Proposal

Project Proposal | Midterm Checkpoint | Final Report

Introduction

Kalshi is a regulated prediction market platform where people can buy and sell shares based on whether or not certain events will occur. Users are able to participate in trades on a wide range of topics such as election results, weather conditions, and current events. The price of a contract reflects a collective estimate on how likely something will actually happen.

Prediction markets aggregate information from many participants into continuously updated prices. The prices on these markets can often serve as useful probability estimates for future events and frequently perform well compared with other forecasting methods [1].

However, these markets are not always perfectly calibrated. For example, prediction-market prices can exhibit a favorite–longshot bias, where high-probability outcomes may be underpriced and low-probability outcomes may be overpriced, especially when there is more time remaining before the event resolves. This suggests that raw market prices can still contain systematic distortions [2].

Recently, Kalshi has emerged as a modern regulated prediction market platform and has been studied as a new source of real-time forecasting data. The Federal Reserve found that Kalshi provides high-frequency, continuously updated expectations data that can be valuable for forecasting, especially in macroeconomic settings [3].

This suggests that while Kalshi markets contain useful predictive information, there may still be room for machine learning to improve prediction or calibration by learning patterns from historical market behavior.

Dataset Description + Links

Our datasets contain structured information about past Kalshi prediction markets. This includes basic information such as identifiers, market categories, and final outcome. Information such as volume, price, and other factors are also tracked. We have found 2 relevant datasets:

• Kalshi Prediction Markets Trades + Labels
• Kalshi Finances/Economics 18 Moth History

Problem Definition

Although prediction markets are useful indicators, they are not perfectly calibrated or efficient forecasters. The core problem this project aims to solve is the following question:

Can we use early market activity and historical trends to predict a market’s final outcome with greater accuracy than the raw probability implied by current prices?

Kalshi’s real-time expectations already provide valuable insights. With the use of machine learning, we may be able to uncover additional patterns from price movements, volume, timing, and other characteristics to forecast better odds. This matters in identifying when market prices are reliable or when they are noisy/inefficient.

Methods

Preprocessing

• Handle rows with missing data by removing them or filling them if possible. This can be accomplished with dropna() or fillna() functions in pandas.

• Ensure formatting consistency and apply categorical encoding to non-numeric features. This can be accomplished with one-hot or mulit-hot encoders in sklearn.

• Normalize numerical features. This can be accomplished with scaling functions in sklearn.

• Detect and decide on what to do with outliers. This can be accomplished with numpy operations.

Potential Algorithms/Models

• Logistic Regression can be used to predict the probability of a binary outcome. It works well as a baseline because it is fast, interpretable, and helps show how individual features influence the prediction.

• Random Forest Classifier combines many decision trees to make more robust predictions. It can help capture nonlinear patterns and interactions between features like price behavior, timing, and market category.

• Gradient Boosting Classifier is a boosting-based model that builds trees sequentially to correct earlier errors and improve performance. It is often effective on structured tabular data and may perform better when the relationships in the market data are more complex.

• K-Means and GMM are clustering methods that allow clusters to have more flexible shapes than K-Means. This makes it useful for modeling market groups that may overlap or have more complex behavior patterns.

Supervised: Logistic Regression, Random Forest, Gradient Boosting, (MLP, LSTM) Unsupervised: K-Means, GMM

Results and Discussion

Our goal is to predict Kalshi market outcomes more accurately than what the current market price suggests alone. Additionally, we will identify which market features are most important when predicting. We aim to use these models to better understand forecasting and market behavior rather than to support manipulative or unfair trading practices.

We expect machine learning models to outperform relying purely on current market price, with tree-based models likely achieving the strongest results. We also expect time-based price features to improve prediction and clustering methods to identify groups of markets with different behavior patterns.

Quantitative Metrics

1. Accuracy — measures the percentage final outcomes that are predicted correctly.

2. F1 Score — balances precision and recall, which is helpful if one class appears more often.

3. ROC-AUC — measures how well the model separates positive and negative outcomes across thresholds.

4. Log Loss — evaluates the quality of predicted probabilities and penalizes confident incorrect predictions more heavily, which is useful because Kalshi prices behave like probabilities.

5. Brier Score — measures the mean squared error of predicted probabilities and is useful for assessing probabilistic predictions.

Gantt Chart

Was too large to use screenshot. Can be found here

Contribution Table

Name	Proposal Contributions
Benjamin Zhao	GitHub setup and documentation
Eshaan Patel	Data preprocessing plan
Liam Weng	Literature review and citations
William Wu	Model selection and evaluation metrics
Ethan Hu	Problem definition, dataset research

References

1. J. Wolfers and E. Zitzewitz, “Prediction Markets,” Journal of Economic Perspectives, vol. 18, no. 2, pp. 107–126, 2004.

2. L. Page and R. T. Clemen, “Do Prediction Markets Produce Well-Calibrated Probability Forecasts?” The Economic Journal, vol. 123, no. 568, pp. 491–513, 2013.

3. A. M. Diercks, J. D. Katz, and J. H. Wright, “Kalshi and the Rise of Macro Markets,” Finance and Economics Discussion Series, no. 2026-010, Board of Governors of the Federal Reserve System, 2026.

Project Award Eligibility

We would like to be considered for the “Outstanding Project” award.