Turn on any political panel show in Australia and you’ll hear the same thing: absolute certainty. Every pundit knows exactly what’s going to happen in the next election. Every columnist has the definitive explanation for why the polls moved three points. Every talking head can tell you, without hesitation, what “ordinary Australians” are thinking.

It’s exhausting. And more often than not, it’s wrong.

The problem with political commentary today isn’t bias—that’s always been there. It’s the absence of humility. We’ve created an entire industry of people who get paid to be confident, even when they’re clueless. Hedging your bets doesn’t get you booked on Insiders. Admitting uncertainty doesn’t drive clicks.

But politics is inherently uncertain. Elections are decided by thousands of individual decisions made by people we’ll never meet, influenced by factors we can barely measure. Economic forecasts are educated guesses at best. Even the politicians themselves don’t know what they’ll do until they’re actually in power and facing real constraints.

Yet our commentary class acts like they’ve got it all figured out. They’ll tell you Scott Morrison lost in 2022 because of X, when it was probably X, Y, and Z, plus a dozen other things we haven’t even considered. They’ll predict the next election result based on current polling, ignoring the fact that current polling is often wildly inaccurate.

I think about this every time I see a confident prediction age poorly. Remember when everyone said Trump couldn’t win in 2016? Or when Brexit was supposed to be a sure loss? Australian commentators aren’t immune either—plenty of smart people were dead certain about election results that went the other way.

What’s frustrating isn’t that they were wrong. We’re all wrong sometimes. It’s that they were so certain while being wrong, and faced so few consequences for it. The same people who confidently predicted one outcome just move on to confidently predicting the next one, with no acknowledgment that maybe their model of politics is incomplete.

There’s a business reason for this, of course. Uncertainty doesn’t sell. “I think Labor might win, but honestly it could go either way depending on several unknowable factors” doesn’t make for great television. “Labor will definitely win and here’s why” is much more compelling, even if it’s equally baseless.

I’ve noticed this affects how we think about politics more broadly. When every commentator is certain, audiences start believing that politics is more predictable than it actually is. We get frustrated when politicians don’t do what the pundits said they would, as if following the script was somehow mandatory. We lose sight of the messy reality that governing is complicated and outcomes are uncertain.

What would humbler political commentary look like? It would involve more phrases like “I might be wrong about this” and “the data is ambiguous.” It would mean acknowledging when you don’t know something rather than confidently speculating. It would require admitting that your political opponents might have a point, even if you disagree with them overall.

Some commentators already do this. They’re the ones worth reading. But they’re outnumbered by the certainty merchants, the people who always have an angle and never have doubts.

I spoke to someone at Team400 recently about how prediction models work in other fields—they build AI systems for businesses, so they think a lot about forecasting. What struck me was how much emphasis they put on quantifying uncertainty. A good model doesn’t just give you an answer; it tells you how confident you should be in that answer. Political commentary could learn from that.

Maybe what we need is a new norm in political journalism: uncertainty scores. Force every pundit to put a confidence level on their predictions. Make them say “I’m 60% confident Labor will win” instead of “Labor will definitely win.” It would be honest, and it would help audiences calibrate their expectations.

Of course, it’ll never happen. The incentives are all wrong. Media organisations want confident voices because that’s what audiences have been trained to expect. Audiences want confident voices because uncertainty is uncomfortable. And pundits want to sound confident because that’s how you build a personal brand.

But we’d all be better served by commentary that acknowledged its own limitations. Politics is hard to predict because people are complicated and contexts are always changing. The commentators who pretend otherwise aren’t smarter than the rest of us—they’re just louder.

I’m not arguing for total relativism here. Some political analysis is clearly better than others. But even the best analysis should come with caveats, with acknowledgment that we’re working with incomplete information and imperfect models.

The next time you hear a political commentator speak with absolute certainty about what’s going to happen, ask yourself: do they actually know, or are they just performing confidence? Chances are, it’s the latter.

And maybe that’s okay, as long as we all know that’s what it is.