許多學者利用多因子選股模型試圖尋找在單一市場能為投資者創造超額報酬的因子，也加入交易策略以求進一步改善績效，然而鮮少有研究比較在相同選股模型基礎下，不同市場的因子與策略選擇是否會有差異，不同性質的股票市場是否可以運用同一個因子進行選股，相同因子在不同市場對績效的影響與差異。 本研究以Python建置回測系統，同時回測台股與美股市場，以統計檢定驗證14種單因子與4組雙因子個別在台股市場與美股市場的選股效果，並以買入持有、逆勢布林通道、逢低買進、逆勢高低通道、順勢布林通道五種策略驗證套用策略是否可以提高獲利或降低風險，比較台股、美股市場在選擇因子與套用策略是否有差異。 ;Many scholars utilize multi-factor stock selection models in an attempt to identify factors that can generate profit for investors in a single market. They also incorporate trading strategies to improve performance. However, few studies compare whether there are differences in factors and strategy selection between different markets on the same stock selection mode. The question arises whether factors that work in one market can be applied to different markets, and the impact and differences in performance of the same factors across different markets. In this study, a backtesting system is developed using Python to simultaneously backtest the Taiwan and US stock markets. Fourteen single-factor and four dual-factor models are statistically tested to verify their stock selection effectiveness in the Taiwan and US markets. Additionally, five strategies including buy and hold, inverse Bollinger Bands, bargaining hunting, inverse high-low channels, and Bollinger Bands are applied to examine whether the utilization of these strategies can enhance profitability or reduce risk. The study aims to compare whether there are differences in factor selection and strategy implementation between the Taiwan and US stock markets.
