In section three titled Weekly Overnight Returns, the authors check if the weekly close-to-open returns persist one, two, three, or four weeks into the future. The authors make use of quantiles to analyze the data. They sort the weekly overnight returns and then divide these into deciles or 10 quantiles. If we jump to table two, titled Short-Run Persistence of Weekly Overnight Returns and Subsequent Weeks’ Total Return, we can see that decile one contains the lowest overnight returns, while decile 10 contains the highest overnight returns. In the other columns, we can see that decile one has low overnight returns one, two, three, and four weeks later. Similarly, decile 10 has higher positive returns in its subsequent weeks. This is an indication that weekly overnight returns may persist for up to four weeks into the future. We now have some ideas for potential factors and how to implement them. First, we’ll want to calculate overnight returns, which are the returns from the closed price of the previous evening to the open price of the following morning. Next, we’ll aggregate overnight returns by week to get weekly overnight returns. Based on the analysis of this paper, the hypothesis we’ll use for this factor is that higher weekly returns for a stock indicate higher subsequent returns in the near future. In other words, our strategy will be to overweight stocks that have high weekly overnight returns and to underweight stocks that have low weekly returns.