26 Sep 2026 · 8 min readBrain

How to Sharpen Your Thinking: A Practical Guide

Thinking clearly is a set of checkable habits, not an intelligence trait. This guide covers the moves that reliably improve judgement: steelmanning the other side, looking for disconfirming evidence, separating correlation from causation, knowing when the base rate matters more than the story, and how to actually make a decision when you will never have complete information.

The first move: state it in one sentence

Most bad thinking starts as a vague cloud. Before analysing anything, write down in one plain sentence what you actually believe and why. If you cannot, you do not yet have a position — you have a mood.

example

Vague:     "AI is taking over everything."
Specific:  "I think AI will reduce entry-level writing and
            design work by more than half within five years,
            because the entry tasks are the most automatable."

Now it can be argued with. The vague version could not.

Argue the other side properly

The fastest way to a bad decision is a private debate where you play both parts badly — a weak case for the other side is not a test, it is a formality.

State the opposing view in its strongest, most sympathetic form. Not a strawman.
Then ask: what is the best explanation for why someone smart holds it?
If you cannot state it well, you have not understood the disagreement, and your confidence is unjustified.
Change your mind cheaply. If new evidence flips you, that is a sign you were thinking, not that you were weak.

💡 Pro Tip: The people with the worst calibrated thinking are usually the most confident, not the least. Confidence is not a measure of correctness — it is a measure of how quickly the words come out.

Check the evidence before the conclusion

Correlation is not cause. Ice cream sales and drownings move together because summer moves both.
Check the base rate. '90% success rate' from 10 trials means something very different from 10,000.
Look for the comparison group. If the claim only reports the good case, there is no finding.
Who benefits if this is true? Follow the incentive when the evidence is thin.
A dramatic single example is almost always less informative than a boring dataset.

The biases that actually cost you

Confirmation bias: you notice evidence for what you already believe and dismiss evidence against it. The fix is to actively look for the strongest case against your position.
Anchoring: the first number or option you see sticks. Ask what the alternative reference point would be.
Sunk cost: continuing because you've already invested, not because the future is good. Ask: if I were starting now, would I do this?
Loss aversion: a 100-point loss feels worse than a 100-point gain feels good, so you take bad trades to avoid admitting one.
Recency: the last thing that happened feels like the pattern. Five data points from last month is not a trend.
Survivorship bias: you only see the people who made it. The thousands who tried the same thing and failed are invisible.

How to decide without complete information

You will almost never have enough information, and waiting for it is usually the more expensive error. These are the moves that help under uncertainty.

Separate the reversible from the irreversible. Test cheaply where you can; deliberate where you can't undo.
Ask what you'd need to be true for this to work, then check whether it is. Often the assumption is the whole decision.
Set a decision date in advance. New information is always available and is the most common reason a decision is never made.
Write down the expected outcome before you commit. Comparing it to reality afterwards is the only way your judgement calibrates.
If two options are close, they're close. Pick one, and stop relitigating.