Parity Probe 01 - Stop/Limit Timing

// This Pine Script® code is licensed under Apache-2.0
// SPDX-License-Identifier: Apache-2.0
// © PineForge contributors 2026
//
// Parity probe 01 — strategy.exit stop/limit fill-bar timing
//
// Purpose: isolate the bar-boundary semantics of strategy.exit's stop=
// and limit= orders. A community-style reverse-on-signal scalper shows a
// consistent pattern where the engine fires Exit-{long,short} ONE 15m bar
// earlier than TV across ~9 of its 14 mismatched exits (e.g. engine 18:00
// vs TV 18:15 on 2025-05-03). The hypothesis is that the engine evaluates
// the stop / limit on the SAME bar as the entry (intra-bar trigger using
// that bar's high/low), while TV defers evaluation to the NEXT bar's open
// onward. Without an isolator, the signal-side cascade in that strategy
// makes it impossible to attribute the drift.
//
// This probe routes a deterministic entry every i_period bars with an
// ATR-derived stop and limit, then lets strategy.exit handle both. No
// MTF, no signals, no regime filter — every exit row in the export is
// either a stop fill, a limit fill, or a forced flatten. Comparing the
// engine's exit timestamps against TV's row-by-row exposes the timing
// semantic directly, with no other moving parts to confound it.
//
// Trade shape: long-only, single contract, pyramiding=0,
// process_orders_on_close=false (TV default), no commission/slippage.
//
// TV setup: 15m chart on BINANCE:ETHUSDT.P, same window as
// data/ohlcv_ETH-USDT-USDT_15m.csv. Export the strategy tester's
// "List of Trades" CSV to ``tv_trades.csv`` in this folder.
//@version=6
strategy("Parity probe 01 - stop/limit timing", shorttitle="par_p01", overlay=true, initial_capital=1000000, commission_type=strategy.commission.percent, commission_value=0, slippage=0, default_qty_type=strategy.fixed, default_qty_value=1, pyramiding=0, process_orders_on_close=false)

// ---- inputs --------------------------------------------------------
int   i_atr_len = input.int(14,  "ATR Length")
float i_stop_x  = input.float(1.0, "Stop = N x ATR", minval=0.1)
float i_tp_x    = input.float(2.0, "TP   = N x ATR", minval=0.1)

// ---- engine -------------------------------------------------------
float atr_val = ta.atr(i_atr_len)

// Wall-clock entry trigger: 00:00 and 12:00 UTC every day. We use
// the bar's UTC time (not bar_index) so engine and TV fire on the
// SAME absolute bars regardless of how many warmup bars precede
// each side's chart-bar-0. bar_index-based triggers would shift the
// firing bar by the warmup-prefix length and produce zero matches
// in `verify_corpus.py`'s align-by-time stage.
bool at_entry_window = (hour == 0 or hour == 12) and minute == 0
bool fire = at_entry_window and strategy.position_size == 0 and not na(atr_val)

if fire
    float stop_lvl = close - atr_val * i_stop_x
    float tp_lvl   = close + atr_val * i_tp_x
    strategy.entry("L", strategy.long, comment="periodic entry")
    strategy.exit("X", from_entry="L", stop=stop_lvl, limit=tp_lvl, comment_loss="stop fill", comment_profit="tp fill")