Strategy family 09 / 12
Relative value & spread trading
Relative value trading holds one asset against a close relative and takes a view on the spread between them, not on the direction of either. In futures markets this means the basis between cash and futures, calendar spreads along a single curve and inter-commodity spreads such as refining and crushing margins; across markets and instruments, it means the same exposure priced in two places, such as dual-listed shares, and related assets priced against each other, such as share classes or a fund and its holdings. Market direction largely cancels out, which leaves the spread and the economics behind it as the main source of risk.
Covers
- Cash–futures basis
- Calendar spreads
- Inter-commodity spreads
- Cross-market spreads
- Related-asset relative pricing
Why it is a demanding test
Spread trades look safe because the two legs usually move together, until they do not. This family tests whether a system understands the economic link that holds a spread in place, including financing, delivery, storage and conversion, and whether it can tell a true mispricing from a spread that has moved for a good reason. Thin margins on each trade, frequent use of leverage and convergence that can arrive late or not at all make it a demanding test of cost accounting and risk control.
What we test
What a system needs to do well in this family. Tasks are built so that each ability can be observed and scored on its own.
Pricing the no-arbitrage relationship
Computing the theoretical basis or calendar spread, or the no-arbitrage band around it, from financing, dividends or coupons, storage and delivery terms, and reading what any remaining gap implies, such as an implied repo rate or convenience yield.
Knowing delivery and contract mechanics
Understanding how the cheapest-to-deliver bond, delivery grades and their premiums or discounts, delivery options, settlement method, roll timing and contract specifications shape a spread, especially as expiry approaches.
Building spreads from physical economics
Constructing inter-commodity spreads from processing relationships such as refining or crushing, with correct conversion ratios and units, and recognizing when the physical link weakens.
Bridging markets and their frictions
Comparing the same or similar exposure across exchanges and jurisdictions while adjusting for currency, trading hours, capital controls, quotas, taxes and settlement, all of which can keep a gap open.
Sizing for delayed or failed convergence
Setting position size, leverage and exit rules so that a position can survive a spread that keeps widening, including the margin calls and funding pressure that come with it.
Example task types
Illustrative designs that show the shape of the work. They are not items from the benchmark itself.
- Track: Analysis
Explain the basis of a stock index future
The system receives futures and index prices, the expected dividend schedule and short-term funding rates for a stock index future on a given date. It must compute the theoretical basis, explain any gap to the observed basis in terms of dividends, funding, hedging demand or short-selling constraints, and judge whether arbitrage between the future and the underlying basket is feasible after costs. A good answer gets the carry arithmetic right and does not assume that every gap can be traded.
- Track: Simulation
Trade a crack spread through changing seasons
In a historical replay, the system trades the spread between crude oil and refined product futures at a stated conversion ratio, with exchange margin, roll schedules and transaction costs applied. It decides entries, position size and exits as inventories, seasonal demand and refinery outages move the spread. A good answer keeps the legs in the right ratio and units, rolls each contract on time, and meets margin calls when the spread moves against it.
- Track: Live
Forecast calendar spreads from live fundamentals
Each week the system forecasts the change over the following month in a set of commodity calendar spreads, as a probability distribution rather than a single number. Forecasts are scored against live outcomes over rolling windows. A good answer ties each forecast to inventories, seasonality and delivery timing, and its intervals are neither too narrow nor too wide.
- Track: Client
Explain a price gap between two listings
A user notices that one company's shares trade at different prices on two exchanges and asks whether buying the cheaper listing is close to a sure gain. The system must explain what keeps such gaps open, including whether shares can move between the two markets, currency, capital controls, liquidity and different investor bases, and why buying one side alone is not a hedged position. A good answer corrects the misconception in plain terms, notes that gaps can persist or widen for long periods, and relates the choice to the user's goals and constraints.
How it is scored
Each run is scored on its outcome and on the process behind it, so a lucky result is not mistaken for a sound one.
- 01
Risk-adjusted return on capital used
Returns are measured after financing costs against the margin and capital a position actually ties up, and reported per unit of risk taken. Leverage can make a thin spread look attractive on margin alone, so risk-adjusted figures are what keep positions with different leverage comparable.
- 02
Captured share of the mispricing
We measure how much of the gap identified at entry was realized after costs. A low capture rate points to poor execution, poor timing or a mispricing that was never real.
- 03
Maximum adverse excursion and survival
We record the largest move against each open position and whether the system's sizing and margin buffer would have withstood it. In this family, surviving adverse moves is a precondition for any average outcome to matter.
- 04
Carry and hedge-ratio arithmetic
Theoretical bases, carry costs, conversion ratios and hedge ratios are checked against expert references, including units, day-count conventions and contract multipliers. Leverage magnifies small arithmetic errors, so they are scored separately.
- 05
Leg ratio, roll and execution discipline
Rubric-based checks test whether legs stay in the intended ratio, whether rolls happen before notice or delivery periods, and whether execution allows for legging risk when both sides cannot be filled at once.
Diagnostics
Failure modes we diagnose
Recurring mistakes the tasks are built to surface. Each one is reported separately, so builders can see which to fix first.
- 01
Assuming convergence is guaranteed
Treating a relationship that has held in the past as one that must hold, and sizing as if the spread cannot widen further. Spreads can widen sharply when the traders who usually close them are forced to unwind.
- 02
Mistaking carry for mispricing
Calling a basis a mispricing without accounting for financing, dividends, storage, or the cost and availability of borrowing the cash leg. Much of an apparent gap is often fair compensation for carry.
- 03
Mismatched units, sizes and ratios
Mixing contract sizes, units, currencies or quoting conventions so that the legs are not actually matched. A position that looks hedged can carry a large outright exposure.
- 04
Drifting into delivery or expiry
Missing a roll date, first notice day or expiry rule, and ending up with a delivery obligation or a forced exit at a poor price.
- 05
Missing a structural break
Trading a spread back toward its historical average after the underlying link has changed, for example through new regulation, a shift in physical trade flows or a revised contract specification.
Environment and data
Tasks use futures curves with full contract specifications, exchange margin schedules and roll calendars, alongside point-in-time cash prices, funding rates, dividend forecasts and inventory data. Cross-market tasks add currency conversion, differences in trading hours and the access rules of each market. Simulations charge bid-ask spreads and fees on every leg and borrow costs where a cash leg is sold short, and include legging risk when legs cannot be filled together, margin calls and the cost of financing positions over time.