What symbols are available depends on the context.
There are also others: ln, log2, sin,
div, etc.
comparisons
symbol
write
<
<
≤
<=
=
=
≠
!=
≥
>=
>
>
basic logic
symbol
write
remark
∧
/\
and (also and
works)
∨
\/
or (also or
works)
¬
!
not (also not
works)
F
FF
false
T
TT
true
U
UU
undefined
→
-->
material
implication
↔
<->
material
equivalence
&&
&&
short-circuit and
||
||
short-circuit or
reasoning
symbol
write
remark
⇒
==>
⇐
<==
⇔
<=>
treats U and
F as equivalent
≡
===
treats U and
F as different
A subproof occurs between subproof and subend.
The first formula of a (sub)proof can be referred to as original.
At the very beginning of a subproof, original refers to the first
formula on the previous level.
A (sub)proof can be made to trust on an assumption by preceding it with
assumeformula;.
quantifiers
symbol
write
∀ x:
AA x:
∀ x; 0 ≤ x < y:
AA x; 0 <= x < y:
∃ x:
EE x:
∃ x; x + 2 ≠ z:
EE x; x+2 != z:
The syntax of the part between ; and : has been restricted.
An Induction Problem Inspired by Alarfaj and Sangwin 2022
Antti Valmari 2025-04-09 (Small language fixes 2025-05-22)
This web page is based on the running example in Maryam Alarfaj & Chris
Sangwin,
“Updating STACK Potential Response Trees Based on Separated Concerns”,
iJET ‒ Vol. 17, No. 23, 2022, DOI 10.3991/IJET.V17I23.35929.
The example consists of three tasks that are related to finding an induction
proof of a certain summation formula.
In this web page the proof is dealt with in full and it has been organized
similarly to how such proofs are typically presented in textbooks.
The tasks have been formulated so that our feedback tool can check the
answers, although it has no feature for representing summations.
While STACK emphasises assessment, our tool has been designed for learning by
trying tasks, getting feedback, and trying again with an improved answer.
Instructions
Our feedback too is called MathCheck.
It does not know who you are and it does not store your answers.
Feel free to try both correct and incorrect answers!
Light brown regionsCongratulations!
You made this visible! contain hidden text that becomes visible
when you move the mouse cursor there.
Help for typing mathematical symbols is available at the top right part of the
screen.
The Goal
The goal is to prove, for every natural number n, the following
formula using mathematical induction:
n
∑
k = 1
(2k − 1)2 =
n (2n − 1) (2n + 1)
3
The Base Case
Why are the left and right hand sides equal in the base case?
AnswerWhen n = 0, the right hand
side of the original formula (1) obviously yields 0.
Also the left hand side yields 0, because then it is a sum over an empty
range.
The Induction Step
In this exercise, a linear factor is something like −3n + 4
or anything simpler.
We must show that the left hand side of (2) is equal to the simplified
right hand side of (2).
We will do so in the next few steps.
What must be done with this result to complete the induction step?
AnswerOne must check that it is equal
to the right hand side of (2).
Alarfaj's and Sangwin's Tasks
The first of the three tasks in [AS22] was, in essence, the same as our
(2).
However, while in our case “∑” and “=” were given and the task consisted of
filling in four answer boxes, in their case (1) was given the name “statement
P(n)”, only one answer box was given, and the task was to write
statement P(n + 1) in it.
Thus realizing that the answer must be an equation (and not just its one or
the other side) was part of the challenge.
Earlier on, almost 60 % of students had failed that part.
This motivated the authors of [AS22] to improve the potential response tree of
the problem.
Originally no feedback and no points had been designed for the case where the
answer is not an equation.
In [AS22], feedback and partial points for this case were specified.
Answers close enough to the LHS of (2) were given 0.3 points, and remaining
answers mathematically equivalent to the RHS (and thus also the LHS) were
given 0.4 points.
Comments
The treatment in [AS22] established the same technical facts as our
induction step, but differed in how they were organized.
We started with the LHS of (2) and derived its RHS, using (1) as the induction
assumption.
In [AS22], the LHS of (1) was subtracted from the LHS of (2), the same was
done with the RHSs, and the results turned out the same.
It is unclear to us why this shared answer was
(2 ⋅ (n + 1) − 1)2 instead of (2 ⋅ n + 1)2.
The assessment and feedback mechanism of STACK relies on knowing or
guessing frequently occurring incorrect answers, and designing feedback for
each of them manually.
The main feedback mechanism of our tool does not need such work at all.
It consists of providing an automatically generated counter-example when the
answer is incorrect.
The teacher has the option, but no obligation, of telling the tool to do
additional checks.
For instance, a limit to the length of answers was used many times above, and
sometimes an answer was required to be in a certain form.
Employing such checks does not require knowledge or guesses of the incorrect
answers that students tend to make.
On the other hand, our tool has no assessment mechanism.
It has, however, sometimes been used as an answer-checking machine of another
tool which keeps track of points.