Learning to Trust My Own Answers

Today was a holiday from work because Ganesh Utsav is starting tomorrow, so technically I had a completely free day. Naturally, my brain interpreted "free day" as "opportunity to make a very ambitious schedule and then not follow it perfectly." Human productivity remains a fascinating experiment.

I had planned quite a lot for today, but the day didn't go exactly according to that plan. Still, I did take a C Programming subjective test, and that gave me something useful to analyse.

The C Programming Test

I scored 26 out of 50, which is around 56%.

The test had 30 questions, and I don't remember exactly how many I attempted, but I made two incorrect answers among the questions I attempted. The questions were also fairly lengthy. Interestingly, some of the one-mark questions took more time than the two-mark questions, which is something worth noticing for future tests.

Considering that this is my first subject-wise C Programming test, I felt reasonably okay about the result. I wasn't particularly happy or disappointed. It was more of a "this is where I currently stand" kind of feeling.

But there was one bigger issue with the test.

The Habit I Need to Get Rid Of

I've now given five tests in total:

• Three topic-wise tests in C
• One topic-wise Digital Logic test
• One subjective C Programming test

And in all five tests, I've used my phone to verify answers.

The pattern is pretty simple. I solve a question on paper, and before actually marking the answer, I sometimes check it using ChatGPT or another AI tool. Most of the time, the answer I worked out myself is correct. I'd estimate that I'm wrong only around 10% of the time.

So why am I checking?

Mostly because I'm afraid of making a silly mistake.

I've made those mistakes before. Sometimes I haven't read a question properly. For example, if a question asks me to mark the false statement, I might accidentally treat it as a true-statement question. There have also been some silly calculation mistakes.

I already maintain an Excel sheet where I record these mistakes. I currently have around two such recurring mistakes listed, but simply having them written down hasn't completely solved the problem.

The bigger issue seems to be trusting my own answer.

And this is something I need to fix now.

In the actual GATE exam, there obviously won't be a phone sitting next to me waiting to provide emotional support every time I become 5% uncertain about an answer. So I need to start practising under actual exam conditions.

For my next test, I'm going to try a simple rule: pause for a few seconds before marking the answer, carefully reread the question, and then commit to my answer without checking it with my phone.

The goal isn't to eliminate every mistake. That's unrealistic.

The goal is to make sure that when I get a question wrong, it's my mistake, not a mistake hidden behind an AI-verified score.

Moving on to TOC

My main focus for this week is Theory of Computation, and particularly Deterministic Finite Automata (DFA).

I actually like this topic quite a lot. I'm studying it from Dev Ashok, and the way he teaches DFA makes the topic easier to understand.

My stronger area is constructing the actual DFA diagrams. I'd say I'm around 70–80% confident when it comes to taking a requirement and building the states, circles, and arrows.

The weaker part for me is the other side of the topic: taking a given set or problem statement and expressing the corresponding language using things like Kleene Star and Kleene Plus.

That's the part where I don't feel nearly as confident yet.

Dev Ashok has around 70 questions on this particular concept, so rather than just watching the explanation and assuming I've understood everything, I'm planning to solve around 5–6 questions every day from this subtopic.

That's probably the better way to get comfortable with it.

Tonight's DFA Goal

I've scheduled a three-hour session from 10 PM to 1 AM tonight, with the goal of completing the DFA chapter.

It's a fairly big topic, so we'll see how much I actually manage to finish. At least the target is clear.

For this week, I'm primarily focusing on TOC and trying to cover as much of it as possible.

Today's C test wasn't spectacular, but it gave me something more valuable than another number: a very obvious thing I need to change in my test-taking habit.

I need to learn to trust the work I do on paper.

And, perhaps more importantly, I need to practise making decisions without having ChatGPT sitting beside me like an overqualified exam invigilator.

Tomorrow

The plan is straightforward: continue with TOC, keep working on DFA, and practise the language-expression questions that I'm currently less confident about.

For now, the goal isn't to make everything perfect.

It's to keep finding the weak spots and actually work on them.