When I worked as a consultant I had give the manager estimations of how long I would need for a specific job. After a couple of months I found out that he multiplied my estimations by 3. I was really upset. But he told me that I was just way to optimistic about my estimations.
Later I found out that people in generally are very bad in estimating their work. As a Scrum Master I came across Scrum Poker and my experience is that this helps the team to be more realistic about their estimations.
People in generally are bad in estimating in real numbers. For example estimate how long would you need to paint your house in hours. But it is easy enough to state that painting your whole house is more work than cleaning your car (there are probably exceptions). So Scrum Poker enforces the idea of relative estimations (one being more work that the other).
It turns out that individual people are not as good in estimating as groups. For individual people it is very hard to estimate the number of jelly beans in a jar. But whenever you have many estimations of different people it is statistically possible to calculate the number of beans quite accurate. Therefore Scrum Poker enforces estimating in groups.
For more information about Scrum Poker and how it works precisely see the following link: https://en.wikipedia.org/wiki/Planning_poker
In SAFe they do not only use Scrum Poker for the sprint planning event but also for get the priority right of the big chunks of work using the WSJF method. We prioritized features for the PI planning event via Weighted Shortest Job First (WSJF). For a detailed description see the SAFe website: https://www.scaledagileframework.com/wsjf/. But the main thing is that you prioritize features relative to each other features. The discussion during the WSJF will lead to consensus among the business stakeholders and a clear list of priorities.