What It Means to Return to Baseline
An accumulation function records the net effect of a rate as the input moves away from a chosen starting point. If some parts of the rate are positive and others negative, their contributions can offset one another. The accumulation returns to its baseline precisely when the net signed contribution since the basepoint is zero.
This is a statement about signed area. A region above the horizontal axis contributes positively, while a region below it contributes negatively. The areas need not have the same shape or occur next to each other. What matters is whether their signed total is zero.
The integral is oriented: if \(x<a\), then \(\int_a^x r(t)\,dt=-\int_x^a r(t)\,dt\). The same criterion applies on either side of the basepoint. Most examples below move to the right, so \(x\ge a\).
Positive and Negative Contributions
For a rate \(r\), define its positive and negative parts by \(r_+(t)=\max(r(t),0)\) and \(r_-(t)=\max(-r(t),0)\). Both are nonnegative, and the pointwise identity \(r(t)=r_+(t)-r_-(t)\) holds. Thus, for \(x\ge a\), the positive area accumulated from \(a\) to \(x\) is \(\int_a^x r_+(t)\,dt\), and the magnitude of the negative area is \(\int_a^x r_-(t)\,dt\).
The theorem follows directly from the identity above: the signed accumulation is the positive contribution minus the negative contribution. It is zero exactly when those two nonnegative quantities are equal. This criterion does not require the positive and negative regions to have any particular shape.
More generally, if the accumulation has the same value at two points \(u<v\), then the signed area between those points must be zero:
Therefore \(A(v)=A(u)\) exactly when the positive and negative contributions on \([u,v]\) balance. For a return to the original baseline, the relevant interval begins at the basepoint \(a\).
Finding Returns from Signed Areas
For rates given by a graph, a return can often be found by adding geometric areas in order. Areas above the axis are added; areas below the axis are subtracted. For a step rate, each constant segment gives a rectangle. For a linear segment, the area can be found from a triangle or trapezoid. These are signed-area calculations, not calculations of the total geometric area without signs.
Worked Example: A Step Rate That Cancels Once
Let the basepoint be \(a=0\), and let the rate be \(r(t)=3\) on \([0,2]\), \(r(t)=-2\) on \((2,5]\), and \(r(t)=1\) on \((5,7]\). Find where the accumulation \(A(x)=\int_0^x r(t)\,dt\) returns to zero on \([0,7]\).
From \(0\) to \(2\), the rate is constant at \(3\). The signed area is a rectangle with width \(2\) and height \(3\), so
From \(2\) to \(5\), the rate is \(-2\). At an input \(x\) in this segment, the added area is \(-2(x-2)\), giving
Setting this accumulated value equal to the baseline \(0\) gives \(10-2x=0\), so \(x=5\). Directly checking the signed areas at that input gives
On the final segment, the rate is \(1\), so \(A(x)=x-5\) for \(5\le x\le7\). This is positive for every \(x>5\), and it equals \(2\) at \(x=7\). On the first segment \(A(x)=3x\), which is positive for \(x>0\); on the middle segment it reaches zero only at \(5\). Thus the baseline returns on \([0,7]\) are \(x=0\) and \(x=5\). Substituting \(x=5\) into the piecewise expression confirms \(A(5)=10-2(5)=0\).
Worked Example: Cancellation Between Two Triangular Regions
Suppose a piecewise-linear rate graph has successive vertices \((0,0)\), \((1,3)\), \((2,0)\), \((3.5,-2)\), and \((5,0)\). Take \(a=0\). The graph first forms a triangle above the axis, then a triangle below it. Determine whether the accumulation returns to zero at \(x=5\), and check its value at \(x=3.5\).
The positive triangle has base \(2\) and height \(3\), so its signed area is
Consequently, \(A(2)=3\). The negative triangle has base \(5-2=3\) and depth \(2\). Its signed area is negative, with magnitude
Adding this contribution gives
To check the intermediate point \(x=3.5\), the portion of the negative triangle from \(2\) to \(3.5\) has base \(1.5\) and depth \(2\). Its signed area is \(-\frac{1}{2}\cdot1.5\cdot2=-1.5\). Therefore
This also checks the accumulated value at that point against the area of the two regions traversed so far. The accumulation decreases through the negative triangle, but its area has not yet fully canceled the positive area. It reaches zero at the end of that triangle, \(x=5\), because the two complete signed areas are \(3\) and \(-3\).
Cancellation Is Not the Same as Accumulated Magnitude
The return criterion uses signed area. If instead we add the magnitude of every contribution, negative parts are counted positively too. For example, in the previous example the two triangles have total signed area \(3-3=0\), but their total geometric area is \(3+3=6\). The signed accumulation returns to baseline; the accumulated magnitude does not vanish.
In general, the accumulated magnitude from \(a\) to \(x\) is \(\int_a^x |r(t)|\,dt\). For \(x\ge a\), this quantity is nonnegative and counts both positive and negative rates as contributions in the positive direction. It equals zero only when the rate is zero except possibly on a set of integral zero. By contrast, a signed accumulation can be zero even when substantial positive and negative areas have occurred.
Worked Example: A Return Followed by a Baseline Plateau
Let \(a=0\), with rate \(r(t)=2\) on \([0,1]\), \(r(t)=-1\) on \((1,3]\), and \(r(t)=0\) on \((3,4]\). Find the baseline returns on \([0,4]\) and compare the signed accumulation with the accumulated magnitude.
On the first segment the positive area is \(2\cdot1=2\), so \(A(1)=2\). On the second segment, the negative area from \(1\) to \(x\) is \(-(x-1)\). Hence
At \(x=3\), the signed area is \(2-1\cdot2=0\), so the accumulation returns to baseline. On \([3,4]\) the rate is zero, and the additional signed area is zero at every input. Thus \(A(x)=0\) for every \(x\in[3,4]\). The full piecewise accumulation is \(A(x)=2x\) for \(0\le x\le1\), \(A(x)=3-x\) for \(1\le x\le3\), and \(A(x)=0\) for \(3\le x\le4\). These expressions agree at the shared endpoints: each gives \(2\) at \(x=1\), and each applicable expression gives \(0\) at \(x=3\).
The accumulated magnitude at \(x=4\), however, is
So the signed accumulation at \(4\) is zero, while the accumulated magnitude is \(4\). The zero-rate segment does not create further cancellation; it simply preserves the baseline value already reached at \(3\).
Sign Conditions and the Zero-Rate Pitfall
A rate that is strictly positive or strictly negative throughout an interval moves the accumulation strictly in one direction on that interval, provided the rate is continuous. In particular, if \(r\) is continuous and strictly positive on \([u,v]\), then \(\int_u^x r(t)\,dt>0\) for every \(x\) with \(u<x\le v\). The corresponding integral is strictly negative when \(r\) is continuous and strictly negative. Under these conditions, the accumulation cannot return to its value at \(u\) at an interior point.
It is not enough to say merely that the rate is nonnegative or nonpositive. A nonnegative rate can be zero over a stretch, and the accumulation may then remain unchanged on that stretch. For instance, if \(r=0\) on \([0,1]\) and \(r=1\) on \((1,2]\), then \(A(x)=0\) for every \(x\in[0,1]\), even though \(r\) is nonnegative throughout \([0,2]\). More generally, without strict sign and regularity conditions, the exact test for a return at \(x\) is always the signed-integral condition \(A(x)=0\).
Why the Balance View Is Useful
Thinking in terms of balance makes the starting point of a calculation explicit. A positive region may build up a positive accumulated value; later negative regions reduce it. A return occurs when the reductions have exactly matched the earlier net contribution. If they have not, the remaining difference is the current accumulated value.
This viewpoint also helps separate two different questions: where has the rate contributed, and how much net contribution remains? The first requires tracking the signed regions encountered so far; the second is their signed sum. Knowing that a graph crosses the horizontal axis does not by itself imply that the accumulation is zero there. The areas accumulated before that crossing still matter.
Record the baseline value \(A(a)=0\), or the specified starting value if the accumulation is expressed relative to another baseline.
Count area above the axis positively and area below it negatively, using rectangles, triangles, or trapezoids where appropriate.
At a proposed return point, verify that the signed total from the basepoint is zero.
If the rate is zero after a return, the accumulation stays at baseline throughout that interval; it does not need a new positive-negative cancellation there.
Check Your Understanding
Use signed contributions from the stated basepoint in each question.
- A rate is \(4\) for \(1\) unit of input, followed by a rate of \(-2\). How far must the second segment extend for the accumulation to return to its baseline?
- A graph has accumulated positive area \(7\) and negative area of magnitude \(5\) by an input \(x\). What is the accumulation at \(x\), relative to its baseline?
- Can an accumulation remain at baseline over an interval even when the rate is only nonnegative there? Give the condition that allows this.
- A positive triangular region has area \(4\), followed by a negative triangular region of area magnitude \(3\). Has the accumulation returned to baseline at the end of the second triangle? What is its value?
- Explain why a signed accumulation can be zero while the accumulated magnitude is positive.
