You know the math.
This finds out where you lose points anyway.

The Execution Diagnostic is an instrument I built myself, because no commercial SAT diagnostic distinguishes knowing the math from executing it under test conditions. Bluebook, Kaplan, Princeton Review, and Khan Academy all hand you a score and a topic breakdown. This hands you the reasons.

Book the diagnostic See a sample report

SAT math, special right trianglesiPad, Zoom session · Aug 2026

Why a practice test can’t tell you this.

A practice test is scored on whether the answer was right. It cannot see that the student solved the problem and then reported the wrong number, or answered in eight seconds because the question looked like one they’d seen, or ran out of module before the questions they could have answered. Those three things cost more points than any topic gap for most students above 600, and a score report files all of them under “Algebra.”

Past a certain point the SAT is not a test of how well you can do the math. It is a test of how well you understand the way the problems are written. I built an instrument to measure that directly.

About an hour, in one sitting, with me watching the whole thing.

What they don’t know

Some points leave because a topic was never learned properly. That part is straightforward, and teaching it is the easy half of the job. Most diagnostics stop here, which is why most plans are a list of topics.

What they know and lose anyway

The rest leave for reasons that have nothing to do with the math. The question that was misread. The answer that was right until it was written down. The twenty seconds that should have been ninety, and the four minutes that should have been one. A practice test cannot see any of this, because it only ever records whether the answer was right.

Which is costing more, and what to fix first

Both at once, weighed against each other and put in order. That order is the plan, and it is the reason the first thing I work on is often not the thing anyone expected.

I sit with the student for the whole hour rather than scoring it afterwards, because half of what I need to know is visible only while it is happening. The items, the way answers are coded, and the scoring behind the report are mine, and they are not published.

Right or wrong is the least of it.

A practice test records one thing per question. This records several more, and that is the only reason it can tell a misconception from a gap, or a rushed guess from a genuine attempt. Two students can miss the same question for opposite reasons and need opposite work; a score report files both under “Algebra.”

The rest of it is me. I watch the student take it, write down what I see, and then put my notes next to the numbers. Neither one is enough on its own — the numbers tell me what happened, and watching tells me why.

What came back for one student, and what I did about it.

His August report, name removed, and the four findings that decided the plan. Reviewed live with the family for up to two hours, then delivered in writing within five business days. He had a 390 in math and no idea why.

Finding one

He could do the hard questions.
He just never got to them.

On the second module he answered the easy and medium questions well and got nothing at all on the hard ones — not because they were beyond him, but because the module ended first. His ceiling was set in the opening minutes, by pacing, not by knowledge.

So the score was never a measure of what he knew. It was a measure of how far into the test he got.

Accuracy by item difficultyModule 1 in rust, module 2 in green

Finding two

Four minutes on question one, then a sprint, then fourteen minutes left on the table.

He spent a fifth of module one on its first question, panicked through the back half at eight to seventeen seconds each, and still handed it in with 14:17 unused. In the other module he did the reverse and ran right to the buzzer. Nobody had ever shown him the clock, so the clock was running him.

So the first thing I changed was not algebra. It was when he is allowed to give up on a question and move on.

Time spent, question by questionAgainst the 35-minute module limit

Finding three

Twelve questions got under twenty seconds.
Ten of them were wrong.

That is not a knowledge problem and it is not bad luck; it is a habit, and habits are what a plan can actually be written against. The write-in answers were worse still: one right out of eleven, and seven of the misses were a number already printed in the problem. He was reading the question, finding a number, and writing it down.

So he got two rules before he got a single new topic: nothing goes down under twenty seconds, and on a write-in he writes what he is solving for before he starts solving.

Every question, by seconds spentGreen correct, rust incorrect

Multiple choice against write-in answersEach dot is one question

Finding four

Five of the twenty-seven misses were a gap in knowledge.
Twenty-two were something else.

The largest group was the one that stings: he did the mathematics correctly and then reported the wrong number. Below are two of them, reconstructed. In the first he found the angle at the center and doubled it when the question wanted it halved. In the second he squared one side of the equation and not the other. Both students who make these errors know the rule perfectly well.

So five of them became lessons, and the other twenty-two became a checklist he runs on every question. That is the difference between a tutor teaching topics and a plan built from evidence.

The 27 misses, by what actually caused themGrouped by reason, not by topic

The relationship, reversedHe doubled where the question halved.

One side squared, not the otherA rule he could recite, applied unevenly.

One report, three numbers, and a plan that isn’t “take more tests.”

5 of 27misses that were actually a gap in knowledge. Those get taught.

10 of 12questions answered in under twenty seconds, wrong. That gets a rule, not a lesson.

7answers that were the number printed on the page. That gets a habit, drilled until it holds on test day.

The plan that came out of this report was very specific and very action-oriented: what I hit first, in what order, with a full-length proctored retest halfway through to check that things are going the way I think they are.

Which one your student takes.

The Execution Diagnostic

Math 650 and above. The instrument described on this page. It answers: I know the math, why am I losing points anyway?

Book →

The Foundation Diagnostic

Math between about 480 and 600. Same length, different question: where should the study time go to reach 650? It maps content gaps across eight topic families and, separately, execution drag on the topics the student already owns. When the content comes back clean, the student moves up to the Execution Diagnostic.

Book →

The free math placement

Below 480, or for anyone before we’ve spoken. Adaptive, sixth grade through Precalculus, results on screen, no sign-up.

Take it now →

The diagnostic is $199, flat, and the whole fee is credited toward any package. It is the first step in every SAT engagement, and there is nothing to commit to beyond it.

Book the diagnostic.

Tell me the student’s current score if you have one and the test date you’re aiming at. I reply within a day. (706) 389-0528 if a call is easier.

You know the math.This finds out where you lose points anyway.