Study · updated 2026-07-10draft — not linked from home
Bilbo Box Options — v2
Full rebuild of the compression-box options study after the bias audit. Every entry
waits for a confirmed candle close (nothing a live trader couldn’t have known in the
moment), every fill is priced from real historical option quotes at the actual
timestamps, and the exit rule was chosen using only 2019–22 data, then tested untouched on
2023–26. The system is now trading forward on the
paper account.
The rules full instructions — everything needed to trade it
Hourly candles built on the on-the-hour grid including extended hours (04:00–20:00 ET), but signals are only acted on during regular hours.
“Daily ATR” below means the 14-day average true range as of the prior day’s close.
The setup — the “box”
Wait for the Saty Phase Oscillator to enter compression on the hourly chart (grey candles).
The box is the high and low of the first 5 compression candles (or fewer, if compression ends sooner). After the 5th grey candle the box is frozen — later grey candles don’t widen it.
The entry signal
An hourly candle that closes (a) out of compression AND (b) above the box high. Wait for the candle close — a poke that closes back inside the box is not a signal.
Long only. Breaks below the box are not shorted (tested: the short side is zero-to-negative under every management tried).
Only take signals whose hourly close lands 10:00–15:00 ET, Monday–Friday.
Enter promptly after the confirming close (within ~5 minutes).
Three gates — skip the trade unless ALL pass
Volume: the signal candle’s volume is at least the median volume of the same clock-hour over the prior 20 sessions.
Trend: the stock is above its daily 21 EMA (as of yesterday’s close). Below it, these breakouts historically carry no edge.
Market quality: the option you’re buying has a live two-sided quote with a bid-ask spread of 5% of mid or less. No bid, or a wide market = no trade.
The position
Buy one call, about one strike out of the money — the strike nearest to (current price + 0.75 × daily ATR) — expiring about 28 days out (nearest expiry in the 21–37 day window).
Size: option premium of 3–5% of the account. One position per signal; no adds, no averaging down.
Exits — keyed to the STOCK price, never the option premium
Invalidation: a 5-minute close below the box low → sell immediately.
Trail: once the stock has moved +1.0 × daily ATR in your favor from entry (“armed”), sell on any 5-minute close at or below entry price + 25% of the peak gain. In other words you tolerate giving back up to 75% of the open move — that looseness is what catches the big runs.
Time cap: still open after 10 trading days → sell.
Never use premium-based stops, profit targets, or “cut losers at end of day 1” — every premium-managed variant tested made results worse.
Equity curve $10,000 start · 4% of account per trade
Starting from $10,000, each trade puts 4% of the account into premium.
Fills are modeled halfway between the quote midpoint and the ask (buying) or the bid (selling).
Log scale.
Monte Carlo 10,000 paths · 150 trades (~1 yr) · $10,000 start · 4% sizing
What might the next ~year of this system look like, starting from $10,000?
Each simulated path draws whole trading days at random (with replacement) from the
7½-year record — drawing days rather than single trades keeps the way wins and losses
cluster together. The outer band holds 90% of the 10,000 paths, the inner band 50%, and the
line is the median path.
Read before trading it
The typical trade loses money; the win rate is only ~41%. The loose trail pays for a minority of large winners with strings of small losers — you have to be able to sit through them. Total loss of the premium happened on ~0.2% of trades.
The exit rule passed a genuine out-of-sample test, but the choice of strike and expiration was made looking at the whole record — so treat the exact +12%-per-trade figure as somewhat flattered. The stronger claim is that both halves of the record (2019–22 and 2023–26) were independently profitable.
The equity curve compounds trades one after another; in reality several positions are open at once, so the path math is a simplification.
This is long-only momentum in megacaps: expect it to bleed in bear markets. 2022 was the worst stretch of the record.
No live track record yet — the forward paper record accrues at /bilbo-paper.html.
Research, not advice. Fills from ThetaData historical NBBO minute quotes; costs modeled as stated; dividends/assignment ignored (calls sold before expiry).