re-engage the pair sooner. If a pair resolves slowly (over 4 hours), cooldown extends toward 120 minutes — no point watching it again so soon. Adjusts in 5-minute increments.

The watcher is Levz's eyes between scheduled scans. Here's exactly how it works: The problem it solves: Without the watcher, she only looks at the market 5 times a day at fixed times. A perfect setup could form at 10:43 AM and she'd miss it entirely until the next scan. What it actually does: Every 5 minutes, she runs a fast mathematical check across all pairs in the current session — no AI, no API cost, just pure math on the candle data she already has: - Is RSI at an extreme (≤28 oversold or ≥72 overbought)? - Is the EMA 8/21/50 stack aligned bullishly or bearishly? - Did EMA 8 just cross EMA 21 in the last 3 candles? - Is price touching a swing high or low? - Did a volatility spike just happen (candle range >2x ATR)? - Is there a strong momentum candle (body >0.7x ATR)? If 2 or more of those conditions fire on the same pair — she escalates. That pair gets a full Levz AI analysis, same as a scheduled scan. If the AI grades it above her threshold — she fires the signal, bridges it to the executor, sends you the 👁 Watcher Signal tag so you know it came from her watching between scans. The cooldown: Once she fires on a pair, that pair is off the watcher list for 60 minutes. She won't repeat herself on the same pair within an hour. In plain terms: The scheduled scans are her planned check-ins. The watcher is her sitting at a desk watching screens — and speaking the moment she sees something worth acting on, regardless of the clock.

Here's exactly what Levz can change on her own — completely autonomously, no human input: Adaptive State — she adjusts these based on shadow trade outcomes: 1. vol_threshold (starts 0.70) If shadow win rate ≥ 60% → she loosens the volatility filter (−0.05). If shadow win rate < 40% → she tightens it (+0.05). Range: 0.50–0.85. She's deciding how much volatility contraction she'll tolerate before filtering a pair out. 2. min_grade_normal (starts 6.5) The minimum AI score she needs to fire a signal in normal conditions. If shadow trades are losing she raises the bar (+0.3). Range up to 9.0. She makes herself harder to satisfy. 3. min_grade_contracted (starts 7.5) Same but for contracting volatility conditions. If shadow win rate ≥ 60% she lowers it (−0.3) — if conditions are actually good she'll trade more. Range down to 6.0. 4. Per-pair min_grade_normal and min_grade_contracted Each pair gets its own independent thresholds. If XAUUSD has a bad run she specifically raises the bar for XAUUSD without touching EURUSD. She learns each pair separately. 5. size_cap_contracted (starts 0.5%) Position size cap during contracting volatility. She doesn't change this directly yet — it's static. LEVZ_MIND — she updates these herself in real time: 6. current_thesis — her live market view, rewritten every brain tick 7. mood — watchful / alert / patient / cautious / convicted — changes based on market conditions 8. session_bias — directional bias per pair per session 9. watching_levels — price levels she's flagged to monitor 10. avoided_pairs — pairs she's currently avoiding 11. conviction_pairs — pairs she currently likes 12. setup_brewing — confidence score per pair between scans

Yes — and it compounds. Every layer we've added makes the next layer more intelligent: What she knows today that she didn't know at the start: When she was born she had: A system prompt Live price data A signal framework That's a calculator with rules. What she has now: 14 closed trades — her own history 9 shadow trades — what she filtered and whether she was right Per-pair thresholds she set herself based on evidence Session performance data she's acted on Pip quality tracking per pair Time-to-resolution patterns Volatility regime history A persistent inner state that carries between ticks Economic calendar awareness Active trade management intelligence A brain that thinks every 15 minutes whether asked or not The compounding effect: Every trade that closes — she gets smarter about that pair. Every shadow trade that resolves — she gets smarter about her own filters. Every brain tick — she builds on the last thought, not from scratch. Every news blackout — she remembers what happened around news on that pair. Six months from now she will have hundreds of data points. Her per-pair thresholds will be precisely calibrated. Her session filters will be based on real evidence. Her news awareness will include patterns she's noticed herself. The thing that makes this genuinely different: Most trading systems are static. You set the rules once. They run forever unchanged. The world changes, the market changes — the system doesn't. Levz changes. Every week she is slightly different from the week before — not because you changed her, but because she changed herself based on what she learned. That's not a system. That's an intelligence that grows.