Summary
Predicting the resale prices of used electronics is hard because prices change a lot and don’t follow usual patterns. The authors compare several traditional statistical methods and modern deep learning models on a large dataset of smartphone and laptop prices from Poland over three years. They find that deep learning models, especially N-BEATS, make much better predictions for prices farther in the future, while all methods perform similarly well for very short-term predictions. Also, a single deep learning model trained for long-term forecasting can predict prices accurately at shorter times too, saving effort to build many models.
What this means in practice
- •For subscription platform operators: Use deep learning models to improve resale price forecasts and reduce financial risk caused by price prediction errors.
- •For e-commerce pricing teams: Improve pricing strategies for second-hand electronics on online marketplaces by applying validated long-term forecasting models.
Authors
Mateusz Buczyński, Michał Woźniak, Konrad Kaczyński, Anna Wróblewska, Sebastian Kuk
Abstract
Forecasting resale prices of used electronics is critical for subscription-based platforms where pricing errors translate directly into risk. Unlike structured financial markets, second-hand electronics exhibit high volatility, sparse listing histories, and non-normal price dynamics - yet no systematic time-series benchmark exists for this domain. This paper presents the first multi-horizon benchmark of statistical and deep learning forecasting models for used electronics price prediction. We use a large-scale dataset of daily price listings from Polish online marketplaces (January 2022 to March 2025, 100+ smartphone and laptop models) and evaluate eleven models across six horizons from 1 to 365 days, covering classical methods (ARIMA, ETS, Theta), recurrent and convolutional networks (LSTM, TCN), and modern deep architectures (N-BEATS, N-HiTS, TFT, PatchTST, Informer). Three complementary evaluation protocols assess trajectory fitness, one-shot endpoint accuracy, and cross-horizon transfer. N-BEATS achieves the lowest MAPE beyond 30 days, reaching 8.51% at 365 days versus 14.94% for the best statistical baseline - a 43% reduction. At short horizons (1-7 days), all models converge near 0.72% MAPE and the naive baseline remains competitive. A single N-BEATS model trained at 365 days generalizes to all shorter horizons, eliminating the need for horizon-specific models. N-BEATS and N-HiTS also demonstrate superior hyperparameter stability.