Introducing PolitBetBuro
A Web App to Keep Score of Your Political Hot Takes
If you’re anything like us, your team or department lunches have a habit of turning into political hot take festivals. Someone confidently predicts an election result, someone else is absolutely certain that a policy will never pass parliament, and inevitably someone has to point out that it is about time for the Dutch government coalition to finally break again.
Now if our similarities don’t end here, you are well accustomed to the problem that nobody keeps track of who said what. Bold claims quietly disappear into the void and we never get to celebrate accurate predictions – nor do we get to shame those who were wildly out of pocket… Bummer!
To fix this, I built PolitBetBuro, my holiday side project for this year. It’s a small web app designed to keep track of your political predictions, measure how confident you were when you made them, and (most importantly, obviously) earn bragging rights when your hot takes turn out to be correct.
Track your political hot takes — Stop losing track of who predicted what Bet in groups — Create private prediction groups with friends or colleagues Confidence-based scoring — Being right matters, being confidently right matters more No money involved — Just points, stats, and well-earned bragging rights
PolitBetBuro is a web application for tracking political bets and predictions in a structured way. It is not a monetary betting platform! No money changes hands here. The only traded currency in this Planned Betconomy is reputation.
The app allows you to create private groups, invite friends and colleagues, and place forecasts together. Each of your predictions is made with an explicit confidence level, which feeds into a simple scoring system. Being right matters. Being confidently right matters more. Being confidently wrong comes at a cost.
To use PolitBetBuro, you need to sign up first.1 Once registered, you can create groups. Each group is assigned a random ID that you can share with your co-predictors, keeping things lightweight and private. Groups are the core unit of the app: all bets, predictions, and scores are organised within these groups.
The creator of a group automatically becomes its moderator and may promote other members to that role. Moderators are responsible for keeping bets well-formed and relevant.
Any member of a group can propose a new bet. To avoid duplicate questions, ambiguous phrasing, or other unwanted issues, proposed bets must be approved by a moderator before they become active.2
Moderators have access to a moderation screen where they can review incoming proposals, approve them, reject them, or adjust their wording before opening them up to the entire group.
Once a bet has been approved, all group members can participate. Currently, PolitBetBuro supports binary outcomes only. A prediction is either YES or NO. Alongside your chosen outcome, you must also declare how confident you are in that prediction.
This confidence plays a central role. It affects how risky your bet is considered and influences both the group’s aggregated prediction and your eventual score. Making bold predictions can pay off, but only if you’re right!
Based on all submitted predictions, the app computes a simple estimate of how likely the outcome is to be YES. Each participant contributes their predicted outcome and their confidence. The system aggregates these inputs to produce a collective probability.
The underlying model at this stage is a simplified one. It’s not meant to be a sophisticated forecasting tool, but rather a quick way of showing whether a bet reflects strong consensus or a highly controversial call. In the future, this could be replaced with a more advanced approach. A pure Bayesian model is one easily implementable option. Regression-based models could also be considered.
Assume an array of bets \(i \in \{1, \dots, n\}\), each placing a predicted outcome \(y_{i} \in \{+1, -1\}\) (where “YES” is \(+1\) and “NO” is \(-1\)) and declaring a confidence \(c_{i} \in [0,1]\).
The predicted probability \(P\) of the outcome being “YES” is calculated as:
\[ P = \frac{1}{n} \sum_{i=1}^{n}(.5+y_i \times \frac{c_{i}}{2}) \]
Once the real world outcome of a prediction is known, a moderator can close the bet by declaring the outcome. At that point, points are automatically distributed.
Correct predictions earn a base score of 100 points, while incorrect ones earn nothing. On top of that, each participant receives a bonus or penalty based on their confidence and how much their prediction diverged from the hive mind. This “hot take factor” rewards confident predictions that go against the grain. That is as long as they turn out to be correct! Otherwise, the same factor is subtracted from your score.
In short: Risk pays off, but only if you’re right. If you’re wrong… well, you get what you deserve.
Based on the previous specifications, the hot take factor \(\Delta_{i}\) for participant \(i\) is calculated as:
\[ \Delta_{i} = 100c_{i} \times \frac{1-y_{i}(2P - 1)}{2} \]
To keep communication manageable in larger or less tightly connected groups, PolitBetBuro supports email notifications. You can choose to be notified about new bets or proposals3 instantly, daily or weekly.4 You can adjust these preferences in the settings menu.
In the settings section, you’ll also find your all-time stats, allowing you to look back and assess whether you are the careful analyst you make yourself out to be or solely a habitual hot take merchant.
Currently, the app is localised in both German and English, with desktop and (a rather imperfect) mobile views available. I am very thankful for input on any issues you might encounter about these fronts, so please do let me know!
That’s PolitBetBuro. It’s an experiment, a side project, and a gentle attempt to bring a bit of accountability to your political lunch-table chit-chat. I hope you find it fun and engaging. If you give it a try, let me know what you think! I am genuinely curious to learn about your experiences and to see whether we are being rewarded for our insight or just going to expose our collective folly.