Data study

Not all SAT skills are worth the same. Here is which ones actually move your score.

We ranked all 24 tested skills by how often they appear and how hard they run, using our full tagged bank of real questions. The skills you see most are not the hardest, the hardest are not the most common, and most study plans spread time evenly across all of them. Here is a better order.

Perfect1600 Research·August 2026·Measured from 3,292 tagged questions in the official bank
3,292
real SAT questions we tagged by skill and difficulty
24
tested skills, and they are not equally worth your time
5
skills that are both common and highly coachable: start there

Most study plans treat the SAT as 24 equal boxes to fill: a bit of this skill, a bit of that one, work through a book front to back. That feels fair, and it is close to the worst way to spend limited hours, because the skills are wildly unequal on the two things that decide your return: how often a skill shows up, and how hard its questions run.

We measured both. Our tagged copy of the official question bank holds 3,292 real College Board questions, each labeled by skill and by College Board’s own Easy, Medium, or Hard rating. That lets us ask two plain questions of the actual test, not the marketing blueprint: which skills appear the most, and which ones are genuinely hard.

1. The skills you will see the most

A handful of skills dominate by sheer volume. In Reading and Writing, Words in Context leads, followed closely by the two grammar clusters and transitions. In Math, Linear Functions is the single most common skill on the section.

The most common skills in the bank
Count of real questions per skill (n = 3,292). Reading & Writing in one color, Math in the other.
Reading & Writing Math
Words in Context
259
Linear Functions
241
Nonlinear Graphs
211
Boundaries
211
Form & Sense
206
Rhetorical Synth.
204
Transitions
194
Text Structure
149
0%questions in the bank100%

2. The skills that are genuinely hard

Now the other axis. Rank skills by the share of their questions that College Board marks Hard, and a completely different set rises to the top. Two geometry skills lead the whole test, and a cluster of Reading skills follows.

The hardest skills, by share of questions marked Hard
Percent of each skill's real questions rated Hard by College Board.
Circles
60%
Right-triangle trig
59%
Inferences
51%
Command of Evidence
49%
Boundaries
45%
Percentages
45%
Cross-Text
44%
Nonlinear Graphs
39%
0%share marked Hard100%

Notice how little the two lists overlap. The skills you meet most often (Words in Context, Transitions, Linear Functions) are not especially hard. The hardest skills (Circles and right-triangle trigonometry, both around 60% Hard) are among the rarest on the test, at roughly 4% of the Math section each. Boundaries is the notable exception that appears on both lists, and that is exactly what makes it valuable.

This is the whole point. “Study the hardest thing” and “study the most common thing” are different advice that point at different skills. Neither is right on its own. What you actually want is the overlap: skills that show up a lot and that you can genuinely improve at in the hours you have.

3. The third axis nobody plots: coachability

Frequency and difficulty are measured. The third factor is a judgment, and we will label it as one: how coachable a skill is, meaning how much it reduces to a short, repeatable procedure that works regardless of the passage or the numbers.

Boundaries is the clearest case. It runs 45% Hard, which sounds intimidating, but nearly all of that difficulty comes from a small, fixed set of punctuation decisions. Learn the decision (when a period, a semicolon, a colon, or a comma is correct) and a hard Boundaries question collapses into an easy one. The same is true for Transitions (a question about the direction of a logical relationship), Words in Context (predict your own word, then match), and the grammar of Form, Structure and Sense (a finite list of rules). These are the skills where a method buys the most points per hour.

Contrast that with Circles or right-triangle trig. A method helps, but each hard question still demands genuine multi-step reasoning on a figure you have not seen before. You improve, just more slowly per hour. That is fine when you are chasing the last few points at the top of the scale. It is the wrong place to start when you have a plateau to break.

SkillHow oftenHard %CoachableWhen
BoundariesHigh (11% of R&W)45%VeryFirst
Words in ContextHighest R&W (14%)23%VeryFirst
Linear FunctionsHighest Math (17%)21%HighFirst
TransitionsHigh (11%)21%VeryFirst
Form, Structure & SenseHigh (11%)30%VeryFirst
Rhetorical SynthesisHigh (11%)21%HighSoon
Nonlinear GraphsHigh Math (15%)39%MediumSoon
Command of EvidenceMedium (8%)49%MediumSoon
PercentagesLow Math (4%)45%HighIf it's a weak spot
CirclesLow (4%)60%LowLast (ceiling)
Right-triangle trigLow (4%)59%LowLast (ceiling)

Dot color marks the tier: start here (primary), get to it soon (brass), save for last (grey). Frequency and Hard % are measured from the bank; Coachable and When are our recommendation.

What to do with the next ten hours

If you have a stack of hours and no plan, spend the first block on the five skills that are common and highly coachable: Boundaries, Words in Context, Transitions, Form and Structure and Sense, and Linear Functions. They appear on nearly every test, they reward a fixed method, and several of them run Hard, which means the points are sitting there for anyone who actually learns the rule.

Then move to the “soon” tier, where the method matters more and the reasoning is a little deeper: Rhetorical Synthesis, Nonlinear Graphs, Command of Evidence. Leave the rare, hard geometry for the end, when you are optimizing rather than plateaued. This is not about avoiding hard material. It is about not paying top price for your first points.

Practice by skill, in priority order

Perfect1600 filters practice to any of the 24 skills and shows you the method on every question, so you can spend your hours where they pay. Free to start.

Already have an account? Pick a skill and drill it

How we counted

Source: our tagged copy of the official Digital SAT question bank, 3,292 questions sourced from College Board and labeled by skill and by College Board’s own difficulty rating. “How often” is a skill’s share of the real questions in its section (Reading & Writing has 1,838 of the 3,292; Math has 1,454). “Hard %” is the share of a skill’s questions marked Hard. Frequency and difficulty are measured; the bank mirrors the real section mix closely but is not the exact blueprint, so treat the counts as directional and the difficulty splits as robust. Coachability and the recommended study order are Perfect1600’s judgment, informed by which skills reduce to a fixed method in our strategy library.

Common questions

Where do these frequency and difficulty numbers come from?

From our own tagged copy of the official Digital SAT question bank: 3,292 real College Board questions, each labeled by skill and by College Board's own Easy/Medium/Hard rating. 'How often' is a skill's share of the questions in its section (Reading & Writing or Math). 'How hard' is the share of that skill's questions marked Hard. Both are measured, not estimated from the published blueprint.

Isn't the hardest skill the one I should study most?

Not if it barely appears. Circles and right-triangle trigonometry are the two hardest skills in the bank, but each is only about 4% of the Math section, so a huge effort there buys you one or two questions. Difficulty tells you where points are hard to get; frequency tells you how many points are on the table. You want skills that score well on both, and that are coachable enough to actually improve in the time you have.

What does 'coachable' mean here, and is it measured?

Coachable is our judgment, not a bank statistic, and we label it that way. A skill is coachable when it reduces to a short, repeatable procedure that works no matter the passage or the numbers: the punctuation decision for Boundaries, predict-then-match for Words in Context, reading slope and intercept for Linear Functions. Skills that need genuine multi-step reasoning on unfamiliar figures, like Circles, improve more slowly per hour even with a good method.

So where should a plateaued student actually start?

With the skills that are common and highly coachable: Boundaries, Words in Context, Transitions, Form/Structure/Sense, and Linear Functions. They appear constantly, they reward a fixed method, and several of them run surprisingly Hard, which means the points are there for people who learn the rule. Save Circles and right-triangle trig for when you are chasing the last few points at the top.