Strategies · Geopolitics

Hormuz disruption tilt

About a fifth of the world’s oil passes the Strait of Hormuz. When the ships counted through it fall well below normal, supply risk is real, not a headline, and energy producers and gold have tended to cushion that kind of shock. The counts arrive weekly and few investors read them.

Pre-registered Sep 26, 2026, before it was tested. Results computed Sep 26, 2026.

Open in the builder

Hypothetical backtest; not investment advice; paper trading only.Past results, real or simulated, do not predict future ones. ARCANE places no live orders from this page.

The rule

Trades on the first trading day of each week

While Hormuz transits are 25% or more below normal, hold 40% energy (XLE), 30% gold (GLD) and 30% the S&P 500 (SPY); otherwise hold the S&P 500. Checked weekly.

Where the numbers came from. The 25% line sits inside the World register’s “Watch” band (20–50% below normal). The weights were set by hand, not fitted.

Hold what is below
If the Hormuz transits vs normal is less than -25
then: Split by set weights (40%, 30%, 30%)
40%: Hold XLE
30%: Hold GLD
30%: Hold SPY
Otherwise
Hold SPY

As text

symphony "Hormuz disruption tilt"
description "Tilt to energy and gold while Hormuz traffic is well below normal."
rebalance weekly
benchmark SPY
weight equal
  if ARCANE:chokepoint/hormuz/vs-normal < -25
    weight specified
      40% asset XLE
      30% asset GLD
      30% asset SPY
  else
    asset SPY

How it would have done

Jan 16, 2020 to Sep 25, 2026, against SPY

The whole history

31 trades

A year
+13.5%
Annualized return
In all
+133%
Jan 16, 2020 to Sep 25, 2026
Sharpe
0.73
Return per unit of risk
Worst fall
−34%
Max drawdown
Swings
20%
Volatility, a year
Beta
0.98
To the benchmark

Each decision reads closing prices up to that day and trades at the next day’s close. It trades on the first trading day of each week. Costs: 5 bp of slippage on every trade and 0.3 bp of fees on sales. Uninvested cash earns nothing. Hypothetical: no real orders were placed.

$0.50$1$1.50$2$2.50$3202120222023202420252026Held out

Growth of $1
SPY, held
Held-out period

−40%−20%0%

What it held: SPY 92% of the time; GLD + SPY + XLE 8% of the time.

In sample and the sealed holdout

The holdout was set before the preset was tested, and opened once

Measure	In sample Jan 16, 2020 – Sep 24, 2024	Sealed holdout Sep 25, 2024 – Sep 25, 2026
Annualized return	+14.1%	+12.0%
Sharpe ratio	0.72	0.79
Sortino ratio	1.01	1.16
Max drawdown	33.7%	18.8%
Calmar ratio	0.42	0.64
Volatility	21.5%	16.1%
Alpha against the benchmark	−0.0%	−2.8%
Beta	1.00	0.88
Correlation	1.00	0.90
Winning days	54%	55%
Turnover a year	0.1×	1.2×
Chance the Sharpe is above zero	94%	87%

Is it real?

Checked on the in-sample window only

Overfitting: Not enough to judge
Held its benchmark itself

There is too little history, or nothing to vary, to measure overfitting.

Nudging its numbers changes nothing in this window (the rule acted the same either way, or never acted), so overfitting cannot be measured here.
It held SPY itself almost every day here, so there is no difference to measure.

Backtests counted	3
Deflated Sharpe ratio (0.95 is the house bar)	—
Probability of backtest overfitting (20% is the house ceiling)	—
Walk-forward Sharpe (choosing, then testing)	—
Walk-forward Sharpe with costs doubled	—
Extra return a year against SPY	0.0%

The nudged versions (2)

Change	Sharpe
threshold -25 on chokepoint/hormuz/vs-normal → -27.5	0.72
threshold -25 on chokepoint/hormuz/vs-normal → -22.5	0.72

The house backtest gate

The gate could not run: its in-sample window gives the overfitting check nothing to measure.

The house backtest gate could not run: the in-sample window gives the overfitting check nothing to measure.
In its in-sample window the rule never switched: it held SPY throughout. The holdout is the only evidence it has.

The pre-registration

Written before the backtest; changing any of it makes a new variant

Claim: Holding 40% XLE, 30% GLD and 30% SPY while 7-day Hormuz transits are at least 25% below normal, and SPY otherwise, beats SPY buy-and-hold on Sharpe after 5 bp costs over the sealed holdout.
Why the edge should exist: A sustained fall in physical tanker traffic is slow, public information about oil supply that headline-driven trading under-weights until it shows in prices.
Where it comes from: slow information
Who is on the other side: Market-weight index holders and risk-parity funds that keep their weights through supply shocks.
Why it is not arbitraged away: The signal is weekly, rarely fires and needs AIS data few funds process; the trade is too small and too rare for large funds to build around.
Universe: Hormuz transits vs normal, GLD, SPY, XLE
In-sample window: Jan 15, 2020 to Sep 24, 2024
Sealed holdout: Sep 25, 2024 to Sep 25, 2026
Evaluation: walk-forward, on sharpe, against SPY buy and hold; 13 variants planned
Costs: 5 bp slippage a trade, no commission
Capacity: $1,000,000: XLE, GLD and SPY trade billions a day; a million dollars is a negligible share of volume.
What stops it: Paper Sharpe significantly below the backtest after 26 weeks of signal-on time. PortWatch stops publishing Hormuz counts or changes their method.

The data

Prices: Yahoo Finance (unofficial chart API): adjusted daily closes for GLD, SPY, XLE. Unofficial and not licensed for redistribution. A licensed vendor (Alpaca, Tiingo or Polygon) is a founder decision before the builder is sold.
Hormuz transits vs normal: IMF PortWatch, from Jan 15, 2020. History: revised. Transits a day over 7 days against the same 7 calendar days averaged over every complete year from 2019 to the one before (2019–2023 from 2024 on, the World register’s normal). Each value counts as public the Wednesday after the week it ends in: PortWatch updates weekly. From 2020: 2019 is the first normal. PortWatch revises recent AIS counts, so history is today’s record.