Tutorial: Smart Curves

Smart Curves are expression-based forward curves that derive their values from a base curve and an optional set of named inputs. They are evaluated on demand — you define the rule, and the platform calculates the result.

How Smart Curves Work

A Smart Curve requires:

A base curve (BASE) — an existing forward curve that determines the valuation dates
An expression — a formula evaluated against BASE and any additional named inputs
An object to attach it to before saving

Creating a Basic Smart Curve

//#region
use training

// Create an object and attach a Smart Curve
CORN_TEST = Object()
CORN_TEST.CLOSE = SmartCurve("#MATBAROFEX.ROS.CORN.FUT:CLOSE", "bootstrapCurve(BASE)")
CORN_TEST.category = "TUTORIAL"

save ${object:CORN_TEST}
print "CORN_TEST saved"
//#endregion

Using a Type

Using a typed object is the preferred approach — it categorises your objects and implies a structure:

//#region
use training

CORN_TEST_TYPE = object as #Agriculture
    CLOSE = SmartCurve("#MATBAROFEX.ROS.CORN.FUT:CLOSE", "interpolate(BASE, 'BACKWARD')")
    category = "TUTORIAL"
end

save ${object:CORN_TEST_TYPE}
print "Typed object saved"
//#endregion

Adding Multiple Smart Curves to One Object

//#region
use training

CORN_TEST_TYPE = object as #Agriculture
    CLOSE = SmartCurve("#MATBAROFEX.ROS.CORN.FUT:CLOSE", "interpolate(BASE, 'BACKWARD')")
    PREMIUM = SmartCurve("CORN_TEST_TYPE:CLOSE", "BASE * 1.15")
    category = "TUTORIAL"
end

save ${object:CORN_TEST_TYPE}
print "CLOSE and PREMIUM curves saved"
//#endregion

Here PREMIUM uses the CLOSE curve as its base and applies a 15% markup.

Testing a Smart Curve Before Saving

Use the .build() method to evaluate the curve for a specific date without saving:

//#region
CLOSE = SmartCurve("#MATBAROFEX.ROS.CORN.FUT:CLOSE", "interpolate(BASE, 'BACKWARD')")
result = CLOSE.build("2022-01-12")
print result
//#endregion

Smart Curves with Multiple Inputs

Some curves require more than one input — for example, a mid-price between a bid and an offer.

Add named references as properties on the Smart Curve object:

//#region
use training

// Fetch the existing object
CORN_TEST_TYPE = ${object:"CORN_TEST_TYPE"}

// Create a MID curve between MIN and MAX
CORN_TEST_TYPE.MID = SmartCurve("#MATBAROFEX.ROS.CORN.FUT:MIN", "(BASE + MAX) / 2")

// Add a named reference to the MAX curve
CORN_TEST_TYPE.MID.MAX = ref("data", "#MATBAROFEX.ROS.CORN.FUT:MAX")

save ${object:CORN_TEST_TYPE}
print "MID curve saved"
//#endregion

The expression (BASE + MAX) / 2 references:

BASE — the MIN curve (the primary input)
MAX — the named reference added as a property

Summary

Feature	What it does
`SmartCurve(base, expression)`	Creates a Smart Curve with a formula
`BASE`	Keyword referring to the base curve in the expression
Named properties (e.g. `.MAX`)	Additional curve inputs referenced in the expression
`.build(date)`	Evaluates the Smart Curve for a specific date without saving
`ref("data", "ID:CURVE")`	Creates a reference to another curve on the platform

How Smart Curves Work​

Creating a Basic Smart Curve​

Using a Type​

Adding Multiple Smart Curves to One Object​

Testing a Smart Curve Before Saving​

Smart Curves with Multiple Inputs​

Summary​