Early prediction of students at risk of doing poorly in CS1 can enable early interventions or class adjustments. Preferably, prediction methods would be lightweight, not requiring much extra activities or data-collection work from instructors beyond what they already do. Previous methods included giving surveys, collecting demographic data, introducing clicker questions into lectures, or using locally-developed systems that analyze programming behavior, requiring some effort by instructors. Today, a widely used textbook / learning system in CS1 classes is zyBooks (used by several hundred thousand students annually), which automatically collects data related to reading, homework, and programming assignments. For our 300+ student CS1 class, we found three data metrics, auto-collected in early weeks (1-4), were good at predicting performance on the midterm exam: non-earnest completion of the assigned readings, struggle on the coding homework, and low scores on the programming assignments, with correlation magnitudes of 0.44, 0.58, and 0.72, respectively. We combined these metrics into a decision tree model combining to predict students at-risk of failing the week 6 midterm exam (<70%, meaning D or F), and achieved 85% accuracy when predicting at-risk students (with 82% sensitivity, 89% specificity), which is higher than previously-published early-prediction approaches. The approach means that thousands of instructors already using zyBooks (or a similar system) can get more accurate early prediction of at-risk students from data being already collected (meaning no extra effort to collect data).