Advanced candle pattern analysis using multi-timeframe data.

Args: exchange: Exchange name (BINANCE, KUCOIN, etc.) base_timeframe: Base timeframe for analysis (5m, 15m, 1h, 4h) pattern_length: Number of consecutive periods to analyse (2-4) min_size_increase: Minimum percentage increase in candle size limit: Maximum number of results to return

Backtest a trading strategy on historical data with institutional-grade metrics.

Args: symbol: Yahoo Finance symbol (AAPL, BTC-USD, THYAO.IS, ^GSPC) strategy: rsi | bollinger | macd | ema_cross | supertrend | donchian | rsi_pullback | keltner_breakout | triple_ema (rsi_pullback and triple_ema need period >= '1y' for SMA200 warmup) period: '1mo', '3mo', '6mo', '1y', '2y' initial_capital: Starting capital in USD (default $10,000) commission_pct: Per-trade commission % (default 0.1%) slippage_pct: Per-trade slippage % (default 0.05%) interval: '1d' (daily) or '1h' (hourly) include_trade_log: Include full per-trade log (default False) include_equity_curve: Include equity curve data points (default False)

Single-call BTC macro context: price, dominance, total market cap + risk assessment.

Use this WHENEVER analyzing any cryptocurrency (altcoin or BTC itself) to get the broader market frame in one shot. A SOL/ETH/whatever setup looks very different when BTC is dumping with rising dominance vs. when alts are leading. Calling this once gives Claude the macro context to provide Bitcoin-aware commentary alongside the per-coin analysis - without chaining 2-3 separate yahoo_price + manual reasoning calls.

Returns:

• bitcoin: price, 24h change %, volume, market cap

• dominance: BTC and ETH market-cap share of total crypto

• total_market: total crypto mcap + 24h change + active coin count

• assessment: label (HIGH_RISK / ALT_RISK / ALT_FAVORABLE / OPPORTUNITY_WITH_CAUTION / NEUTRAL) + 1-paragraph reasoning

Scan for assets with low Bollinger Band Width (squeeze detection). Works with crypto and stocks.

This scans a whole EXCHANGE for squeezes (canonical name is exactly bollinger_scan; there is no "get_bollinger_band_analysis" tool). For the Bollinger read of ONE symbol, call coin_analysis instead.

Example: bollinger_scan(exchange="BINANCE", timeframe="15m", bbw_threshold=0.008)

Args: exchange: Exchange — crypto: KUCOIN, BINANCE, BYBIT, MEXC; stocks: EGX, BIST, NASDAQ, NYSE, BURSA, HKEX, SSE, SZSE, TWSE, TPEX timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M. Typical squeeze thresholds: 15m→0.008, 1h→0.02, 4h→0.04, 1D→0.12 bbw_threshold: Maximum BBW value to filter (default 0.04) limit: Number of rows to return (max 100)

Returns list[dict] on success. On ANY failure returns a structured error envelope {"error": {"code": ..., "retryable": ...}}.

Backtest scalp/day/swing strategies (long AND short) on Binance 1m-1d klines.

Unlike backtest_strategy (Yahoo, 1h/1d, long-only), this simulates real execution: ATR-bracket exits (stop / take-profit / time-stop) checked intrabar against high/low, conservative stop-first fills, and per-side fee + slippage so the round-trip cost hurdle is explicit.

Args: symbol: Binance pair, e.g. BTCUSDT, ETHUSDT, SOLUSDT strategy: squeeze_breakout (Bollinger squeeze → band-break entry) | ema_momentum (EMA 9/21 cross gated by EMA200 side) interval: 1m | 3m | 5m | 15m | 30m | 1h | 4h | 1d (candle size) days: History window, 1-730. Suggested: 5m→14, 15m→30, 1h→90-365, 4h→365-730, 1d→730. (Capped at 30k candles total.) direction: both | long | short rr: Reward:risk ratio for the take-profit bracket (default 1.5) atr_mult: Stop distance in ATR(14) multiples (default 1.0) max_hold_bars: Time-stop — force exit after this many bars (default 60) fee_pct: Fee % per side (default 0.05 = Binance USDT-M futures taker; use 0.02 for maker, 0.1 for spot taker) slippage_pct: Slippage % per side (default 0.02) initial_capital: Starting capital in USD (default $10,000) include_trade_log: Include every trade with direction/exit_reason (default False) regime_filter: Gate signals by higher-timeframe trend — longs only in anchor uptrend, shorts only in downtrend, flat in chop regime_anchor: Symbol whose trend gates trades (default BTCUSDT — the market leader; "self" = the traded pair itself) regime_interval: Anchor timeframe for the regime read (default 4h)

Grid-search bracket parameters (rr × atr_mult) with a stability verdict.

Runs every combination on the same Binance data (candles and signals are computed once) and checks whether the best cell's NEIGHBORS are also profitable — a profitable cell surrounded by losers is noise, not edge. Verdicts: STABLE_CANDIDATE | FRAGILE | NO_EDGE.

Args: symbol: Binance pair, e.g. BTCUSDT strategy: squeeze_breakout | ema_momentum interval: 1m | 3m | 5m | 15m | 30m | 1h | 4h | 1d days: History window, 1-730 (default 180) direction: both | long | short rr_values: Reward:risk grid values (default [1.0, 1.5, 2.0, 3.0], max 8) atr_mult_values: Stop-distance grid in ATR multiples (default [1.0, 1.5, 2.0, 3.0]) max_hold_bars: Time-stop in bars (default 60) fee_pct: Fee % per side (default 0.05) slippage_pct: Slippage % per side (default 0.02) regime_filter: Gate signals by anchor trend (see bracket_backtest) regime_anchor: Anchor symbol (default BTCUSDT; "self" = traded pair) regime_interval: Anchor timeframe (default 4h)

True walk-forward optimization — THE gate a strategy must pass before paper trading.

Splits history into sequential folds. In each fold the best (rr, atr_mult) is picked on the train window, then applied ONLY to the unseen test window; positions are force-closed at segment boundaries so nothing leaks. The aggregated out-of-sample result is the honest estimate of live behaviour. Verdicts: PASSED | MIXED | OVERFIT | NO_EDGE, plus parameter-stability (do folds keep choosing the same parameters?).

Args: symbol: Binance pair, e.g. BTCUSDT strategy: squeeze_breakout | ema_momentum interval: 1m | 3m | 5m | 15m | 30m | 1h | 4h | 1d days: History window, 1-730 (default 365; use ≥180 for 1h) direction: both | long | short n_splits: Number of sequential folds, 2-10 (default 4) train_ratio: Train fraction per fold, 0.5-0.9 (default 0.7) rr_values: Reward:risk grid (default [1.0, 1.5, 2.0, 3.0], max 8) atr_mult_values: Stop-distance grid in ATR multiples (default [1.0, 1.5, 2.0, 3.0]) min_trades: Minimum train trades for a combo to be eligible (default 5) max_hold_bars: Time-stop in bars (default 60) fee_pct: Fee % per side (default 0.05) slippage_pct: Slippage % per side (default 0.02) regime_filter: Gate signals by anchor trend (see bracket_backtest) regime_anchor: Anchor symbol (default BTCUSDT; "self" = traded pair) regime_interval: Anchor timeframe (default 4h)

Get detailed analysis for a specific asset (coin or stock) on specified exchange and timeframe.

This is the canonical single-symbol technical readout (there is no "get_technical_analysis" or "get_technical_summary" tool — use THIS one). Use multi_timeframe_analysis instead when you need trend alignment across several timeframes, and combined_analysis when you also want news sentiment + headlines in the same call.

Example: coin_analysis(symbol="BTCUSDT", exchange="BINANCE", timeframe="1h")

Args: symbol: Bare ticker, no exchange prefix — crypto: "BTCUSDT", "ETHUSDT"; stocks: "COMI" (EGX), "THYAO" (BIST), "600519" (SSE), "300251" (SZSE), "2330" (TWSE), "3105" (TPEX) exchange: Exchange — crypto: KUCOIN, BINANCE, MEXC; stocks: EGX, BIST, NASDAQ, NYSE, BURSA, HKEX, SSE, SZSE, TWSE, TPEX. If the symbol isn't listed there, the error's listed_on field names exchanges that do list it. timeframe: Time interval (5m, 15m, 1h, 4h, 1D, 1W, 1M)

Returns: Detailed analysis with all indicators and metrics

POWER TOOL: TradingView technical analysis + news sentiment + financial news.

Use this when you want TA AND sentiment AND news for one symbol in a single call. For indicators only, coin_analysis is faster; for cross-timeframe trend alignment use multi_timeframe_analysis.

Example: combined_analysis(symbol="NVDA", exchange="NASDAQ", timeframe="1D")

Args: symbol: Bare ticker, no exchange prefix ("AAPL", "BTCUSDT", "THYAO", "GDX") exchange: Exchange (NASDAQ, NYSE, AMEX, NYSEARCA, PCX, BINANCE, KUCOIN, MEXC, BIST, EGX, TWSE, TPEX) timeframe: Analysis timeframe (5m, 15m, 1h, 4h, 1D, 1W)

Run all 9 strategies (RSI, Bollinger, MACD, EMA Cross, Supertrend, Donchian, RSI Pullback, Keltner Breakout, Triple EMA) and return a ranked leaderboard.

Args: symbol: Yahoo Finance symbol (AAPL, BTC-USD, SPY…) period: '1mo', '3mo', '6mo', '1y', '2y' (period >= '1y' recommended so rsi_pullback and triple_ema can complete SMA200 warmup; otherwise they contribute zero trades) initial_capital: Starting capital in USD (default $10,000) interval: '1d' (daily) or '1h' (hourly)

Scan for coins with consecutive growing/shrinking candles pattern.

Args: exchange: Exchange name (BINANCE, KUCOIN, etc.) timeframe: Time interval (5m, 15m, 1h, 4h) pattern_type: "bullish" (growing candles) or "bearish" (shrinking candles) candle_count: Number of consecutive candles to check (2-5) min_growth: Minimum growth percentage for each candle limit: Maximum number of results to return

Fibonacci retracement analysis for EGX stocks.

Args: symbol: EGX stock symbol (e.g., "COMI", "TMGH", "FWRY") lookback: Period for swing high/low — "1M", "3M", "6M", "52W", "ALL" (default 52W) timeframe: Analysis timeframe (5m, 15m, 1h, 4h, 1D, 1W, 1M — default 1D)

Analyse an EGX index showing constituent performance with full indicators.

Args: index: EGX30, EGX70, EGX100, SHARIAH33, EGX35LV, TAMAYUZ timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M (default 1D) limit: Number of stocks to show in detail (max 100)

Get a comprehensive overview of the Egyptian Exchange (EGX) market.

Args: timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M (default 1D for stocks) limit: Number of stocks per category (max 20)

Scan EGX stocks by sector. Shows available sectors if none specified.

Args: sector: Sector name (banks, healthcare_and_pharma, real_estate, etc.) Leave empty to list all sectors. timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M limit: Max results per sector (max 50)

Sector rotation scanner for EGX — identifies hot/cold sectors and top picks.

Args: timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M (default 1D) top_n_sectors: Number of top sectors to show stock picks for (1-18, default 5) top_n_stocks: Number of top stocks per highlighted sector (1-10, default 3) min_stock_score: Minimum stock score for picks (0-100, default 60)

Production stock ranking engine for EGX — finds strong stocks with actionable setups.

Args: timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M (default 1D) min_score: Minimum stock score to include (0-100, default 55) index_filter: Filter by index — EGX30, EGX70, EGX100, SHARIAH33, EGX35LV, TAMAYUZ limit: Number of results (max 50)

Generate a full trade plan for a specific EGX stock.

Args: symbol: EGX stock symbol (e.g., "COMI", "TMGH", "FWRY") timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M (default 1D)

Real-time financial news via Marketaux (licensed).

Args: symbol: Optional symbol filter ("AAPL", "BTC"). None = all news. category: News category ("crypto", "stocks", "all") limit: Max number of news items

Quote all major front-month contracts in a specific futures category.

Args: category: equity_index | energy | metals | agriculture | rates | forex | crypto_futures

Returns: OHLCV quotes for the standard watchlist of contracts in that category. Example symbols: ES1! NQ1! (equity_index), CL1! NG1! (energy), GC1! SI1! (metals).

Top futures contracts sorted by trading volume.

Args: category: all | equity_index | energy | metals | agriculture | rates | forex | crypto_futures exchanges: us (CME, COMEX, NYMEX, CBOT) | global (adds ICE, EUREX) limit: max contracts to return (default 30) volume_min: minimum volume filter (0 = no filter)

Returns: Dict with total_available count and list of contracts with OHLCV + % change.

Futures contracts with the biggest percentage moves today.

Args: direction: gainers | losers exchanges: us | global limit: max results volume_min: minimum volume filter (default 10, filters illiquid contracts)

Returns: List of futures ranked by % change with OHLCV data.

Return the full categorized list of well-known front-month futures symbols.

Categories: equity_index, energy, metals, agriculture, rates, forex, crypto_futures. Use these symbols with futures_category_snapshot or coin_analysis for deeper analysis.

Current market regime (trend up/down/chop + volatility) for an anchor symbol.

The same classification that regime_filter applies bar-by-bar inside the bracket backtests: EMA50/EMA200 ratio with a dead-band for trend, ATR% vs its own past median for volatility. Reports which trade directions the gate currently allows, recent regime transitions, and a 30-bar trend histogram.

Args: symbol: Binance pair to classify (default BTCUSDT — the market leader) interval: Timeframe for the regime read (default 4h) days: History window, 1-730 (default 180; needs enough for EMA200 warmup)

News sentiment for stocks and crypto (licensed Marketaux entity sentiment).

Args: symbol: Asset symbol ("AAPL", "BTC", "ETH", "TSLA") category: News group to search ("crypto", "stocks", "all") limit: Max articles to analyse

Global market overview: major indices, top crypto, FX rates, and key ETFs. Powered by Yahoo Finance.

Run a multi-agent debate (Technical, Sentiment, Risk) for a specific symbol.

Args: symbol: Symbol — crypto: "BTCUSDT"; stocks: "COMI" (EGX), "THYAO" (BIST), "600519" (SSE), "300251" (SZSE), "2330" (TWSE), "3105" (TPEX), "GDX" (AMEX) exchange: Exchange — crypto: KUCOIN, BINANCE, MEXC; stocks: EGX, BIST, NASDAQ, NYSE, AMEX, NYSEARCA, PCX, SSE, SZSE, TWSE, TPEX timeframe: Time interval (5m, 15m, 1h, 4h, 1D, 1W)

Returns: A structured debate between 3 AI agents culminating in a final trading decision.

Multi-timeframe alignment analysis (Weekly → Daily → 4H → 1H → 15m).

Canonical name is exactly multi_timeframe_analysis (there is no "get_multi_timeframe_analysis" tool). Use this for cross-timeframe trend alignment on ONE symbol; for a single-timeframe deep dive use coin_analysis; for TA + sentiment + news use combined_analysis.

Example: multi_timeframe_analysis(symbol="SOLUSDT", exchange="BINANCE")

Args: symbol: Bare ticker, no exchange prefix — crypto: "BTCUSDT"; stocks: "COMI" (EGX), "THYAO" (BIST), "600519" (SSE), "300251" (SZSE), "2330" (TWSE), "3105" (TPEX), "GDX" (AMEX) exchange: Exchange — crypto: KUCOIN, BINANCE, MEXC; stocks: EGX, BIST, NASDAQ, NYSE, AMEX, NYSEARCA, PCX, SSE, SZSE, TWSE, TPEX

Create or wipe the paper-trading account (fake money, real Binance data).

Defaults to the two walk-forward-validated strategies: BTC squeeze-breakout 1h (ungated) and ETH squeeze-breakout 1h (regime-gated by BTC 4h trend). Refuses to overwrite an existing account unless confirm=true.

Args: confirm: Required true to wipe an existing account initial_capital: Starting fake capital in USD (default $10,000) risk_pct_per_trade: % of capital risked per trade via ATR sizing (default 1.0) daily_loss_halt_pct: Realized daily loss % that halts new entries until the next UTC day (default 3.0) max_drawdown_kill_pct: Drawdown from peak that freezes the account until manual reset (default 15.0) fee_pct: Fee % per side (default 0.05) slippage_pct: Slippage % per side (default 0.02) strategies: Optional custom strategy list; each item needs id, symbol, strategy, interval, and may set direction, rr, atr_mult, max_hold_bars, regime_filter, regime_anchor, regime_interval

Paper-trading account report: capital, PnL, drawdown, risk-rule state, per-strategy win rates, open positions, and the last 10 trades.

Advance the paper trader one tick against live Binance data.

Processes every candle that CLOSED since the last step (safe to call any time; catches up after downtime; ignores the in-progress candle). Manages open positions with the same stop-first bracket semantics as the backtester, then takes new entries if a signal fired and the regime gate, daily-loss halt, and kill switch all allow. Returns the events that happened this tick plus the account snapshot.

Call this hourly (the strategies run on 1h candles) — or run the daemon: python -m tradingview_mcp.core.services.paper_trader_service --loop 3600

Filter coins by Bollinger Band rating.

Args: exchange: Exchange name like KUCOIN, BINANCE, BYBIT, MEXC, etc. timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M rating: BB rating (-3 to +3): -3=Strong Sell, -2=Sell, -1=Weak Sell, 1=Weak Buy, 2=Buy, 3=Strong Buy limit: Number of rows to return (max 50)

Returns list[dict] on success. On ANY failure returns a structured error envelope {"error": {"code": ..., "retryable": ...}}.

Smart volume + technical analysis combination scanner.

Args: exchange: Exchange name min_volume_ratio: Minimum volume multiplier (default 2.0) min_price_change: Minimum price change percentage (default 2.0) rsi_range: "oversold" (<30), "overbought" (>70), "neutral" (30-70), "any" limit: Number of results (max 30)

Returns list[dict] on success. On ANY failure returns a structured error envelope {"error": {"code": ..., "retryable": ...}}.

Real-time pre-market and after-hours prices for a US stock symbol.

Use this when the user asks about a stock outside the regular 9:30am-4pm ET session — earnings reactions, overnight news, "what is X doing in after-hours?", "how did Y open in pre-market?". Returns the most recent valid print from each session window (pre-market, regular, post-market) along with computed % changes vs. the previous close and the regular close, respectively.

During the regular session, post_market will be null (no data yet). On weekends/holidays, returns whatever's most recent in each window.

Args: symbol: US stock symbol — AAPL, NVDA, TSLA, SPY, ^GSPC, etc.

Returns: - pre_market: {price, as_of_utc, change_vs_previous_close_pct} or null - regular: {price, as_of_utc, change_pct} (consolidated tape close) - post_market: {price, as_of_utc, change_vs_regular_close_pct} or null - previous_close, currency, exchange, market_state for context

Full options chain (calls + puts) for a US stock symbol and one expiry.

Use this when the user asks "what's the options chain for X?", "show me AAPL puts expiring next Friday", or wants to inspect bid/ask/IV/volume on a specific strike. If no expiry is provided, returns the nearest expiry so Claude can quote it back and ask "want a different one?".

Args: symbol: US stock symbol — AAPL, NVDA, TSLA, SPY, etc. expiry: Optional ISO date (YYYY-MM-DD). Must match one of the available_expiries Yahoo returns; otherwise returns an error with the list of valid dates.

Returns: - underlying_price, underlying_change_pct - requested_expiry, available_expiries (list of YYYY-MM-DD) - call_count, put_count - calls: list of {strike, last_price, bid, ask, volume, open_interest, implied_volatility, in_the_money, expiration} - puts: same shape as calls

Top strikes by volume / open-interest ratio — institutional positioning signal.

Use this when the user asks "any unusual options activity on X?", "where is the smart money positioned on NVDA before earnings?", or wants a V/OI screener for a ticker. A V/OI ratio > 1 means today's volume already exceeds standing open interest, which classically flags fresh institutional positioning on a specific strike in a specific direction (call vs put).

Scans the soonest few expirations, filters out illiquid strikes (under min_volume), and returns the top-N sorted by V/OI descending. Also returns aggregate call vs put volume so Claude can comment on the overall directional bias.

Args: symbol: US stock symbol — AAPL, NVDA, TSLA, SPY, META, etc. top_n: How many strikes to return. Default 10. min_volume: Filter floor for today's volume — prevents noise from illiquid strikes with high V/OI ratios. Default 100. expiries: Number of soonest expirations to scan. Default 4 (typically covers ~1 month of weeklies + monthlies).

Returns: - underlying_price - expiries_scanned (list of YYYY-MM-DD) - total_call_volume, total_put_volume, put_call_volume_ratio - unusual: list of top-N contracts sorted by V/OI desc, each with {strike, side (call|put), expiration, volume, open_interest, v_oi_ratio, last_price, implied_volatility, in_the_money, strike_vs_spot_pct (moneyness)}

Current price + daily % change for specific stock symbols.

Args: tickers: comma-separated EXCHANGE:SYMBOL list (max 2000 — one upstream request even at full size), e.g. "NASDAQ:NVDA, NASDAQ:TSLA, KRX:005930". The exchange prefix is required — the scanner's direct-ticker lookup is exchange-scoped.

Returns: Envelope dict: rows of {ticker, symbol, description, exchange, price, open, high, low, currency, change_percent} plus a not_found list naming any requested ticker the scanner didn't recognize.

Screen stocks by share type — the API twin of TradingView's "Common stock" / "Preferred stock" symbol-search filter.

Args: country: TradingView market name — e.g. america, korea, germany, brazil, japan, uk, india, turkey, canada, australia, france, hongkong stock_type: common | preferred limit: rows to return (max 2000, single upstream request), ranked by market cap descending exclude_otc: default True — drop OTC listings (foreign companies traded over-the-counter); "america" otherwise means "US venue", not "US company" compact: default False — True returns only ticker/symbol/price/currency/ change_percent per row (light payload for bulk price feeds) sort_by: market_cap (default) | dividend_yield | change | price — server-side descending sort over the WHOLE market, so e.g. sort_by=dividend_yield with limit=20 is the market's true top-20 dividend payers, not just the biggest companies re-sorted

Returns: Envelope dict: total_matches (market-wide count), returned, and rows of {ticker, symbol, description, exchange, price, open, high, low, currency, change_percent, dividend_yield, market_cap} — price is the current/last close; open/high/low are the current session's daily bar. Prices are in the market's local currency (e.g. KRW for korea).

Return top gainers for an exchange and timeframe using Bollinger Band analysis.

Args: exchange: Exchange name — crypto: KUCOIN, BINANCE, BYBIT, MEXC; stocks: EGX, BIST, NASDAQ, NYSE, BURSA, HKEX, SSE, SZSE, TWSE, TPEX timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M limit: Number of rows to return (max 50)

Returns: list[dict] on success. On ANY failure returns a structured error envelope {"error": {"code": ..., "retryable": ...}} — never a raw exception string.

Return top losers for an exchange and timeframe. Supports crypto (KUCOIN, BINANCE, MEXC) and stocks (EGX, BIST, NASDAQ).

Returns list[dict] on success. On ANY failure returns a structured error envelope {"error": {"code": ..., "retryable": ...}}.

Detect coins with volume breakout + price breakout.

Args: exchange: Exchange name like KUCOIN, BINANCE, BYBIT, MEXC, etc. timeframe: One of 5m, 15m, 1h, 4h, 1D, 1W, 1M volume_multiplier: How many times the volume should be above normal level (default 2.0) price_change_min: Minimum price change percentage (default 3.0) limit: Number of rows to return (max 50)

Returns list[dict] on success, or an error envelope on total upstream failure ({"error": {"code": "ALL_BATCHES_FAILED", ...}}). The empty list now strictly means "no matches today"; rate-limit cliffs surface explicitly.

Detailed volume confirmation analysis for a specific coin.

Args: symbol: Coin symbol (e.g., BTCUSDT) exchange: Exchange name timeframe: Time frame for analysis

Walk-forward backtest to detect overfitting — validates strategy on unseen data.

Args: symbol: Yahoo Finance symbol (AAPL, BTC-USD, SPY…) strategy: rsi | bollinger | macd | ema_cross | supertrend | donchian | keltner_breakout (rsi_pullback and triple_ema not supported here — SMA200 warmup exceeds typical fold size; use run_backtest with period='2y') period: '1mo', '3mo', '6mo', '1y', '2y' (recommend '2y') initial_capital: Starting capital per fold in USD (default $10,000) commission_pct: Per-trade commission % (default 0.1%) slippage_pct: Per-trade slippage % (default 0.05%) n_splits: Number of walk-forward folds (default 3, max 10) train_ratio: Fraction of each fold used for training (default 0.7) interval: '1d' (daily) or '1h' (hourly)

Real-time price quote from Yahoo Finance for any stock, crypto, ETF or index.

Args: symbol: Yahoo Finance symbol — e.g. AAPL, BTC-USD, SPY, ^GSPC, EURUSD=X, THYAO.IS