AP Statistics · AP Physics 1 · AP Calculus AB
Months before exam day, Marlio estimates the band you are heading for, names the concepts moving it most, and schedules the work that shifts them. It will not show you a number until the evidence behind it is real.
Free to start. Your estimate and your gap map stay free.
Score estimate · Statistics
Early estimateTracking
3–5
most likely a 4 · 62% chance of 4+
Chance of each band
68% exam-day ready · 96 graded attempts
24 of 44 concepts assessed (63% by weight)
An estimate from your Marlio work — not affiliated with or endorsed by the College Board, and not based on College Board scoring. Sample student.
Caught in practice · Unit 1
3 matches“Small, medium and large go in order, so I can call them 1, 2 and 3 and work out the average rating.”
Resolution: order is real here, and it licenses the median, the mode and every “more than” comparison. It stops where an operation needs the gaps to be the same size. A mean needs that. Ask which summary survives a re-coding.
The concept tree
Sampling distribution of a sample mean
Open. Both prerequisites are mastered.
Real Marlio surfaces, rebuilt. The student in them is a sample, and the data is illustrative.
Counted from the shipped content when this page loaded. No user numbers, because a user count is not a reason to trust a forecast.
142
Concepts in the graph
Each with its prerequisites, its objective and its own items.
793
Misconceptions catalogued
184 of them carry the full diagnostic: the claim, what it predicts, the check, the resolution.
3
Subjects live
AP Statistics, AP Physics 1, AP Calculus AB.
19
Units mapped
475 practice and transfer items written against them.
What it does
The forecast
You get the score band you are heading for, the chance attached to each band, and the concepts ranked by leverage: weight against headroom. That is what turns a number into a place to start on a Tuesday night.
The refusal
Under the floor the dial withholds itself and says what it is waiting for, down to how many more concepts need evidence. A confident score off four questions is the exact failure this design exists to prevent.
Diagnosis
Distractors are tied to catalogued misconceptions, so a miss comes back as the belief in a student’s own words and the check that contradicts it.
Scheduling
FSRS-6 schedules every concept against your own recall curve, so revision week is a list of dates rather than a panic.
The tree
Every concept states what it is waiting on. Nothing opens because a week passed. It opens because the thing underneath it is held.
The forecast
A point score would be wrong about half the time, so Marlio shows the credible band, the distribution behind it, and the evidence it was computed from. The number is an instrument you can read, not a verdict handed down.
Score estimate · Statistics
Early estimateTracking
3–5
most likely a 4 · 62% chance of 4+
Chance of each band
68% exam-day ready · 96 graded attempts
24 of 44 concepts assessed (63% by weight)
An estimate from your Marlio work — not affiliated with or endorsed by the College Board, and not based on College Board scoring. Sample student.
The diagnosis
Every distractor is written against a specific wrong idea, so a miss is evidence about a belief rather than a dent in a score. Marlio replays the belief, states what it predicts, and shows the check that breaks it.
Caught in practice · Unit 1
3 matches“Small, medium and large go in order, so I can call them 1, 2 and 3 and work out the average rating.”
Resolution: order is real here, and it licenses the median, the mode and every “more than” comparison. It stops where an operation needs the gaps to be the same size. A mean needs that. Ask which summary survives a re-coding.
The schedule
Each concept carries a memory model. Marlio puts a review in front of you as recall drops toward the edge, which is the point where the work pays. A week looks uneven because forgetting is uneven.
Mon
12
due
Tue
4
due
Wed
0
clear
Thu
9
due
Fri
0
clear
Sat
15
due
Sun
2
due
Review dates come from FSRS-6, scheduled per concept against your own forgetting curve. Sample student.
The tree
A unit is drawn as its real prerequisite subgraph. A locked concept names the concept it is waiting on, and mastering that one opens it. There is no hidden progression and no unlock that a calendar handed you.
Variability in a statistic
Mastered
Sampling distribution of a sample mean
Open now
Confidence interval for a mean
Locked until the sampling distribution is mastered
Every edge is a real prerequisite between two concepts. Sample student.
Method
Marlio has not yet been run against a cohort of real exam results, so there is nothing here to quote and no student to put in a pull quote. Every number on this page counts our own content, and nothing on it claims an outcome.
What we can state is the mechanism. The estimate stays hidden until four conditions hold at once, and when any of them fails the dial says which one and how far off it is.
The commitment: when the first cohort sits their exams, we publish the forecast’s accuracy against their actual results, including the cases where it was wrong. Until that exists, this page will not pretend it does.
the graded attempts a concept needs before it counts as assessed at all.
of the course by weight has to be assessed before any band appears.
the ceiling on measurement error, in logits. Above it the estimate is too loose to be worth showing.
the oldest your median evidence may be. Stale work is not current knowledge.
Live
Marlio content graph
As of
Concepts in the prerequisite graph
142
Misconceptions catalogued
793
Of those, with a confront and a resolution written
184
Units mapped
19
Practice and transfer items
475
Subjects live
3
Counted from the shipped content on every request. Subjects authored privately for a single account are excluded, so this is what a visitor can actually reach.
Pricing
$89per year
or $11.99 a month · 7-day free trial
$199per year
or $23.99 a month
Free stays free and permanent: the score estimate, the gap map, 10 tutor messages a day, 20 practice questions a day, 3 graded free-response answers a month. There is also a $24 exam-season pass for one subject, April 1 to exam day. Full pricing.
Questions
A credible band for the score you are currently heading for, the chance attached to each band, and the concepts with the most leverage over it. Then lessons, practice, a tutor and a review schedule aimed at those concepts.
Because one of four conditions has not been met yet: 3 graded attempts per concept, 60% of the course covered by weight, measurement error at or under 0.50 logits, and median evidence no older than 30 days.
When it withholds, it tells you how many more concepts it needs. A number produced from too little work is worse than no number, because you would act on it.
AP Statistics, AP Physics 1, AP Calculus AB. Coverage is uneven and the product says so per concept rather than hiding it: where teaching content has not been written yet, the concept opens on an open-ended tutor prompt instead of pretending.
No. Marlio is not affiliated with, endorsed by, or sponsored by the College Board. Nothing in it is College Board material: no released items, no rubric wording, no scoring tables. Every concept, question and explanation is written for Marlio and aligned to publicly published topic outlines.
Wrong answers are written on purpose. Each distractor is tied to a catalogued misconception, so choosing it is evidence about a specific belief. Marlio replays that belief, states what it predicts, and shows the check that contradicts it.
No. The tutor works Socratically against the concept you are stuck on, and the evidence it produces feeds your mastery estimate. Free-response answers are graded against written criteria with the reasoning shown.
Nothing. The score estimate and the gap map are free permanently, with daily caps on tutoring and practice. Plus is $89 a year for one subject with a 7-day trial; Premium is $199 a year for every subject.
Take the diagnostic in one sitting. You get the band, the gap map and the first week of work, free.