A continuous futures series joins or blends observations from successive contracts under a stated selection rule. Its values depend on that rule, any adjustment and data timing; they are not automatically executable prices or investment returns.
Separate the reference from its constituents
An individual futures contract has a finite life, while a continuous series can extend across many expirations. The series therefore needs a rule identifying which contract supplies each observation. Its label describes a statistical construction, not a perpetual exchange-listed obligation. Before interpreting a change, determine whether both observations came from the same contract or whether the reference changed its constituent.
CME's continuous price-series description distinguishes active-contract and front-contract views. This is direct evidence that continuous does not identify a unique selection method. Different legitimate constructions can produce different values for the same economic market. A comparison should name its construction rather than implying that one uninterrupted price history exists independently of these choices.
Recognize front and active selection
Front generally refers to a nearby expiry, subject to the provider's switching convention. Active selection instead emphasizes a specified liquidity criterion. These descriptions can point to different contracts during a roll transition. The rule needs to explain eligible months, the switching trigger and whether selection uses information already available. A familiar generic symbol does not answer those questions by itself.
The EIA's futures-price definitions define Contract 1 by the earliest delivery date and describe the crude-oil expiry convention. Its historical RCLC1 reference is labelled Cushing, Oklahoma crude oil future Contract 1 in dollars per barrel. That identity supports reference-series interpretation, not a claim that a particular account held or filled each constituent contract.
See what a switch can introduce
Consider hypothetical NYMEX Light Sweet Crude Oil months A and B. On one date their reference prices are $70 and $72 per barrel; on the next date they are $71 and $73. Both contracts have increased by $1. A stitched series that displays A first and B next moves from $70 to $73, creating a displayed increase of $3.
The displayed increase combines a within-contract movement with a difference between delivery months. It is not evidence that either contract rose by $3. The numbers also do not describe a roll fill or account result. Identified observations for both contracts are what make the decomposition possible; a single generic column without constituent data cannot establish the same attribution.
| Hypothetical observation | Month A | Month B | Selected reference |
|---|---|---|---|
| First | $70/barrel | $72/barrel | A: $70/barrel |
| Second | $71/barrel | $73/barrel | B: $73/barrel |
| Change | $1/barrel | $1/barrel | $3/barrel |
Distinguish stitching from blending
A stitched series selects one contract at a time. A blended series combines several contracts with explicit weights. Suppose two hypothetical reference levels are 70 and 74 in the same quote unit. A blend containing 75% of the first and 25% of the second equals 71. Equal weights produce 72. Changing the weights changes the reference even if both contract prices stay fixed.
That weighted average is a constructed price statistic. It is not automatically the value of a self-financing portfolio or a return index. Portfolio bookkeeping would need to account for positions, rebalancing and a capital denominator separately. A blended line can be useful for a defined analytical purpose, but its smoothness does not remove the need to describe how it was formed.
Document adjustment as another layer
After choosing constituents, a provider may alter earlier levels to reduce visible switch gaps. Selection and adjustment are distinct operations: the first chooses observations, while the second transforms their values. Two datasets can use identical roll dates yet disagree because one is adjusted. Conversely, two unadjusted datasets can disagree because their switch schedules select different contracts on the same date.
The description should therefore state whether levels are unadjusted, additively adjusted, multiplicatively adjusted or constructed another way. It should identify the adjustment anchor and how later rolls affect earlier history. These details determine which arithmetic is meaningful. A label such as continuous close is too incomplete to establish whether price differences, ratios or fixed thresholds preserve their intended interpretation.
Keep historical reference use bounded
A historical public reference such as RCLC1 can illustrate how deterministic directional formulas react to a documented sequence. Its use should disclose dates, units, missing observations and the construction limitation. Referencing the official source is different from republishing its raw history as a public mirror. An educational discussion can present synthetic worked examples and source links without distributing the underlying dataset.
Nor should a reference-only exercise imply actual contract ownership, broker execution or investment performance. A two-observation signal delay does not supply those missing records. The bounded question is how the stated rule labels the chosen reference. That remains a useful question when the result is clearly separated from any claim about trading the changing contract constituents.
Read the construction before the chart
A compact provenance description should identify provider, series code, price field, units, eligible dates, selection policy and adjustment status. When available, retain the constituent identity and switch markers. Missing documentation should be described as missing rather than inferred from the shape of the line. Smooth prices do not prove adjustment, and visible gaps do not prove data errors.
The most useful comparison starts with the question each series can answer. An unadjusted front reference describes successive selected quotes. An adjusted series emphasizes a chosen continuity property. A contract-level record identifies a particular expiry. Treating these as complementary but different objects prevents a convenient chart from acquiring evidential powers that its underlying data construction does not support.
Research checkpoints
- Continuous is a family of construction methods, not a unique contract.
- Contract switches and changing blend weights can move the reference without equivalent contract movement.
- A historical reference is not evidence of actual fills or investment returns.
Source record
Related methods
Contract Identity and Expiry: Know What a Futures Price Describes
Identify exchange, product, delivery month, units and settlement terms before interpreting a futures observation.
Back-Adjustment and the Return Traps Hidden in Smooth Charts
Compare additive and ratio adjustment with checked examples, and learn which differences, ratios and signals remain meaningful.
Contango, Backwardation and Roll: Read the Curve Carefully
Separate a futures curve observed at one instant from price changes through time, roll transactions and realized outcomes.
Published 16 October 2026. Illustrations and reference diagnostics do not establish strategy profitability or execution quality.