feat: shadow book + shadow-vs-manual performance comparison

The manual paper book only contains trades taken by hand, inside a 20
minute window, on days someone was available. The backtest that validated
this strategy auto-takes the top-ranked qualified setups up to capacity
every session. The forward record was therefore measuring strategy plus
discretion plus availability -- and degrading silently on busy days.

The shadow book closes that gap: it mirrors the backtest's selection rule
(top strategy_rank qualified, up to capacity, 1% fixed-fractional risk)
and shares the manual book's exit policy, so the only difference between
the two books is which setups get taken. Selection ordering reuses the
strategy_rank the scanner already stores rather than recomputing it, so
the two cannot drift apart. It runs as a near-close pipeline step right
after the scan, marking entries at the same prices a human would see.

Gate-reset re-entry state is now scoped per book -- the books diverge as
soon as their entries differ, and each must see only its own stops.

Performance view rewritten around the comparison:
  - three series (shadow, manual, SPY) from a new endpoint
  - SPY changes from a per-trade cost-basis counterfactual to plain
    buy-and-hold %, since one line has to serve two books
  - headline stats are R-multiples, not currency: the books size
    differently, so only R compares across them
  - configurable start date, because the strategy has been revised
    repeatedly and pre-cutover trades ran under rules that no longer
    exist

Migration 024 also repairs the numeric weekday crons written by 023,
rewriting only rows still holding the broken form so hand-corrected
settings survive. Its literals are inlined because bound parameters
render as NULL under 'alembic upgrade --sql'.

The shadow book is opt-in and writes nothing until enabled. Verify its
first selections match a backtest of that day's cross-section before
trusting any point on the curve.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

app/services/paper_trade_service.py

@@ -3,6 +3,7 @@
 from __future__ import annotations
 
 import bisect
+import logging
 from datetime import date, datetime, timezone
 
 from sqlalchemy import and_, func, select
@@ -20,7 +21,9 @@ from app.services.outcome_service import (
     Bar,
     evaluate_setup_against_bars,
 )
-from app.services.trade_policy import get_reentry_gate_locks
+from app.services.trade_policy import MANUAL_BOOK, SHADOW_BOOK, get_reentry_gate_locks
+
+logger = logging.getLogger(__name__)
 
 # Exit policy for OPEN paper trades (auto-close). Production defaults to the
 # July 2026 promoted strategy: initial stop + 3x ATR trailing stop, with a max
@@ -690,6 +693,66 @@ def build_equity_curve(
     return out
 
 
+KEY_PERFORMANCE_START = "performance_start_date"
+
+
+async def get_performance_start(db: AsyncSession) -> date | None:
+    """Date the performance view starts from, or None for 'all history'.
+
+    The strategy has been revised repeatedly, so early trades were taken under
+    rules that no longer exist. Pinning a start date keeps the comparison inside
+    one regime instead of averaging across configurations that were replaced.
+    """
+    raw = await settings_store.get_value(db, KEY_PERFORMANCE_START, "")
+    if not raw or not str(raw).strip():
+        return None
+    try:
+        return date.fromisoformat(str(raw).strip())
+    except ValueError:
+        logger.warning("invalid %s: %r", KEY_PERFORMANCE_START, raw)
+        return None
+
+
+def trade_r_multiple(trade, mark: float | None) -> float | None:
+    """Result in R — profit measured in units of the trade's own initial risk.
+
+    R is the only sizing-independent yardstick available here: the shadow book
+    sizes at a fixed 1% of equity while manual trades were sized by hand, so
+    currency P&L cannot compare them. Open trades are marked to ``mark``.
+    """
+    risk_per_share = abs(trade.entry_price - trade.stop_loss)
+    if risk_per_share <= 0:
+        return None
+    exit_price = trade.close_price if trade.status == "closed" else mark
+    if exit_price is None:
+        return None
+    per_share = (
+        exit_price - trade.entry_price
+        if trade.direction == "long"
+        else trade.entry_price - exit_price
+    )
+    return per_share / risk_per_share
+
+
+def book_stats(trades: list, marks: dict[int, float]) -> dict:
+    """Sizing-independent summary of one book: counts, win rate, R-multiples."""
+    rs = [
+        r
+        for r in (trade_r_multiple(t, marks.get(t.ticker_id)) for t in trades)
+        if r is not None
+    ]
+    closed = [t for t in trades if t.status == "closed"]
+    wins = [r for r in rs if r > 0]
+    return {
+        "trades": len(trades),
+        "closed": len(closed),
+        "open": len(trades) - len(closed),
+        "win_rate": round(100.0 * len(wins) / len(rs), 1) if rs else None,
+        "total_r": round(sum(rs), 2) if rs else 0.0,
+        "avg_r": round(sum(rs) / len(rs), 3) if rs else None,
+    }
+
+
 async def equity_curve(db: AsyncSession, user_id: int) -> list[dict]:
     """Equity-curve series for a user's paper book (empty without benchmark data)."""
     trades = (
@@ -714,3 +777,112 @@ async def equity_curve(db: AsyncSession, user_id: int) -> list[dict]:
     for tid, day, close in rows.all():
         ticker_closes.setdefault(tid, {})[day] = float(close)
     return build_equity_curve(list(trades), ticker_closes, benchmark_closes)
+
+
+def _cumulative_pnl(trades: list, ticker_closes: dict, days: list[date]) -> list[float]:
+    """Cumulative realized + mark-to-market P&L of one book on each day."""
+    sorted_dates = {tid: sorted(c) for tid, c in ticker_closes.items()}
+    out: list[float] = []
+    for d in days:
+        total = 0.0
+        for t in trades:
+            if t.opened_at.date() > d:
+                continue
+            closed_on = (
+                t.closed_at.date()
+                if (t.status == "closed" and t.closed_at is not None)
+                else None
+            )
+            if closed_on is not None and closed_on <= d and t.close_price is not None:
+                ref = float(t.close_price)
+            else:
+                ref = _value_on_or_before(
+                    sorted_dates.get(t.ticker_id) or [],
+                    ticker_closes.get(t.ticker_id) or {},
+                    d,
+                )
+                if ref is None:
+                    continue
+            per_share = (
+                ref - t.entry_price if t.direction == "long" else t.entry_price - ref
+            )
+            total += per_share * t.shares
+        out.append(round(total, 2))
+    return out
+
+
+async def performance_summary(db: AsyncSession) -> dict:
+    """Shadow book vs discretionary book vs SPY, from the configured start date.
+
+    Currency P&L is reported per book but is *not* the comparison — the books
+    size differently, so the honest read is the R-multiple stats. SPY is a plain
+    buy-and-hold reference over the same window rather than a per-trade
+    counterfactual, so one line serves both books.
+    """
+    start = await get_performance_start(db)
+    stmt = select(PaperTrade)
+    if start is not None:
+        stmt = stmt.where(func.date(PaperTrade.opened_at) >= start)
+    trades = list((await db.execute(stmt)).scalars().all())
+
+    benchmark_closes = await benchmark_service.load_benchmark_closes(db)
+    empty = {
+        "start_date": start.isoformat() if start else None,
+        "series": [],
+        "stats": {},
+    }
+    if not trades or not benchmark_closes:
+        return empty
+
+    first = min(t.opened_at.date() for t in trades)
+    if start is not None:
+        first = max(first, start)
+    days = [d for d in sorted(benchmark_closes) if d >= first]
+    if not days:
+        return empty
+
+    ticker_ids = {t.ticker_id for t in trades}
+    rows = await db.execute(
+        select(OHLCVRecord.ticker_id, OHLCVRecord.date, OHLCVRecord.close).where(
+            OHLCVRecord.ticker_id.in_(ticker_ids), OHLCVRecord.date >= first
+        )
+    )
+    ticker_closes: dict[int, dict[date, float]] = {}
+    for tid, day, close in rows.all():
+        ticker_closes.setdefault(tid, {})[day] = float(close)
+
+    books = {
+        MANUAL_BOOK: [t for t in trades if (t.book or MANUAL_BOOK) == MANUAL_BOOK],
+        SHADOW_BOOK: [t for t in trades if t.book == SHADOW_BOOK],
+    }
+    pnl = {
+        name: _cumulative_pnl(book_trades, ticker_closes, days)
+        for name, book_trades in books.items()
+    }
+
+    bench_dates = sorted(benchmark_closes)
+    spy0 = _value_on_or_before(bench_dates, benchmark_closes, days[0])
+    spy_pct = [
+        round(100.0 * (benchmark_closes[d] / spy0 - 1.0), 2) if spy0 else 0.0
+        for d in days
+    ]
+
+    # Latest close per ticker, for marking open positions in the R stats.
+    marks = {
+        tid: closes[max(closes)] for tid, closes in ticker_closes.items() if closes
+    }
+    stats = {name: book_stats(bt, marks) for name, bt in books.items()}
+    for name in books:
+        stats[name]["pnl"] = pnl[name][-1] if pnl[name] else 0.0
+    stats["spy"] = {"pct": spy_pct[-1] if spy_pct else 0.0}
+
+    series = [
+        {
+            "date": d.isoformat(),
+            "manual_pnl": pnl[MANUAL_BOOK][i],
+            "shadow_pnl": pnl[SHADOW_BOOK][i],
+            "spy_pct": spy_pct[i],
+        }
+        for i, d in enumerate(days)
+    ]
+    return {"start_date": start.isoformat() if start else None, "series": series, "stats": stats}