SPX · carry-to-back · 2017–2026 · account $100k · click a point for structure
show:
Click anywhere on a curve to inspect that week's position (payoff · legs · entry chain). Toggle any
strategy with the checkboxes above (saved locally). Overlays (Medusa / SVOL) aren't clickable. ■ V1■ V4 (adaptive strikes/expiry by VIX, ½ short back call; 13:00 adds an asymmetric Δ-hedge, hedge-once)■ QQQ SVOL■ SPY SVOL (external daily short-vol tracks, integer-scaled). ■ 0DTE PS ×68 (native SPX 0DTE credit put spread, 10:00 tab only; the line & stats are ×68 contracts, integer-scaled to ≈ V4's P&L for comparison — Sharpe/win-rate are scale-invariant. The per-session table & modal stay 1 contract to match the strategy doc).
metric
V1
V4
0DTE PS ×68
Medusa ×5
QQQ+Med 60/40
Sharpe
1.02 [0.35, 1.67]
1.45 [0.80, 2.04]
0.93 [0.01, 2.01]
3.02 [2.41, 3.55]
3.94 [3.32, 4.64]
Sortino
1.39 [0.45, 2.45]
1.82 [0.93, 3.27]
1.23 [0.02, 2.74]
4.44 [2.64, 9.58]
7.39 [5.34, 11.37]
Calmar ($/yr per $DD)
0.71 [0.15, 1.91]
1.13 [0.50, 3.34]
0.68 [0.01, 2.17]
3.32 [3.30, 18.26]
8.93 [4.91, 16.05]
Win rate
66% [61%, 70%]
67% [62%, 72%]
80% [78%, 83%]
83% [76%, 88%]
68% [65%, 71%]
Reward/risk
0.82 [0.66, 1.03]
1.11 [0.83, 1.49]
0.29 [0.26, 0.32]
1.40 [1.08, 1.86]
1.46 [1.29, 1.70]
EV / trade
$1,407 [$494, $2,309]
$1,662 [$927, $2,442]
$608 [-$74, $1,244]
$4,135 [$3,460, $4,820]
$681 [$579, $779]
Max DD
-$102,633 [-$191,208, -$54,668]
-$69,759 [-$109,281, -$31,105]
-$209,508 [-$445,808, -$116,144]
-$35,675 [-$33,550, -$8,000]
-$12,228 [-$22,060, -$8,071]
Avg DD
-$14,360 [-$33,878, -$12,561]
-$9,691 [-$15,068, -$5,790]
-$41,188 [-$119,153, -$31,889]
-$10,222 [-$7,896, -$3,056]
-$2,863 [-$3,920, -$1,998]
Avg DD len
7d [6d, 23d]
8d [4d, 13d]
6d [5d, 38d]
8d [3d, 7d]
3d [2d, 4d]
Max DD len
61d [32d, 226d]
79d [24d, 145d]
45d [23d, 191d]
23d [7d, 33d]
33d [6d, 25d]
Winner MAE (avg)
$0 [$0, $0]
$0 [$0, $0]
—
—
—
Stop @ winner p95
$0 [$0, $0]
$0 [$0, $0]
—
—
—
Final equity
$732,964
$768,312
$770,412
$790,550
$763,544
Trades
450
402
1102
167
974
Corr vs Medusa (Δeq)
-0.06
-0.04
-0.05
—
—
95% CI in brackets (bootstrap; Wilson for win rate). Ratios annualized (√252 daily / √52 weekly-carry). ■ Medusa overlaid + column, integer-scaled ×5 to a $100k book so its final ≈ V4's. Corr vs Medusa = Pearson of weekly Δequity per variant over Medusa's window (~2020-06 → 2026-03); ≈0 = uncorrelated. V4+Med ×5 = both at full size (≈2× capital, pooled trades). ½V4+½Med ×5 = the same two at equal capital (½ each on one $100k book) — directly comparable to V4: similar final equity, far higher Sharpe. Ratios are scale-invariant, so its Sharpe = the full combined's. All Sharpes weekly-52 (√52); combined daily-sampled ann-252 = 1.91.
Full-history combined (2.3) is diluted by the ~3 pre-Medusa years — over 2020–2026 where both are live: V4 1.42 · Medusa 3.02 · Combined 2.96. Medusa's exit-date track has no daily marks, so weekly is the honest cadence; a naive per-trade×√252 would overstate it ~4×. ■ QQQ SVOL / ■ SPY SVOL = external daily short-vol tracks (their cum_equity), each integer-scaled (×4 / 7) to a $100k book so finals ≈ V4's. These have real daily marks, so their Sharpe is honest at any cadence; Corr vs Medusa shows whether they diversify the existing sleeves too. ■ QQQ+Med 60/40 = the max-Sharpe dollar mix from a 0→100% weight sweep (objective: weekly Sharpe over the 2021–2026 overlap where both trade). The two sleeves are ~0-correlated, so the blend beats either alone — overlap Sharpe: QQQ 3.48 · Medusa 2.83 → blend 4.21; full-history weekly 3.94, maxDD -$12,228.