How this calculation works
The factor b acts once per unit of x. For a two-unit interval, the ratio of the y-values is b².
b = exp[(ln y₂ − ln y₁)/(x₂ − x₁)], A = y₁/bˣ¹A worked example
(0,3) and (2,12) give y = 3 × 2ˣ.
Good to know
Use distinct x-values and positive y-values.
Reference: OpenStax Algebra and Trigonometry (opens in a new tab) ↗
Recover the per-unit multiplier from the endpoint ratio
For points (0,3) and (2,12), the y value grows by a factor of 12/3 = 4 across two x units. A constant per-unit multiplier b therefore satisfies b² = 4, giving b = 2. Because the first x-coordinate is zero, the coefficient A in y = Abˣ is immediately 3.
The fitted model is y = 3 × 2ˣ. It gives 3 at x = 0, 6 at x = 1 and 12 at x = 2. For arbitrary distinct x-values the calculation uses log growth rate k = [ln(y₂) − ln(y₁)]/(x₂ − x₁), then b = eᵏ and A = e^[ln(y₁) − kx₁].
Interpret the time or distance unit behind the multiplier
A multiplier is tied to one x unit. If x measures years, b is the yearly multiplier. Changing to months would replace it by its twelfth root. A value above one describes growth, a value between zero and one describes decay, and b = 1 describes a constant positive level.
Both observations must have positive y-values because the method takes their logarithms. The x-values must differ. The Single-base exponential equation calculator can find where the fitted model reaches a positive target after dividing that target by A. Two points establish an exact exponential interpolant, while assessing whether future observations follow it requires additional evidence.
A few questions, answered
Does reversing the order of the points change the fit?
No. Both the logarithmic difference and the x difference reverse sign, preserving their ratio.
Is the log growth rate the same as the multiplier?
The rate k is ln(b). The multiplier is b = eᵏ. A multiplier of 2 has log rate ln 2.