Four studies investigated individuals' confidence in predicting near future and distant future outcomes. Study 1 found that participants were more confident in theory-based predictions of psychological experiments when these experiments were expected to take place in the more distant future. Studies 2–4 examined participants' confidence in predicting their performance on near and distant future tests. These studies found that in predicting their more distant future performance, participants disregarded the format of the questions (e.g., multiple choice vs. open ended) and relied, instead, on their perceived general knowledge (e.g., history knowledge). Together, the present studies demonstrate that predictions of the more distant future are based on relatively abstract information. Individuals feel more confident in predicting the distant future than the near future when the predictions concern outcomes that are implied by relatively abstract information.