Slow Stochastic Cross

// This Pine Script® code is licensed under Apache-2.0
// SPDX-License-Identifier: Apache-2.0
// © PineForge contributors 2026
//
// PF probe 103 — slow stochastic %K/%D crossover
// Clean-room: written from validation goal, not from source.
//
// Purpose: exercises `ta.stoch(close, high, low, length)` (which
// returns the raw fast %K) plus the standard slow-stochastic
// smoothing chain: `slowK = ta.sma(rawK, smooth)` and then
// `slowD = ta.sma(slowK, dLen)` (D is an SMA of the slow K, not of
// the raw K — otherwise slowK and D would be the same series and
// never cross). Trade trigger uses `ta.crossover` / `ta.crossunder`
// of slowK against slowD, with a band gate to avoid entries at
// extreme overbought/oversold. The engine surface under test is the
// stacked-SMA recurrence and `ta.crossover`/`ta.crossunder` ordering
// against TradingView's reference.
//
// Validation goal: bit-exact trade list against TradingView's broker
// emulator on the reference 15m feed. Pass = every long/short entry
// fires on the same bar TradingView places it.
//
// TV setup: 15m chart, ETH-USDT-USDT (Binance USDT-M Perpetual). Full
// window matching corpus/data/ohlcv_ETH-USDT-USDT_15m.csv. Strategy
// properties: leave defaults from script header. Export "List of
// trades" → tv_trades.csv.
//@version=6
strategy("PF probe 103 - stoch slow cross", shorttitle="PF_p103_STO",
     overlay=false, initial_capital=1000000,
     commission_type=strategy.commission.percent, commission_value=0,
     slippage=0, default_qty_type=strategy.fixed, default_qty_value=1,
     pyramiding=1, process_orders_on_close=false)

// ---- inputs --------------------------------------------------------
int i_k_len   = input.int(14, "Stoch %K length",     minval=1, maxval=100)
int i_smooth  = input.int(3,  "Slow %K smoothing",   minval=1, maxval=20)
int i_d_len   = input.int(3,  "%D smoothing",        minval=1, maxval=20)

// ---- raw %K from ta.stoch, then slow-smoothed --------------------
// ta.stoch in v6 returns the raw fast %K. Slow-stochastic recipe:
//   slowK = SMA(rawK, smooth)         (3 by default)
//   slowD = SMA(slowK, dLen)          (3 by default)
// D is computed from slowK, NOT from rawK, so the two series can
// actually cross.
float raw_k  = ta.stoch(close, high, low, i_k_len)
float slow_k = ta.sma(raw_k,  i_smooth)
float slow_d = ta.sma(slow_k, i_d_len)

// ---- crossover entries with band gate ------------------------------
// Band gate prevents entries at extreme readings (long while already
// at overbought ≥ 80, short while at oversold ≤ 20). Entries flip
// position so longs and shorts can alternate.
bool bull_cross = ta.crossover(slow_k,  slow_d) and slow_k < 80
bool bear_cross = ta.crossunder(slow_k, slow_d) and slow_k > 20

if bull_cross
    strategy.entry("L", strategy.long, comment="slowK cross above D")

if bear_cross
    strategy.entry("S", strategy.short, comment="slowK cross below D")