Seeing what you expect to see
Suppose you favour a particular opening in a card game, a build in a strategy title, or a way of splitting your resources in a board game. After a few weeks you feel sure it works. Ask yourself how you know. For most players, the honest answer is a mix of memorable wins and a general impression.
That is where confirmation bias operates. It is the tendency to seek, notice, remember and interpret information in ways that support what you already believe. It is not deliberate dishonesty. The mind quietly gives more weight to evidence that fits and less to evidence that does not.
How it shows up in play
Memory. Wins with your favoured approach are vivid and get replayed in your head. Losses with it get filed under “bad luck”, “a strong opponent” or “one mistake”. Wins with an approach you dislike, by contrast, get attributed to luck. The same record can tell opposite stories depending on which explanation you reach for.
Attention. Players tend to look for reasons their approach is right. After a win, they check what went well. After a loss, they often check nothing, or they search specifically for outside causes.
Interpretation. Ambiguous results get read in your favour. A narrow loss is “proof it nearly worked.” A narrow win is “proof it works.”
Search. When looking up advice or community opinion, people often pick sources that already agree with them and skim those that do not.
A small example
Imagine tracking 20 games with a favourite tactic, and you won 11. That is 55 percent, which feels like confirmation. But if you were honestly unsure whether the tactic was any better than your usual play at 50 percent, 11 wins in 20 is well within what chance alone would produce, and nowhere near statistical significance. A fair coin flipped 20 times lands on 11 or more heads quite often, roughly four times in ten. Short-run results can look like confirmation or contradiction without meaning either, and the pull of confirmation bias is to treat the result as the former.
This is why player statistics drawn from a handful of games deserve caution, and why the question to ask is not “did I find support?” but “how likely would I be to see this if the tactic made no difference?”
Why games make it worse
Games where luck and skill mix give biased reasoning a lot of room. When outcomes are noisy, almost any record can be explained either way. A skill game with clear feedback, such as a puzzle with a fixed answer, corrects mistaken beliefs quickly. A game of mixed skill and chance does not, so a belief can survive for years without evidence for it.
Hidden information adds to the problem. If you never see an opponent’s hand or the cards left in a deck, you cannot check what would have happened had you played differently. Players then reason from the cases they do see, which are not a random sample.
Methods that make evidence harder to bend
- Decide the test before you start. Write down what result would make you doubt your approach, and how many games you will play before judging. If you set the bar after seeing the results, bias can set it for you.
- Record every game, not only memorable ones. Include the losses and the dull results. Many tracking mistakes come from logging only when you remember to, which means logging what stood out.
- Compare against an alternative. Play a different approach for a similar number of games and compare, instead of asking whether your favourite “works” in isolation.
- Look for disconfirming evidence on purpose. Ask what a smart opponent of your strategy would point to. Try to make the best case against it.
- Separate decision from outcome. A good decision can lose and a poor one can win. Judge the reasoning at the time, using the information then available.
- Ask someone who disagrees. A friendly critic who reviews your records often sees what you have filtered out.
What to expect
Even with careful records, noisy games rarely give clean answers quickly. Sometimes the right conclusion is that you do not yet know. That is an uncomfortable place to be, because the mind prefers a story, but it is more accurate than a tidy belief built on selective memory.
A strategy may well be sound. The point is to hold it with confidence matched to the evidence, to update when the record changes, and to remember that the games you most want to prove right are the ones you should examine most carefully. Outcomes in games involving chance remain uncertain, and a well-kept record reduces, but never removes, that uncertainty.