Most political arguments fail long before anyone makes a bad point. They fail because the two people are not arguing about the same thing, or because each is answering the worst version of the other, or because somebody felt cornered and stopped listening. None of those are about intelligence or knowledge. They are habits, and habits can be practised.
What follows is what we have learned from watching thousands of live, timed, one-on-one debates between strangers who were matched precisely because they disagree. Some of it is old debating advice; some of it is what an AI judge, scoring the argument rather than the side, turns out to reward.
1. Agree on the question before you argue about the answer
"Immigration" is not a question. "Should the country admit more refugees next year than it did this year?" is. A surprising share of heated arguments are two people defending different propositions: one is talking about asylum, the other about labour migration, and both are right about the thing they mean.
So start by stating the question as precisely as you can, and check that your opponent would state it the same way. If you cannot agree on the wording, you have found the first real disagreement, and it is often the most interesting one. This is why every Policon debate is held on a single, written proposition that both debaters voted to argue: it removes this failure entirely.
2. Steelman before you strike
A strawman is the weakest version of your opponent's view, built to be knocked down. A steelman is the strongest version: the one a thoughtful person on their side would actually defend. Restate their position in a form they would accept ("so your concern is that a higher minimum wage costs the least experienced workers their first job, is that fair?") before you answer it.
This does three things. It proves you were listening, which lowers the temperature immediately. It forces you to answer the real argument, which is the only answer that persuades anybody watching. And it occasionally reveals that you agree on more than you thought, which is worth knowing. When Policon's judge sees a debater do this, it marks the moment as a steelman, and it is one of the moves that reliably wins the room.
3. Ask before you attack
Questions are underrated weapons. "What would change your mind?" "How would that work in practice?" "Where does that number come from?" A good question makes your opponent do the work of exposing a weak point, and it is much harder to resent than an accusation. It also buys you time to think.
Ask one question at a time and then actually wait for the answer. A barrage of five questions is a speech pretending to be a question, and everyone can tell.
4. Make claims you can support, and say where they come from
"Studies show" is not a source. "The Congressional Budget Office estimated" is. You do not need a citation for everything, but when you put a number on the table, be ready to say roughly where it came from and how confident you are in it. Saying "I think it was around a third, but I would want to check" is stronger than inventing a precise figure, because the precise figure is the one that gets you caught.
The flip side: when your opponent cites something you think is wrong, say what you believe the real figure is, rather than just calling theirs a lie. A correction is an argument; an accusation is not.
5. Concede the points you have lost
Conceding a point feels like losing ground. It is the opposite. "You are right that the pilot programmes were small, but" tells everyone listening that you are honest about the evidence, which makes the rest of what you say more credible. Refusing to concede an obvious point is the fastest way to lose an audience, and with it the debate.
It also keeps the argument on the ground that matters. If you concede the small points quickly, the debate moves to the question you actually disagree about.
6. Talk about the argument, never the person
Motive is the easiest thing to attack and the least persuasive. "You only think that because you are rich" might even be true, and it still says nothing about whether the policy works. Ad hominem attacks are also the moment most debates stop being debates: the other person stops arguing the question and starts defending themselves.
If you notice you are about to describe your opponent rather than their reasoning, stop and find the reasoning instead.
7. Keep it short, and take turns
Long monologues are hard to follow and easy to stop listening to. Make one point, support it, and hand the floor back. Timed formats help enormously here: when each side gets a fixed opening and then a shared floor, nobody can filibuster, and nobody has to fight to be heard. That structure is a large part of why debates between strangers on Policon stay civil far more often than comment sections do.
8. Know what winning means
You will rarely change your opponent's mind in fifteen minutes, and that is not the point. You win a debate by making the better case: clearer, better supported, more responsive to what the other side actually said. Sometimes the person you change is a listener. Sometimes it is you.
The aim of argument, or of discussion, should not be victory, but progress.
Practise it
Reading about debate helps the way reading about swimming helps. The habits above only become habits when you use them against a real person who disagrees with you and is waiting to answer. If you want somewhere to practise, Policon matches you with someone whose politics differ from yours and scores the argument rather than the side; our guide to debate practice has more on getting better over time, and the debate topics page has hundreds of questions to argue.