IB Maths AI SL November 2023 TZ2: what the examiners reported

Paper 1 was largely straightforward, with inverse functions, bounds, asymptotes and conditional probability as the weak points and many marks lost to misreading instructions. Paper 2 and the IA comments match the other time zone: a tough but fair paper hurt by premature rounding, and IAs held back by templates and weak reflection.

By Pietro Meloni, PhD · Summarised from the IB subject report on

Maths AI SL grade boundaries, November 2023 TZ2: minimum mark for each grade. Every session
ComponentMax4567
Overall10034486478
Internal assessment209121518
Paper 18026364960
Paper 28027395262

SL Paper 1

A mostly straightforward paper with a few demanding parts; careless reading of instructions and algebraic routes instead of the GDC cost many students marks.

Found difficult

  • inequalities
  • upper and lower bounds
  • inverse functions
  • asymptotes
  • conditional probability

Well prepared

  • reading values from a graph
  • distance between two points
  • tree diagrams
SL Paper 1: question by question, in my words.
QTopicHow it wentWhere marks were lost
1regression line, correlation and validity of a predictionAnswered wellweaker students made errors finding the regression line; imprecise wording when explaining why the model should not be usedTip: Justify unreliable predictions using correct terms such as extrapolation and the variables in the question.
2bounds, unit conversion and volume of a pyramidMost struggledcould not find upper and lower bounds; did not cube the length conversion factor for volume; used a formula for the wrong solidTip: When converting volume units, cube the length conversion factor.
3vertical asymptote, power rule and increasing intervalMost struggledwrong vertical asymptote equation; did not simplify the last term of the derivative; found the minimum but not the increasing interval; left the question blankTip: After finding a minimum, state the interval of x values where the derivative is positive.
4Voronoi diagram and perpendicular bisectorMixedwasted time converting a correct point-gradient equation to another form; drew the bisector through an endpoint instead of the midpointTip: A perpendicular bisector passes through the midpoint; point-gradient form is fine unless another form is requested.
5sine rule with bearingsMixedassumed a right-angled triangle and used a basic trig ratioTip: Check that a triangle really has a right angle before using SOH-CAH-TOA; otherwise use the sine rule.
6value of an inverse function and its domainMost struggleddifferentiated instead of inverting; attempted algebraic inversion unsuccessfully; wrote h(1) instead of solving h(x) = 1; wrong domain for the inverseTip: To find an inverse value, solve f(x) = value on the GDC; do not confuse inverse with derivative notation.
7quadratic model from coordinatesMixedused slow algebraic methods instead of solving the system on the GDCTip: Set up the simultaneous equations from points and solve them with the GDC.
8normal to a curve and distanceMost struggledchain rule differentiation errors when finding the gradient; left a decimal in the denominator of a gradient; did not use the negative reciprocal for the normalTip: Get the tangent gradient from the GDC, then take its negative reciprocal for the normal.
9box plot statements and t-testMixedchose three statements when asked for two; compared the p-value with the wrong inequalityTip: Select exactly the number of options asked for; extra choices can score zero.
10tree diagram and conditional probabilityMixederrors with the dice probabilities; gave probabilities greater than one; did not recognise conditional probabilityTip: Check every probability lies between 0 and 1 before moving on.
11binomial distribution with a compound success probabilityMixeddid not recognise a binomial situation; could not find the compound probability needed as pTip: When p is not given, first compute the single-trial probability, then use it in the binomial.
12Spearman rank correlation and hypothesis testMixedsquared r_s; compared r_s with a chi-squared value or p-value instead of the critical valueTip: For Spearman tests compare r_s directly with the critical value given.

What to practise

  • Read instructions carefully, such as how many options to select.
  • Use correct terminology, especially outlier and range.
  • Simplify final answers but do not rearrange line equations unless asked.
  • Give answers to 3 significant figures with the unrounded value shown.
  • Prefer the GDC to lengthy algebra.

SL Paper 2

A stern but fair paper with a slope of difficulty in each question; many respected accuracy instructions, but premature rounding still cost marks.

Found difficult

  • trapezoidal rule
  • unit conversion
  • hypothesis testing
  • chi-squared test
  • loans and amortization
  • cost, revenue and profit
  • analytic methods
  • mathematical modelling

Well prepared

  • evaluating functions
  • definite integrals and area
  • right-angled trigonometry
  • percentage error
  • arcs and sectors
  • mean from a frequency table
  • box plots and outliers
  • loans and amortization
SL Paper 2: question by question, in my words.
QTopicHow it wentWhere marks were lost
1fitting a curve, trapezoidal rule, integral and percentage errorMixedwrong strip width h; missed the brackets around the middle terms; swapped integral and trapezoidal methods between parts; changed the lower limit or omitted dxTip: Compute h as interval width divided by number of strips, or add individual trapezoids if unsure.
2lengths, angles, arc length, sector area and volumeMixedrounded an angle prematurely; forgot to double the half-angle for the arc; subtracted the wrong region from the sector; multiplied lengths in different units for a volumeTip: Convert all lengths to the same unit before computing a volume, and reuse unrounded angles.
3grouped data, quartiles, outliers and chi-squared goodness of fit for a normal distributionMost struggledtreated data as discrete; GDC not set to use frequencies; reversed or poorly worded hypotheses, omitting parameters; could not compute expected frequencies from normal probabilitiesTip: In a goodness of fit test, name the distribution with its parameters, state df, p-value and the comparison.
4car loan with the TVM solver and final paymentMost struggledgave the down payment as the loan; used the compound interest formula for a payment problem; gave a non-integer number of periods; multiplied the payment by the decimal part of nTip: Problems with regular payments need the TVM solver; the number of payments must be a whole number.
5income and profit functions for mugs and optimal selling priceMost struggledincluded n in the income expression; used the given result in its own justification; insufficient justification that a point was not a maximum; could not build the new cost functionTip: Never use the answer you are asked to show; and use show-that results to re-enter later parts.

What to practise

  • Allow about one minute per mark and move on when stuck.
  • Use unrounded values in later steps and give final answers to 3 significant figures.
  • Check units and the reasonableness of answers against the constraints.
  • For goodness of fit, always state the distribution type and parameters.
  • Strengthen analytic skills alongside GDC use.

Internal assessment

What moderators raised about the exploration this session, criterion by criterion.

  • A. lengthy introductions, repeated basic calculations and plans that hurt conciseness
  • A. processes irrelevant to the aim, such as extra tests in a correlation study
  • B. inconsistent case, accuracy and notation; uncited internet diagrams and GDC screenshots
  • C. primary data followed only by routine processes, or functions chosen mechanically
  • D. reflection limited to obvious description and a final list of generic limitations
  • E. assumptions of statistical tests not checked against the data
  • E. models chosen by trial with R squared, often very high degree polynomials, ignoring data trends
  • E. correlation interpreted as causation

About this summary

After every session the IB sends schools a subject report: how candidates did on each paper and question, and what teachers should work on. This page is my reading of the Maths AI SL report for November 2023 TZ2, in my own words. No text from the report or from the exam questions is reproduced, and it is not endorsed by the International Baccalaureate.

Session codes: TZ is the time zone variant of the papers. Other variants of the same session: November 2023 TZ1. Corrections: pietro@pietromeloni.com.

Frequently asked questions

Which questions were hardest in Maths AI SL November 2023 TZ2?

+
The questions most candidates struggled with were on bounds, unit conversion and volume of a pyramid, vertical asymptote, power rule and increasing interval, value of an inverse function and its domain and normal to a curve and distance. The tables above list every question discussed in the report, paper by paper.

What did candidates find difficult in November 2023 TZ2?

+
Across the papers the examiners listed inequalities, upper and lower bounds, inverse functions, asymptotes, conditional probability, trapezoidal rule, unit conversion and hypothesis testing, among others.

What were the grade boundaries for Maths AI SL November 2023 TZ2?

+
Maths AI SL: a 7 from 78 out of 100 overall in November 2023 TZ2. The component boundaries are in the table on this page.

Is this the official IB subject report?

+
No. It is my summary of the report the IB issues to schools after each session, rewritten in my own words with a teacher's reading of it. No text from the report or from exam questions is reproduced. Your school's IB coordinator can access the original.

Find out what is costing the marks.

One to one, online, on Google Meet.

Book a first session