Futures Trading Basics: Contracts, Leverage, Margin and Roll

Futures trading is the systematic use of standardized, exchange-cleared contracts to take leveraged exposure to an asset for delivery or cash settlement at a future date. For quants, futures are the preferred vehicle for most directional and macro exposure: they are capital-efficient, deeply liquid, centrally cleared, and — crucially — they expose the term structure, which is itself a tradeable object. This article treats futures as an instrument to be modeled precisely: contract specification, the margin and mark-to-market mechanics that determine your cash flows, the cost-of-carry relationship that prices the curve, and the roll dynamics that quietly dominate the returns of any held position.

The central reason to understand futures mechanically rather than casually is that the headline price move is rarely the whole P&L. Daily variation margin, roll yield, and funding (for crypto perpetuals) are first-order contributors that a naive backtest ignores and a live account does not.

Contract specification

A futures contract is fully defined by its specification, and you cannot size or risk it without those numbers. The fields that matter:

• Contract size / multiplier — the notional one contract controls. The E-mini

S&P 500 (ES) has a multiplier of 50 USD per index point; crude oil (CL) controls 1,000 barrels.

• Tick size and tick value — the minimum price increment and its dollar value.

ES ticks in 0.25 points, so one tick is 0.25 × 50 = 12.50 USD.

• Expiry and last-trade date — when the contract settles and when you can no

longer trade it.

• Settlement type — physical delivery (CL, agricultural) versus cash settlement

(equity index, most rates).

• Trading and delivery calendar — first notice day matters: holding a

physically-settled contract past it can assign delivery obligations.

The mapping from price to P&L is exact: for ES, a 10-point move on one long contract is 10 × 50 = 500 USD. Everything downstream — stop placement, position sizing in dollar-risk terms — depends on getting tick value right.

Margin and mark-to-market

Futures are leveraged because you post margin (a performance bond), not the full notional. Two thresholds govern the position:

• Initial margin — collateral to open. If initial margin on a 100,000 USD

notional contract is 5,000 USD, you are running 20:1 notional leverage.

• Maintenance margin — the equity floor; breaching it triggers a margin call or

forced liquidation.

The mechanic that newcomers from an equities background miss is daily mark-to-market: the clearinghouse settles gains and losses in cash every day via variation margin. Your paper loss is not deferred until you close — it is debited from your account that evening. A correct long-horizon thesis can still produce a margin call mid-path if you cannot meet interim variation margin. This path dependence is why futures sizing must be done against worst-case interim moves, not just terminal outcomes.

The discipline is to size off notional exposure and your risk budget, never off the margin requirement. Margin tells you what you must post; notional tells you what you actually risk. Anchoring on margin is the single most common path to over-leverage in futures. This connects directly to disciplined position sizing and risk management and to the mechanics covered in leverage and margin.

The term structure and cost of carry

A futures price is tied to spot by cost-of-carry. Under no-arbitrage for a storable, financeable asset:

F(t, T) = S(t) * exp((r + u - y) * (T - t))

where S is spot, r the financing rate, u storage/insurance cost, and y the convenience yield (the benefit of holding the physical asset now). The sign of (r + u - y) determines the shape of the curve:

• Contango — F > S. Futures above spot; the normal state when carry costs

exceed convenience yield (financeable, storable assets).

• Backwardation — F < S. Futures below spot; typical when convenience yield is

high (tight physical supply, strong demand to hold now).

As t approaches T, F converges to S. The path of that convergence is roll yield, and for a continuously-held position it is not a side effect — it is a primary return driver. The curve also connects directly to the carry trade and to cash-and-carry arbitrage, where you lock in F - S by holding spot against a short future when the basis is rich.

Term structure	Curve	Long roll impact	Driver	Typical carry signal
Contango	Upward (F > S)	Negative (sell low, buy higher)	r + storage > convenience yield	Short the carry
Backwardation	Downward (F < S)	Positive (sell high, buy lower)	Convenience yield > carry	Long the carry

Roll and continuous-contract construction

Because contracts expire, maintaining exposure requires rolling: closing the front contract and opening the next. Roll economics follow the curve — rolling in contango costs you, rolling in backwardation pays you. For quants this matters in two distinct places.

First, roll yield can dominate spot returns over long horizons. A commodity ETF perpetually rolling long contracts in steep contango can bleed materially even when spot is flat — a structural drag, not a trading mistake. Annualized roll drag of 5–15% on persistently contango'd commodities is not unusual.

Second, continuous-contract construction is a backtesting trap. When you splice contracts to build a continuous price series for research, the price gap at each roll is a methodological artifact, not a tradeable return. Failing to back-adjust correctly injects fake P&L. The standard fix is panama-canal (difference) or ratio back-adjustment, applied consistently:

import pandas as pd

def back_adjust(front: pd.DataFrame, roll_dates: list, method: str = "diff"):
    """Build a back-adjusted continuous series from stitched front-month bars.

    `front` has a DatetimeIndex and columns ['contract', 'close'].
    `roll_dates` are the dates we switch to the next contract (decision dates).
    We adjust history so the splice introduces no artificial jump.
    """
    s = front.copy().sort_index()
    adj = s["close"].astype(float).copy()
    for d in sorted(roll_dates, reverse=True):
        # gap between the new and old contract at the roll boundary
        idx = s.index.get_indexer([d], method="pad")[0]
        if idx < 1:
            continue
        gap = s["close"].iloc[idx] - s["close"].iloc[idx - 1]
        if method == "diff":
            adj.iloc[:idx] = adj.iloc[:idx] + gap      # shift older history
        elif method == "ratio":
            ratio = s["close"].iloc[idx] / s["close"].iloc[idx - 1]
            adj.iloc[:idx] = adj.iloc[:idx] * ratio
    return adj

Difference adjustment preserves absolute dollar moves (good for P&L); ratio adjustment preserves percentage returns (good for volatility estimation). Pick deliberately based on what you are computing, and never compute roll P&L from the raw spliced series.

Crypto perpetual futures

Crypto's dominant instrument is the perpetual future ("perp"), which has no expiry and therefore no roll. Instead, a periodic funding rate tethers the perp to spot: when the perp trades above spot, longs pay shorts; when below, shorts pay longs. Funding plays the structural role that the curve plays for dated futures.

For a held position, funding is a continuous cash flow you must model as carry. Persistently positive funding is a drag on a held long and an income stream for a hedged short — the entire funding-rate carry strategy monetizes exactly this. A long-spot/short-perp basis position harvests funding while neutralizing price, which is the crypto analogue of cash-and-carry.

Using futures: speculation, hedging, spreads

• Directional and trend exposure. Capital-efficient

trend-following is built almost entirely on futures because of liquidity and breadth across asset classes.

• Hedging. Short index futures to neutralize the beta of a long book without

liquidating it. A fund holding 500,000 USD of S&P-like stocks, with ES notional around 250,000 USD per contract, shorts roughly two contracts to flatten directional exposure, then lifts the hedge cheaply when the concern passes.

• Calendar and inter-commodity spreads. Trading the curve itself (front vs.

back) or related products isolates term-structure and relative-value views with far lower directional risk than outright positions.

A worked hedge ratio matters here: the correct number of contracts is (portfoliovalue × portfoliobeta) / (index_level × multiplier), not a round guess. Sizing the hedge by notional alone, ignoring beta, leaves residual exposure.

Honest limits

Futures are powerful but carry mechanics that punish imprecision:

• Path dependence via variation margin. You can be right on the destination and

still be liquidated on the journey. Stress positions against worst-historical-day interim moves, not just terminal P&L.

• Roll/funding drag is structural. Long-only held exposure in persistent

contango (or positive funding) loses to carry regardless of spot direction. Model it explicitly; do not treat the continuous series as a tradeable return stream.

• Liquidity is concentrated in the front month. Back-month and spread liquidity

can be thin, so realistic transaction costs and slippage differ sharply from the headline front-month spread.

• Leverage amplifies process errors, not edge. Leverage multiplies an existing

edge and equally multiplies the damage of a flawed one; it manufactures nothing.

Used with precise specification awareness, correct margin and mark-to-market modeling, and explicit roll/funding accounting, futures are the most efficient exposure vehicle available to a systematic trader. The discipline is to treat every contributor to P&L — price, variation margin, and carry — as a modeled quantity, and to size off notional and a real risk budget rather than off the deceptively small margin requirement.