When to put Desmos down: the fifteen-second rule and the questions it slows you on
The rule: if you cannot set a question up in Desmos cleanly in about fifteen seconds, stop, and do algebra or work backward from the choices. Everything below is why, and where it applies.
I wrote earlier about the places Desmos turns a two-minute problem into a ten-second one. This is the other half, and in a session it comes first, because a student who has just learned six tricks will try all six on the next question whether or not they fit.
The test has moved on this. More of the math questions on recent forms are written so that graphing does nothing for you. The difficulty is in the setup or the interpretation, and the calculator can only help once the setup is right. Bluebook’s spring 2026 update made the calculator easier to resize and see; it did not make it the right tool more often. Top scorers use it as a scalpel.
Can you use Desmos on every SAT math question?
You can open it on every question. It helps on roughly a third of them, and on the rest it costs time you will want later.
Desmos is two things: a shield against careless errors, and a speed tool for messy algebra. It is not a substitute for understanding, and it is not always the fastest path. The skill is not “using Desmos”. It is choosing between head, paper and calculator in the first five seconds of a question, and being right about the choice.
The fifteen-second rule
Say the number. It turns a vague feeling of this is taking a while into a rule you can check. On a section that gives you about ninety-five seconds a question, fifteen spent on a setup that is not working is a bad trade, and forty-five is a disaster you then have to make up somewhere else.
The companion rule: one sentence before you type. I am graphing both sides and reading the intersection. I am putting three points in a table and fitting a parabola. If the student cannot say the sentence, they do not have a method; they have a hope, and typing until something appears is gambling with time rather than calculating.
When should you not use Desmos?
When the choices are fractions or radicals, when the data is a frequency table, when the question is about meaning rather than value, and when you cannot say your method in one sentence. The full list:
The answer choices are ugly fractions. Desmos gives you 0.5833. The choices say 7/12. You will end up converting all four to test them. Do the algebra.
The answer choices contain radicals. Same problem. 1.414 against √2 is a conversion, and 2.236 against √5 is a slower one.
The data is a frequency table, a histogram or a dot plot. This is the big one, and it is how the SAT presents almost every median and spread question. Desmos’s statistics functions want a list of numbers; to use them you would have to type fourteen appears nine times as nine fourteens. Do the median by hand with a running count instead: total the frequencies, find the middle position, count up to it.
The question compares two standard deviations. The SAT never asks you to compute one. It asks which of two data sets has the larger spread, and it draws them so the answer is visible: the more clustered set has the smaller deviation. Reaching for a function here means you have misread the question.
A short, ordered list wants a median. Under ten values, cross off from both ends on paper. Faster than typing them in.
The question is about meaning, not value. What does the slope represent in this context? Which expression is equivalent? Graphing tells you nothing. Read.
The trap is in the setup. A word problem where the hard part is writing the right equation. Desmos will faithfully solve the wrong one.
Plain arithmetic. Slower than your head.
Six traps that turn a right method into a wrong answer
These are not judgment calls. They are mechanical errors I see repeatedly in the timing logs, each one a point lost on a question the student knew how to do.
- Degrees versus radians. The wrench icon sets the mode.
sin(30)is 0.5 in degrees and −0.988 in radians. Check it before any trigonometry. - Not zooming out. The most common Desmos mistake. If a graph seems to show no intersection or no zeros, the feature is very often just off-screen. Never conclude no solution from the default window; zoom out once first.
- Format mismatch. 1.75 is 7/4. Use the fraction toggle in the output box before comparing to the choices, or you will pick the wrong one of two that both “look right”.
- Solving for x when the question asked for something else. If 4x + 7 = 23, what is 8x + 3? Desmos hands you x = 4 with total confidence. The answer is 35. Read the last line of the question again before you answer it.
- Leftover lines. An expression from question 7 still on the graph corrupts question 8. Clear the calculator between questions until it is a reflex.
- Tangent or crossing? At default zoom a curve that touches a line and one that crosses it look identical. When the question turns on one solution versus two, zoom in on the point before deciding.
What this looks like in a log
When I look at a student’s per-question times on the math section, Desmos misuse has a signature: a cluster of slow correct answers on questions that had a fast algebraic route, next to fast wrong answers on the frequency-table statistics questions. The first group is time thrown away. The second is the calculator applied to the one category it is worst at.
Both are fixable in a session, and the fix is not more tricks. It is the rule at the top of this post, said out loud until it is automatic.