1998issue C091-10
Percentage back-adjusted futures series
TradersWeek editorial: a continuous futures series is an execution-cost construction, not a chart convenience. How rolls are spliced and how bars are scaled decides whether historical profit and loss is comparable and whether a printed price can be treated as executable.
- TradersWeek editorial: a continuous futures series is an execution-cost construction; the roll splice and the bar scale decide whether historical profit and loss is comparable.
- A nearest-futures series injects false roll gaps because adjacent months trade at different prices, while a perpetual-forward series can smooth those gaps only by printing prices that were never traded.
- A merged back-adjusted series matches the dollar result of a one-contract buy-and-hold that rolls at each expiration, but a constant-contract test still misstates percentage risk when face value changes.
- A querged series adds each merged open, high, low, and close change as a fraction of the prior querging close; the wrong close can misalign dates or replace the divisor with a non-traded price.
Why the splice is an execution problem
TradersWeek editorial: a continuous futures series is an execution-cost construction, not a chart convenience. How you splice rolls and scale bars decides whether historical profit and loss is even comparable.
A fixed dollar loss on an index future is a much larger percentage of contract face value when the index level is low than when the same contract later trades at a much higher level.
Four common continuous-series constructions
Four common continuous-series constructions are a single-contract series, a nearest-futures series, a perpetual-forward series, and a merged back-adjusted series.
A single-contract series is one expiration’s full life, from thin early trade through last print. It typically starts with insufficient liquidity for reliable system tests and only becomes usable once that month is the active contract.
A nearest-futures series concatenates contracts at a chosen roll date, so month-to-month spreads appear as price gaps. Adjacent months trade at different prices, and the join injects a false roll gap: a chart discontinuity from splicing two expirations, not from tradable market movement. Those gaps can mislead indicators and system reports.
A perpetual-forward series is a weighted blend of nearby contracts looking a fixed horizon ahead. It can smooth roll gaps, but it prints prices that were never traded, so those prices need not be executable and a stop-style breakout cannot be mapped to an exact executable order price.
A merged back-adjusted series adds the old-to-new difference at the rollover so a one-contract buy-and-hold that rolls at each expiration matches the continuous-series dollar result.
Dollar match is not percentage match
Face-value sizing chooses exposure from allocated capital and leverage first, then converts to contract count at the then-prevailing multiplier.
Holding allocated face value fixed while the underlying level and point multiplier change produces different integer contract counts across eras, so a constant-contract test misstates percentage risk.
The merged back-adjusted series can match the dollar result of a one-contract rolled hold to expiration. It does not by itself keep percentage exposure comparable when face value changes.
Build a querged series from merged bars
A querged series is a percentage-scaled continuous open-high-low-close series built from the merged moves. Add each merged open, high, low, and close change divided by the prior querging close, then multiply by a positive scale factor.
The querging close is the unadjusted prior close of the contract used to scale that bar’s merged move into a percentage increment.
Choose the querging close with the same roll rule
Using the nearest-contract close as the querging close can misalign dates if the merged series rolls on a different rule than the nearest-futures series.
Using a perpetual-forward close as the querging close substitutes an averaged, non-traded price for the percentage divisor. That is a rough first approximation mainly when month-to-month gaps are small.
All readings on this track · 8 readings
- 1989Treasury auction-rollover as yield-to-price series construction
- 1991Calendar spread execution for futures month rolls
- 1992Converting OHLC, session, and rollover series for analysis
- 1992Linked futures series for mechanical system evaluation
- 1998Percentage back-adjusted futures series
- 2001Construct a volume-weighted continuous-lead contract
- 2013Construct a continuous futures series from the roll you can execute
- 2018Commodity-basket exposure as a contract-design problem