TCDformer-based Momentum Transfer Model for Long-term Sports Prediction

TL;DR

TM2 tackles long-horizon sports prediction by encoding momentum from unstructured time series using LLSA and MODWT, then transferring momentum into forecasts via a trend-oriented MLP and a wavelet-attentive seasonal component. The model decomposes the momentum signal into a trend part x_t and a seasonal part x_s, and combines them to yield P(t+1)=x̄_t(t+1)+x_s(t+1). Final outcomes are determined by comparing momentum across players, with historical rankings as tie-breakers. On the 2023 Wimbledon dataset, TM2 significantly outperformed baselines, reducing MSE by 61.64% and MAE by 63.64%, demonstrating strong predictive performance and interpretability for long-term sports outcomes.

Abstract

Accurate sports prediction is a crucial skill for professional coaches, which can assist in developing effective training strategies and scientific competition tactics. Traditional methods often use complex mathematical statistical techniques to boost predictability, but this often is limited by dataset scale and has difficulty handling long-term predictions with variable distributions, notably underperforming when predicting point-set-game multi-level matches. To deal with this challenge, this paper proposes TM2, a TCDformer-based Momentum Transfer Model for long-term sports prediction, which encompasses a momentum encoding module and a prediction module based on momentum transfer. TM2 initially encodes momentum in large-scale unstructured time series using the local linear scaling approximation (LLSA) module. Then it decomposes the reconstructed time series with momentum transfer into trend and seasonal components. The final prediction results are derived from the additive combination of a multilayer perceptron (MLP) for predicting trend components and wavelet attention mechanisms for seasonal components. Comprehensive experimental results show that on the 2023 Wimbledon men's tournament datasets, TM2 significantly surpasses existing sports prediction models in terms of performance, reducing MSE by 61.64% and MAE by 63.64%.

Figures (6)

• Figure 1: An overview of TCDformer-based momentum transfer model for long-term sports prediction.
• Figure 2: Model result output process based on TCDformer.
• Figure 3: Wavelet attention structure diagram
• Figure 4: Iterative adjustment curve of TM$^{2}$ prediction sequence length.
• Figure 5: Comparison of TM$^{2}$ with existing basic tennis models in terms of F1-score, precision and accuracy metrics, in the task of predicting tennis match results, with data from the 2023 Wimbledon tournament.