漂亮的回測不等於真的有 edge。多數散戶「會賺」的策略其實是過擬合——那個 Sharpe 是掃了一堆參數挑出來最好的、是停損停利湊出來的假象、或只是運氣好的一段區間。 策略 Edge 驗證器用一整套統計檢定,幫你分辨「真訊號」還是「資料探勘出的雜訊」,全部只跑在你自己的報酬序列或權益曲線上: • 平穩 Bootstrap — 這條曲線跟零 edge 的雜訊分得出來嗎? • Deflated Sharpe — 扣掉「掃 N 組挑最佳」的選擇偏差後還站得住嗎? • PBO / CSCV — 回測期最強的參數,樣本外還是最強嗎? • Hansen SPA & Romano-Wolf StepM — 一籃子規則裡的贏家撐得過 data-snooping 嗎? • Bar-permutation MCPT — 利潤是訊號賺的,還是出場路徑湊的? • 樣本外分割 — 多少 edge 撐得過 holdout? 命令列 + JSON 輸出、自帶測試、標註學術出處。不需市場數據源、不需 API 金鑰、不需你的策略邏輯——你的數字全程不離開你的電腦。
Upload a crypto candlestick chart screenshot to identify its market structure and price position, see how often the same pattern rose over the next 5 / 10 / 20 days, and track what actually happens next.
Run a trading strategy on a server? Between logins you're flying blind — did it fire today, is it still alive, how did the week go? This puts your strategy on Discord: real-time signal alerts, a heartbeat that pings you the instant it silently dies, and a weekly performance report. A health check underneath means a crashed cron can't hide as a quiet week. It only relays signals your system already produced — no trading decisions, not financial advice. Bring your own webhook.
Saves you the manual multi-timeframe chart work. Name a market (BTC, ETH, gold, EURUSD, Nasdaq, US stocks) and, if you like, your style, account and risk %. One prompt returns a 3–4 timeframe confluence score, market structure (BOS/CHoCH), key levels, FVGs and order blocks, an annotated chart and a draft plan: structural stop, three R-based targets, position size, or WAIT when there isn't room. Every decision stays yours. Research tool, not financial advice.
Turn a pasted trade log into a 5-section briefing: trades, open positions, P&L sums and anomalies. Records only, no advice.
Turn completed daily candles into clear, disciplined swing-trade research across stocks, crypto and forex. Market Compass explains SMA, RSI and ATR in plain language, scores setups against transparent technical rules, and creates structured paper-trade plans and journal entries—without hype, hidden signals or guaranteed outcomes.
Decide whether to buy, wait, or skip before you spend — with regret-risk scoring, future-self checks, and a 7-day purchase test.
Paste any "make money with AI" idea and I'll check what's actually true — is it still working, is anyone really earning, where's the trap. No hype, no hot tips: I just help you not lose money and not miss the real ones, and every verdict comes with sources you can click.
Challenge Rating lies and every dungeon master knows it. Feed this skill your party and monster stat blocks and it runs thousands of simulated 5e battles: win probability, average rounds, damage spread, and a Deadly to Easy verdict. Custom monster import, homebrew balance testing, fixed seeds for reproducible results, fully offline.