The complete map of the SAT Math section as nineteen recurring question types, ordered by their measured frequency in a 1675-question College Board bank: what each type looks like on screen, the one opening move that beats the others, and the wrong answer built for you. Covers nonlinear and linear functions, equations and systems, equivalent expressions, area and volume, ratios and percentages, one- and two-variable data, inequalities, right-triangle trigonometry, circles, conditional probability, margin of error and statistical claims, and closes with a ranked study order and a four-week plan.
Subject: SAT Prep · 62 slides · symbolic lesson
Open the interactive version of this deck
Title
SAT Math · The complete map
Nineteen types, ranked by how often they actually appear
Objectives
This deck is a map, not a textbook. It will not teach you to factor; it will teach you to recognise which of nineteen questions you are looking at, and what to do first.
The one idea underneath all of it: you cannot study everything, so study in frequency order.
Naruhodo Tutoring — SAT Math question bank export (sat-question-index.json) — the ordering in this deck is counted from the site's own tagged question bank
Section
Section 1
Concept
Before types, the container. The Math section of the digital SAT is two modules, and the second one adapts to how you did on the first.
| Module 1 | Module 2 | |
|---|---|---|
| Questions | 22 | 22 |
| Time | 35 minutes | 35 minutes |
| Difficulty | mixed, fixed for everyone | easier or harder, based on Module 1 |
None of that changes what is asked. The same nineteen types fill both modules.
College Board — Digital SAT Suite: test description and format — test format and timing
Picture it
College Board publishes how much of the section each of the four domains is worth. This is the blueprint every form is built to.
Figure (svg): Stacked bar of the four Math domains at their blueprint weights
Read the consequence: Algebra and Advanced Math together are about seventy per cent of your Math score. Geometry, which students fear most, is the smallest slice.
Picture it
Here is the same picture counted from the 1675 tagged questions in the bank you will actually practise on.
Figure (svg): Stacked bar of the four Math domains as counted in the local question bank
It is close, but not identical — the bank carries more Problem-Solving and Data Analysis than a real form does. Worth knowing so you do not over-weight statistics because the practice set does.
Prediction
A question about where attention goes, not where points are.
Predict first
Students routinely spend the largest share of their prep on one domain that is worth the least. Which one?
Correct: Geometry and Trigonometry
Why: Geometry is about 15% of the section — roughly 6 or 7 questions — but it is the domain students remember as hard from school, so it eats prep time out of proportion. Meanwhile Advanced Math, at 35%, gets skipped because it looks intimidating. Study the weighting, not the memory.
Picture it
This is the whole Math section on one slide. Every type you can be asked, sized by how often it appears in the bank.
Figure (svg): Horizontal bar chart of all nineteen Math question types ordered by frequency
The shape of that chart is the argument of this entire deck. The top seven bars are more than half the section; the bottom six together are less than one seventh of it.
Ranking
Before we start. Order these from most common to least common in the bank.
Put in order
Why: Nonlinear functions is the single most common type at 14.0% — one question in seven. Linear functions follows at 9.2%. Percentages is 4.5%, and Circles is 3.0%. Nonlinear functions alone outweighs Percentages and Circles put together, nearly twice over.
Concept
About a quarter of the section is student-produced response: no choices, you type the answer into a box. The maths is not harder. The formatting rules are where points leak.
Read that list twice. A correct answer typed as 1,500 scores zero, and no amount of algebra recovers it.
College Board — Digital SAT Suite: test description and format — student-produced response entry rules
Concept
Every geometry formula you are expected to use is provided in the app. Memorising them is wasted effort; knowing which ones are missing is not.
| given to you | you must know it yourself |
|---|---|
| Area and circumference of a circle | Slope from two points |
| Area of a triangle and a rectangle | The quadratic formula |
| The Pythagorean theorem | Vertex form and what h and k mean |
| Special right triangles, 30-60-90 and 45-45-90 | SOH-CAH-TOA |
| Volume of box, cylinder, sphere, cone, pyramid | Percent change as a multiplier |
| 360 degrees and 2 pi radians in a circle | Exponent and radical rules |
Everything in the right-hand column is algebra, which is exactly the domain worth seventy per cent. That is not a coincidence.
College Board — Digital SAT Suite Assessment Specifications, Math section: domain weightings and skill definitions — reference sheet contents
Two truths and a lie
The built-in graphing calculator wins points on some types and quietly wastes minutes on others. Three of these claims are true. Eliminate them, and the one left standing is the false one.
Eliminate the wrong options
Which claim about Desmos is FALSE?
Survives elimination: c
Why: Equivalent expressions is the one type where Desmos is the slow road. The fastest route is to substitute a small number into the original and into all four choices, which is arithmetic you can do in your head. Typing five expressions into a calculator takes longer and introduces transcription errors. The rule of thumb: reach for Desmos when the question is about a picture — an intersection, a zero, a region — and leave it closed when the question is about a rewrite.
Pattern
This is the engine of the deck. For all nineteen types, you are learning exactly three things, in this order.
Nineteen types times three facts is fifty-seven things to know. That is a week of work, and it is worth more than a month of untargeted practice.
Every slide from here on is one card with those three panels: teal tell, green move, red trap.
Warm-up
A warm-up you cannot yet do well. Come back to this slide at the end of the deck.
Discussion prompt
A stem reads: "The function f is defined by f(x) = 2(3)^x. What is the value of f(0)?" Which of the nineteen types is this, and what is the first thing you write?
Hint: Ask what shape the function is before you ask what is being computed.
Answer:
Type 1, Nonlinear functions — the tell is the variable sitting in the exponent, which makes it exponential rather than linear.
The first move is to recognise the form: in a times b to the x, the number a is the starting value, which is what f(0) asks for. The answer is 2, with no arithmetic at all.
The trap is computing 2 times 3 and answering 6, which is f(1), not f(0). Anything to the power zero is one.
Section
Section 2 — 57.9% of the section
Concept
The top seven types are 57.9 per cent of the bank between them — call it 25 of the 44 questions you will see.
| # | type | share | running total |
|---|---|---|---|
| 1 | Nonlinear functions | 14.0% | 14.0% |
| 2 | Linear functions | 9.2% | 23.2% |
| 3 | Nonlinear equations and systems | 8.7% | 31.9% |
| 4 | Linear equations in two variables | 7.3% | 39.2% |
| 5 | Systems of two linear equations | 6.6% | 45.8% |
| 6 | Equivalent expressions | 6.1% | 51.9% |
| 7 | Linear equations in one variable | 6.0% | 57.9% |
Notice what they have in common: every one of the seven is Algebra or Advanced Math. Geometry does not appear until type 8.
Picture it
The most common question on the test, by a clear margin: a function that is not a straight line. Quadratics, exponentials, polynomials and radicals all live here, and so does every question about a curved graph.
Figure (svg): Nonlinear functions: the tell, the move, and the trap
The single highest-value hour in your prep: learn to read the three forms of a quadratic and say what each one hands you for free. That hour pays for one question in seven.
Picture it
A constant rate of change, dressed as a story. Almost always a word problem, and almost always solvable by naming two numbers.
Figure (svg): Linear functions: the tell, the move, and the trap
If you can answer what is true at zero and what happens each step, you have solved every linear-function question on the test.
Picture it
Solving, rather than interpreting, when a curve is involved. Includes systems where one equation is a line and the other is not, and every question that asks how many solutions exist.
Figure (svg): Nonlinear equations and nonlinear systems: the tell, the move, and the trap
The discriminant is worth memorising cold: positive means two solutions, zero means one, negative means none. It turns a hard question into a ten-second one.
Picture it
The line itself, rather than a story about it. Two points, a slope and a point, a table, or a graph — and the question wants the equation, or something you can read off it.
Figure (svg): Linear equations in two variables: the tell, the move, and the trap
Check the axis scale before you read anything off a graph. This is the most reliable way the test converts a student who knows the maths into a student who got it wrong.
Picture it
Two unknowns, two conditions. The most predictable type on the test, and the one Desmos handles best.
Figure (svg): Systems of two linear equations: the tell, the move, and the trap
The solution-count rule, in one line: same slope and different intercept means no solution; same slope and same intercept means infinitely many; different slopes means exactly one.
Picture it
No solving at all — just rewriting. Factoring, expanding, exponent rules, and combining rational expressions.
Figure (svg): Equivalent expressions: the tell, the move, and the trap
Substitution beats manipulation here. If you take one tactic from this deck into the test, take this one — it converts an algebra question into arithmetic.
Picture it
The simplest type on the test, and it still appears in six per cent of questions. One unknown, one equals sign, straight lines only.
Figure (svg): Linear equations in one variable: the tell, the move, and the trap
Watch for the question that solves for x and then asks for 2x plus 1. The value of x is always one of the wrong answers, deliberately.
Check
One question, and the point of it is the trap, not the algebra.
Check your understanding
The function g is defined by g(x) = (x minus 3)(x plus 5). What is the x-coordinate of the vertex of the graph of g?
Answer: A
Why: The factored form hands you the zeros for free: x equals 3 and x equals negative 5. A parabola is symmetric, so its vertex sits exactly halfway between its zeros. Halfway between 3 and negative 5 is negative 1. No expanding, no formula, about eight seconds.
This is Type 1 answered with the Type 1 move: the form you were given already had the answer in it.
Trap
The trap. A question ends: ...what is the value of 3x?
You solve carefully, get x equal to 4, see 4 among the choices, and pick it. The algebra was perfect. The answer is wrong.
This is not carelessness in the ordinary sense. The test places the value of x among the choices on purpose, on hundreds of questions, precisely because solving feels like finishing.
The fix, and it costs two seconds: before you start solving, underline or say aloud the thing being asked.
On a 44-question section, students typically lose two to four points to this single habit. It is the cheapest fix available to you.
Discrimination
No solving. Read each stem and name the type — this is the ten-second skill the whole deck is built around.
Sort into buckets
Which type is each stem?
Error analysis
A student's work on an equivalent-expressions question. One line is illegal. Find it before reading the notes.
Annotate
On: \( \frac{x^2 + 6x}{x} \;=\; x + 6x \;=\; 7x \)
Every type has one signature error. For Equivalent expressions it is partial cancelling, and it survives into calculus if nobody names it.
Section
Section 3 — 27.9% of the section
Concept
Types 8 to 13 are 27.9 per cent of the bank — roughly 12 of your 44 questions. Together with tier one that is 85.8 per cent, or about 38 questions.
| # | type | share | running total |
|---|---|---|---|
| 8 | Area and volume | 5.3% | 63.2% |
| 9 | Ratios, rates, proportions, and units | 4.9% | 68.1% |
| 10 | Lines, angles, and triangles | 4.7% | 72.8% |
| 11 | Percentages | 4.5% | 77.3% |
| 12 | One-variable data: centre and spread | 4.3% | 81.6% |
| 13 | Linear inequalities in one or two variables | 4.2% | 85.8% |
This is where the other three domains finally appear. If tier one gets you to roughly 600, tier two is the difference between 600 and 700.
Picture it
Shapes and solids. The formulas are all handed to you, so this type is never really about recall — it is about units and scaling.
Figure (svg): Area and volume: the tell, the move, and the trap
Say the scaling rule until it is automatic: lengths scale by k, areas by k squared, volumes by k cubed. One sentence, worth a question on most forms.
Picture it
Proportional reasoning and unit conversion. Mechanically easy, and the errors are almost always in the setup rather than the arithmetic.
Figure (svg): Ratios, rates, proportions, and units: the tell, the move, and the trap
Units are a free correctness check. If you cancel them and end up with miles per hour when the question wanted hours, stop — you have not made an arithmetic slip, you have set it up upside down.
Picture it
Angle chasing and similar triangles. The geometry that appears most often, and it rewards marking up the figure over reasoning in your head.
Figure (svg): Lines, angles, and triangles: the tell, the move, and the trap
Similar triangles are the high-frequency half of this type. The tell is two triangles sharing an angle, and the move is always the same: corresponding sides, in the same order, as a ratio.
Picture it
Percent of, percent change, discount, tax, interest. Ordinary arithmetic that becomes reliable the moment you stop computing percentages additively.
Figure (svg): Percentages: the tell, the move, and the trap
That 0.96 is the whole type in one number: a rise and an equal fall leave you four per cent down, every time, regardless of the starting amount.
Picture it
Mean, median, range, and standard deviation, read off a list, a table, a histogram or a box plot. Interpretive rather than computational.
Figure (svg): One-variable data: centre and spread: the tell, the move, and the trap
The one sentence that answers most of this type: the mean chases outliers, the median ignores them. Standard deviation is about width, never about centre.
Picture it
The same algebra as equations, with one extra rule and one extra way to be asked. Regions in the plane belong here too.
Figure (svg): Linear inequalities in one or two variables: the tell, the move, and the trap
Testing (0, 0) answers nearly every shaded-region question in one substitution. Use another point only when the boundary passes through the origin.
Check
A percentages question built the way the test builds them.
Check your understanding
A jacket is marked up 25 per cent from its cost, then put on sale at 20 per cent off the marked price. The sale price is what percent of the original cost?
Answer: A
Why: Use multipliers and let the cost be 1. Up 25 per cent is times 1.25; down 20 per cent is times 0.80. Together, 1.25 times 0.80 equals exactly 1.00, so the sale price is 100 per cent of the cost — the shop breaks even. Choosing a friendly number like 100 dollars gives the same result: 125, then 100.
This is the rare case where the multipliers cancel exactly. That is precisely why it is asked — the intuitive answer, plus 5 per cent, is wrong and feels right.
Trap
The trap. A triangle is drawn on screen. One angle looks like a right angle. You use the Pythagorean theorem, and every step after that is clean.
Underneath the figure, in small grey type, are the words not drawn to scale. There was no right angle. The theorem never applied.
The same trap runs on lengths: a side that looks twice another is not twice another, and a point that looks like a midpoint is not a midpoint.
The fix. Treat a figure as a diagram of relationships, not a measurement, unless you are told otherwise.
Note the symmetry: the same caption that forbids you from measuring also licenses estimation when it is absent. Both directions are worth points.
Comparison
Fill the blanks from memory. These four turn up across types 11, 12 and 14, and mixing them up is a reliable way to lose points.
Comparison matrix
| term | what it measures | does an outlier move it? |
|---|---|---|
| Mean | the balance point of the data | yes, strongly |
| Median | the middle value in order | barely |
| Range | largest minus smallest | yes, by definition |
| Standard deviation | typical distance from the mean | yes, it widens |
The median is the only one of the four that an outlier leaves essentially alone. That single fact answers most what-happens-to-the-data questions.
Section
Section 4 — 14.2% of the section
Concept
The last six types are 14.2 per cent of the bank — about 6 questions. Individually each is rare; together they are the gap between a strong score and a top one.
| # | type | share | running total |
|---|---|---|---|
| 14 | Two-variable data: models and scatterplots | 3.8% | 89.6% |
| 15 | Right triangles and trigonometry | 3.0% | 92.6% |
| 16 | Circles | 3.0% | 95.6% |
| 17 | Probability and conditional probability | 2.5% | 98.1% |
| 18 | Sample statistics and margin of error | 1.3% | 99.4% |
| 19 | Evaluating statistical claims | 0.7% | 100% |
Here is the useful property of the tail: these types have narrow, memorisable rules. Types 18 and 19 in particular are almost pure vocabulary — an hour spent there is close to guaranteed points.
Picture it
A scatterplot with a fitted line or curve, and questions about what the model says. Interpretation, not calculation.
Figure (svg): Two-variable data: models and scatterplots: the tell, the move, and the trap
Whenever a choice says one thing causes another, look hard. Scatterplot data almost never licenses that word, and the test knows it is tempting.
Picture it
Right triangles, the three ratios, and the special triangles. A small, closed, memorisable topic that repays study out of proportion to its size.
Figure (svg): Right triangles and trigonometry: the tell, the move, and the trap
One relationship gets tested directly and surprises people: sin of an angle equals cos of its complement. If sin x equals 0.6, then cos of (90 minus x) is also 0.6.
Picture it
Circle equations, arcs, sectors and radians. Two quite different sub-topics wearing one label.
Figure (svg): Circles: the tell, the move, and the trap
Both halves reduce to one habit: get to standard form, then read. Everything a circle question asks is visible once the equation is in that shape.
Picture it
Almost always a two-way table. The arithmetic is trivial; identifying the correct denominator is the entire question.
Figure (svg): Probability and conditional probability: the tell, the move, and the trap
Underline the words after given that and circle the row or column they name. The question becomes two numbers and a division.
Picture it
Surveys, confidence intervals and margins of error. Rare, narrow, and almost entirely a vocabulary test.
Figure (svg): Sample statistics and margin of error: the tell, the move, and the trap
Two rules cover nearly every question here: bigger sample, narrower interval, and conclusions stop at the edge of the population you sampled.
Picture it
The rarest type on the test, at 11 questions in 1675 — and the one with the highest ratio of points to study time, because it is a single distinction.
Figure (svg): Evaluating statistical claims: the tell, the move, and the trap
Learn this pair and you own the type: random assignment licenses cause; random selection licenses generalisation. Neither one gives you the other.
Check
A conditional probability question with the standard trap in place.
| Passed | Failed | Total | |
|---|---|---|---|
| Studied | 42 | 8 | 50 |
| Did not study | 15 | 35 | 50 |
| Total | 57 | 43 | 100 |
Check your understanding
A student is chosen at random from those who passed. What is the probability that the student studied?
Answer: A
Why: Chosen from those who passed restricts the population to the Passed column, whose total is 57. Of those 57, the number who studied is 42. So the probability is 42 out of 57.
Notice that choice C is the same two facts in the opposite order. Conditional probability is not symmetric, and the test builds a choice for the reversal every time.
Trap
The trap. A scatterplot shows that towns with more libraries have higher reading scores. A choice reads: building more libraries would raise reading scores.
It matches the data, it sounds sensible, and it is the answer most students pick. It is wrong, and it is wrong for a reason the test tests deliberately.
Observational data cannot separate the libraries from everything that travels with them — funding, income, parental education. Any of those could produce the same plot.
The fix. Treat causal language as a red flag word, the way you treat always and never.
This single habit covers type 14, type 18 and type 19 — about 5.8 per cent of the section, from one reading rule.
Matching
The whole deck compressed. Six types, six opening moves — no solving, just the first line you write.
Match the pairs
Why: Every one of these openings replaces a page of algebra with a single decision. That is the entire claim of this deck: the first line you write matters more than how well you execute the rest.
Elimination
A rectangle has length 3 metres and width 2 metres. Every length is then tripled. What happens to the area?
Eliminate the wrong options
Which choice survives, and can you rule the others out before computing anything?
Survives elimination: b
Why: Lengths scale by 3, so area scales by 3 squared, which is 9. Checking: the original area is 6 square metres and the new one is 9 by 6, which is 54 — and 54 divided by 6 is 9. The three wrong answers are the three predictable confusions: using the factor, using its cube, and multiplying the sides.
Section
Section 5 — the plan
Concept
This is the slide to photograph. It is the whole nineteen-type list in study order, with the running total of the section you have covered by the time you finish each one.
| order | study this | domain | share | you now cover |
|---|---|---|---|---|
| 1 | Nonlinear functions | Advanced Math | 14.0% | 14.0% |
| 2 | Linear functions | Algebra | 9.2% | 23.2% |
| 3 | Nonlinear equations and systems | Advanced Math | 8.7% | 31.9% |
| 4 | Linear equations in two variables | Algebra | 7.3% | 39.2% |
| 5 | Systems of two linear equations | Algebra | 6.6% | 45.8% |
| 6 | Equivalent expressions | Advanced Math | 6.1% | 51.9% |
| 7 | Linear equations in one variable | Algebra | 6.0% | 57.9% |
| 8 | Area and volume | Geometry | 5.3% | 63.2% |
| 9 | Ratios, rates, proportions, units | Problem-Solving | 4.9% | 68.1% |
| 10 | Lines, angles, and triangles | Geometry | 4.7% | 72.8% |
| 11 | Percentages | Problem-Solving | 4.5% | 77.3% |
| 12 | One-variable data: centre and spread | Problem-Solving | 4.3% | 81.6% |
| 13 | Linear inequalities | Algebra | 4.2% | 85.8% |
| 14 | Two-variable data and scatterplots | Problem-Solving | 3.8% | 89.6% |
| 15 | Right triangles and trigonometry | Geometry | 3.0% | 92.6% |
| 16 | Circles | Geometry | 3.0% | 95.6% |
| 17 | Probability and conditional probability | Problem-Solving | 2.5% | 98.1% |
| 18 | Sample statistics and margin of error | Problem-Solving | 1.3% | 99.4% |
| 19 | Evaluating statistical claims | Problem-Solving | 0.7% | 100% |
Work down it. Do not skip ahead to the topic you find interesting, and do not start at 19 because it is short.
Naruhodo Tutoring — SAT Math question bank export (sat-question-index.json) — shares counted from 1675 tagged questions
Picture it
The same nineteen types grouped into the three tiers, sized by what they are worth.
Figure (svg): Stacked bar showing tier 1 at 57.9 per cent, tier 2 at 27.9 per cent and tier 3 at 14.2 per cent
Read it as a schedule: tier one is most of your score, tier two is most of the rest, tier three is the polish. Spend your weeks in that ratio and nothing else.
Concept
Four weeks, about five hours a week. The plan is built on the ranking, not on a textbook's chapter order.
| week | types | what you do |
|---|---|---|
| 1 | 1-3 | Three forms of a quadratic; a times b to the x; the discriminant. Drill 30 questions per type until the tell is instant. |
| 2 | 4-7 | Lines, systems, and substitution-instead-of-algebra. Add a timed 22-question module at the end of the week. |
| 3 | 8-13 | Reference sheet, scaling rule, multipliers, mean-versus-median, the sign flip. Mixed drill, no type labels. |
| 4 | 14-19 | The tail, plus a full timed section. Two hours on types 17-19 alone, since they are almost pure vocabulary. |
Khan Academy — Official Digital SAT Prep, Math — free official practice organised by these same skill tags
Sorting
Three students describe how they lost a point. Sort each into which of the three facts failed them — this is exactly how your error log should work.
Sort into buckets
Tell, move, or trap?
Trade off
Not all study hours are worth the same. Fill in what one focused hour returns.
Comparison matrix
| one hour spent on | questions it can affect | how learnable it is |
|---|---|---|
| Types 18 and 19 (stats vocabulary) | about 1 question | very high — two rules, closed topic |
| Type 1 (nonlinear functions) | about 6 questions | moderate — three forms to internalise |
| Grid-in formatting rules | up to 11 questions | very high — it is a list to read once |
| Memorising the reference sheet | 0 questions | irrelevant — it is given to you |
The bottom row is the point. Effort is not the same as return. Grid-in formatting takes ten minutes and protects a quarter of the section; memorising provided formulas protects nothing.
Concept
The 1675 questions behind this deck are tagged with the same nineteen skills, so you can practise in exactly the order this deck recommends.
Practising a type you already own feels productive and teaches nothing. The error log tells you where to go; follow it rather than your preferences.
Naruhodo Tutoring — SAT Math question bank export (sat-question-index.json) — 1675 questions, tagged by domain, skill and difficulty
Explain it to yourself
Close the deck for this one. Retrieval, not recognition.
Discussion prompt
Without looking, name the top five Math question types in order, and give the tell for each.
Hint: The first three are worth 31.9 per cent between them. Two of the five are about lines.
Answer:
1. Nonlinear functions (14.0%) — a squared, cubed, rooted or exponential x; a curve on screen.
2. Linear functions (9.2%) — a per, a rate, a starting amount, a constant-step table.
3. Nonlinear equations and systems (8.7%) — solve with a square or a root, or how many solutions.
4. Linear equations in two variables (7.3%) — two points, a slope, parallel or perpendicular.
5. Systems of two linear equations (6.6%) — two equations at once, or no solution / infinitely many.
If you got the first three, you have named 31.9 per cent of the section from memory.
Explain it
Two minutes, out loud, to a friend who is about to start studying.
Discussion prompt
A friend says: I am going to start with geometry because it is my weakest area. What do you tell them, and what do you tell them to do instead?
Answer:
Weakest is the wrong axis. The right question is weakest times most common, because a topic you are bad at that is worth 3 per cent costs you less than a topic you are mediocre at that is worth 14 per cent.
Geometry and Trigonometry is about 15 per cent of the section — around 6 or 7 questions. Advanced Math is about 35 per cent.
What to do instead: start at type 1 and work down the ranking, but spend longer on the ranked types you are weak at. The order comes from the test; the pace comes from you.
The one exception worth granting: if a whole domain is genuinely blank for you, a short foundation pass first is reasonable. But it is a detour, not the plan.
Commit first
Commit to a number before you check.
Predict first
If you can reliably name the type and the first move for types 1 through 7, what share of the Math section have you covered?
Correct: About 58 per cent
Why: Types 1 to 7 sum to 57.9 per cent of the bank — roughly 25 of 44 questions. That is what seven types buys you, and it is why the ranking is the plan. Adding types 8 to 13 takes you to 85.8 per cent.
Exit ticket
One question. Answer it before you close the deck.
Predict first
You have three hours of study left before the test and no idea where to spend them. What do you do first?
Correct: Re-read the error log and drill the types it names
Why: With three hours left, new practice generates data you no longer have time to act on. The error log is data you already paid for. Reviewing the reference sheet is worth nothing, since it is given to you during the test. And the tail types are only worth prioritising if the log says so — shortness is not a reason.
Recap
Nineteen types. Three facts each. One ranking that decides your order.
| never do this | do this instead |
|---|---|
| Start with your weakest topic | Start at the top of the frequency list, slow down where you are weak |
| Memorise the reference sheet | Memorise the algebra it does not give you |
| Expand everything into standard form | Ask which form already answers the question |
| Do algebra on equivalent expressions | Substitute a small number into all five expressions |
| Add and subtract percentages | Multiply multipliers |
| Trust the figure | Check for not drawn to scale, then use only what is marked |
| Answer x | Re-read the underline: it may have wanted 2x |
Naruhodo Tutoring — SAT Math question bank export (sat-question-index.json) — and the site's SAT Math page to drill these nineteen types in order
Want this taught 1-on-1? Alexander tutors SAT Prep — $55/session, free consultation.