We need it on this date. How much can we have by then?
There are there steps here :
1. Determine how many iterations you have
2. Estimate velocity as a range
3. Use that "range x the number of iterations" to partition the backlog into Will Have, Might Have, and Won't Have.
In general you want to balance for two types of risk.
Always determine the range even if you need to communicate a single value. So you can show the truth about the project by sharing the math above. Then find an appropriate balance between delivery and expectation risk. And you can communicate that as your point estimate.
You definitely don’t want to always promise a boss, client, or customer that you will deliver at either of these extremes. In most cases you will want to promise to deliver some amount of the might-have region of the product backlog. Exactly how much to promise should be determined by finding an appropriate balance between failing to meet initial expectations and failing to deliver all you’ve promised.