I obtained the H2 Math papers shortly after the respective examinations were over, and went ahead to do them. I finished P1 in 1 hour 32 mins, and P2 in 1 hour 52 mins (including the correlation question). I’ll write up some of my thoughts of the papers here!
The questions in the papers, as always, have a wide range of difficulty. There are questions that almost all students should be able to do, questions that very few students should be able to do, and everything in between. This is something that sets it apart from most JC prelim papers. Many papers set by JC teachers comprise of just medium to extra difficult questions, without any room for routine questions for students to build up some confidence and get a suitable number of marks as a floor. I understand why schools tend to do this; they want to prepare their students as much as possible through the assessments they set so that students are exposed to as many different types of non-trivial problems as possible. This objective, however, just doesn’t make for a well-balanced paper.
Another good thing in the papers I did is that the challenge of the papers span across a large variety. There are questions difficult due to the algebraic manipulation, due to the way they are phrased, or because students have just never seen them before. Application questions require students to decipher what is written carefully, and convert what they read into mathematical components so that they can deal with the problem. Questions that require qualitative answers are also not that routine (see P1 Q11iii, P2 Q4iiib, P2 Q9v, P2 Q10ii), requiring students to really be able to infer physical conclusions from their mathematical results and to express these conclusions properly in words. In comparison, school papers do not usually express this level of creativity in the type of difficulty used. There is an overemphasis on making the problems as algebraically tedious as possible, which to be fair is necessary in small doses, but not for too many questions! School paper problems are usually derived from past year A Level problems, just twisted to make them even more difficult to approach. It can be difficult to find original ways of phrasing questions, or original qualitative questions as well from these prelim papers.
There are some things, however, that I didn’t really like in the A Level papers this year as well. First of all I think there is an overemphasis on some very specific skills, failing which students would be heavily penalised for. An example will be the use of quotient rule, which was present in P1 Q2, P1 Q3i and P1 Q11ii. Maxima / Minima problems were also tested twice, once in P1 Q11 and again in P2 Q4. I felt that the papers would have better differentiated if attempts were made to test students on a spread of more diverse skills.
I think that P1 Q10 deserves its own specific criticism. It’s a 14 marks application question on differential equations, specifically on population size changes. I felt that right from the very first part, that the question would not have been very clear for students as to how to model the differential equation. The most important sentence, “the difference every year between the death rate and the birth rate for the population of sheep on the island is 3%” (and before that, the death rate being greater than the birth rate), is supposed to let students deduce that 
Of course, one can also say that if students use the wrong DE and obtain an overly simple and unrealistic expression afterwards, it’s their fault for not having the metacognition to realise that what they have is nonsense and that they need to rethink of their model. I agree to this, and sadly, many students do not possess the level of metacognition needed to realise this. This is something I always try to bring up during tuition when students obtain an answer that obviously doesn’t make sense, but they happily accept it and move on to the next part.
There’s a question which I thought was rather vague: P2 Q10, the hypothesis testing question. The question asked for the “critical region for this test”. Essentially, what they wanted was the range of values of 
There’s also the issue of P2 Q2, where students were tested on a recurrence relation. I believe SEAB has stated that even though recurrence relations are not formally in the syllabus (the deeper parts about them, e.g. convergence and general solving in explicit form, are now in Further Math), however students could be still given such expressions and to tackle problems based on these expressions. However, this very first part Q2ai still required students to describe how the sequence behaves, given a starting value. Schools would almost certainly not have taught how to use the SEQ mode on the GC to list out the terms of a recurrence relation, given that as a chapter itself this is no longer in H2 Math syllabus. Are students supposed to calculate a first few possible terms, and see for themselves that the sequence is increasing and tends to infinity? But with only a first few terms, how are they supposed to logically come up with this conclusion or deduction?
What I thought, however, is a great differentiating question is P1 Q11. This is a application question on maxima / minima, where the first part requires students to show a trigonometric expression which isn’t that trivial to show. Even if students cannot do this part, they still can do the differentiation afterwards, so they don’t get penalised too heavily for not being able to perform the trigonometry. The subsequent parts are minor but also interesting and different, where students need to give a qualitative explanation, consider limits and to pick out specific information in the question to produce a relevant inequality, all of which may be things that students might not have dealt with much in an application question setting but are things that they could come up with provided they think about the question deeply. This question would truly elevate the students who can handle tedious algebraic manipulation and differentiation, as well as those who are able to think deeply about the consequences of the information provided in a question. It is also a test of mental stamina, given that it’s the last question of a 3 hour paper.
P1 Q5 and Q6 also would be a deciding factor in the grades that students may achieve. P1 Q5 is a rather challenging vector properties question which requires students to not only know their basic properties of dot and cross products, but also to see how certain equations simplify to vector equations of a line or of a plane, and to describe them as such. P1 Q6 was a 8 mark complex roots problem, which involves a quadratic equation with complex coefficients. Students who can only handle the most typical problems might lose many marks here, whereas students who are more flexible in their complex number algebra will know how to approach this. Being able to apply the sum / product of roots formulas might be slightly useful as well. I also liked P2 Q8, which was a pretty interesting hybrid P&C question where students need to know how to handle binomial coefficient algebra in order to handle the question competently. Students might easily be misled to think that this was a binomial distribution question when actually it is not.
That’s most of the thoughts about the paper I’ve come up with! I’ll slowly type out the solutions and my comments on the questions and upload them, hopefully by the end of the year!
ZhengMath 