Five books, named for what they hunt.
A book is a rule set with a cash floor. The code tells you the mechanism before you read a single number: a wick, a level, a range, a sigma band. Two of them trade the same idea on both indices, which is why they share a root and differ by one letter.
| Book | Underlying | DTE | Trades | Trades/mo | Win | PF | Net $ | Net/mo | Max DD | Cash floor |
|---|---|---|---|---|---|---|---|---|---|---|
| SkyQ WICK | SPY | 1 | 651 | 50.9 | 83.3% | 2.05 | 10,184 | 850 | 736 | 1,863 |
| SkyQ LEVEL | QQQ | 1 | 1,782 | 139.2 | 76.4% | 1.49 | 29,993 | 2,493 | 3,297 | 5,904 |
| SkyQ RANGE | SPY | 2 | 2,998 | 234.2 | 84.5% | 1.37 | 11,542 | 1,154 | 607 | 2,467 |
| SkyQ SIGMA-S | SPY | 3 | 2,018 | 157.7 | 86.4% | 1.65 | 16,433 | 1,454 | 896 | 1,746 |
| SkyQ SIGMA-Q | QQQ | 2 | 1,802 | 140.8 | 86.0% | 1.55 | 16,931 | 1,475 | 946 | 2,121 |
| Portfolio | 5 books | — | 9,251 | 722.8 | 83.6% | — | 95,038 | 7,426 | 6,482 | 14,101 |
Bar length is net income, not capital efficiency. LEVEL earns the most and needs the most room.
Win rate is not the ranking
RANGE wins 84.5% of its trades and still has the lowest profit factor of the five, because a typical winner is smaller than a typical loser. SIGMA wins more often still, at a better factor, but it trades a three-day option whose stop is wider. The books are ranked by what they contribute after fees and after the capital they lock up — not by how often they are right.
What each one is actually betting on.
Mechanism, instrument and the specific way each book fails. A strategy description that cannot name its own failure mode is not a description, it is a brochure.
The bar where the wicks carry the information
A completed five-minute bar whose combined wick length clears ATR(14) while its body stays above a tenth of ATR. The next open is the entry, the bracket is ATR-scaled, and the trade is flat inside the session. 651 trades in 12.8 months — the lowest frequency on the desk and the highest profit factor at 2.05.
Betting on: a bar that rejects its own extreme is followed by continuation at the next open, often enough to pay for the debit.
Fails when: the setup is rare, so a bad quarter is a small sample; and in a regime where a long wick marks exhaustion rather than continuation, the edge inverts.
A rolling low, armed as a stop order
No indicator: a rolling minimum of typical price, offset by a fixed number of bars, armed as a stop order one bar at a time. Close below the level, buy the reclaim; close above it, sell the breakdown. The bracket is a fixed percentage of the level and goes in with the entry. 1,782 trades, 76.4% wins, PF 1.49, and the largest cash floor on the desk at $5,904.
Betting on: a level that price has respected becomes a decision point, and the first move through it runs for minutes rather than reversing.
Fails when: the tape chops straight through the level in both directions — the stop entry fills at the worst tick of the rotation, repeatedly.
Buy the bottom decile, sell the top
A channel built from a normalised-price volatility measure. Long in the bottom tenth of the channel, short in the top tenth, target a fixed fraction of the channel width, stop just outside it. 2,998 trades — 234 a month, the busiest book — 84.5% wins, PF 1.37, and the smallest drawdown on the desk at $607.
Betting on: intraday extremes of a volatility channel revert more often than they break, and the reversion is fast enough to exit before the debit decays.
Fails when: the session trends. The channel edge gets run over and the stop is hit at the outside of the range on trade after trade.
Session VWAP, and how far price has strayed
A session-anchored VWAP with one- and two-sigma bands. Buy the lower band, sell the upper, target a fraction of the way to the opposite band, stop on ATR, refuse entries when the bar is stale. 2,018 trades, 86.4% wins — the best win rate on the desk — PF 1.65, floor $1,746.
Betting on: price stretched away from the session's own average trades back toward it, and the reversion usually arrives before the option loses its premium.
Fails when: it is a trend day. Price pins outside the band, the stop carries the loss, and the book takes several of them in a row.
The same band logic, on the higher-volatility index
Identical rule set to SIGMA-S, re-tested on QQQ at a shorter expiry: 1,802 trades, 86.0% wins, PF 1.55, $1,475 a month on a $2,121 floor. It shares the mechanism but not the trades — the two indices produce different band touches on the same minutes, which is the point of running both.
Replay 2025-10-06 → 2026-10-31, one contract per signal, long premium only, net of $0.21 per round trip and the $10 monthly platform fee. Cash floor = maximum concurrent premium + that book's maximum drawdown. Hypothetical throughout; no live or paper fills exist yet.
Why these codes, and where they came from.
Systematic firms name the method once and let the books carry short codes. The code should survive a due diligence call: you should be able to guess the mechanism from the name before reading the deck.
| Code | Mechanism | Instrument | Trades | Lineage |
|---|---|---|---|---|
| WICK | Rejection-wick continuation | SPY · 1 DTE | 651 | Published five-minute wick rule, re-derived on OPRA prints |
| LEVEL | Reclaim or breakdown of a rolling level | QQQ · 1 DTE | 1,782 | HPotter's Iron Support Resistance Automat |
| RANGE | Channel mean reversion | SPY · 2 DTE | 2,998 | Published channel scalper, re-parameterised for one contract |
| SIGMA-S | VWAP band reversion | SPY · 3 DTE | 2,018 | NQ Phantom Scalper Pro, ported to SPY |
| SIGMA-Q | VWAP band reversion | QQQ · 2 DTE | 1,802 | Same port, re-tested on QQQ |
One root per mechanism
If two books trade the same idea, they share the root and differ by index. A reader learns four mechanisms, not five names.
No invented mythology
No Roman gods, no planets, no animals. The code states the pattern being traded, which is also what survives an operational review.
Lineage stays visible
Every book descends from a published study. Saying so is cheaper than being asked, and the full-history test is what separates a port from a copy.