Ask HN: How to Plan Research Projects and Careers

I am wondering how successful computer scientists working in research positions plan their research projects and careers in the face of uncertainty and career pressures such as external deadlines and publication expectations. Time estimation is a notoriously difficult problem in software development and in research, yet managers, deans, and other people who evaluate the performance of researchers for hiring/firing/compensation purposes need researchers to be able to make accurate estimates of the time and effort it takes to complete their work. "Publish or perish" is the reality for researchers to earn a living doing research, but competitive conferences have acceptance rates in the 10%-30% range. How do researchers manage these performance expectations without "perishing"?