The create a new node and blow out the old one is correct. But going forward a different structure will avoid the issue entirely.
Bottom line with any kind of database programming (for us): never tie keys directly to data. Always use some kind of 'random' or other key structure to reference your data.
Firebase provides a perfect tool for this in childByAutoId, which creates a 'random' child node name key.
Chrillewoodz answer is the answer... and to add some additional clarity
teams
randomKey_0:
team_name: "thunder"
wins: 3
draws: 2
loss: 1
randomKey_1:
team_name: "warrior"
wins: 3
draws: 2
loss: 1
randomKey_2:
team_name: "super dudettes"
wins: 100
draws: 0
loss: 0
The randomKey_x is a firebase generated childByAutoId node name.
With this structure, you can change the team name, wins etc on the fly without altering the structure. And because you now have a reference, you can use it in other nodes to reference back to that team. For example, you could have a separate node to keep info about team members
team_members:
randomMemberKey_0
team_member_name: "Scott"
height: "7.5"
belongs_to_team: "randomKey_0"
randomMemberKey_1
team_member_name: "Billy"
height: "6.5"
belongs_to_team: "randomKey_0"
randomMemberKey_2
team_member_name: "Frank"
height: "5.0"
belongs_to_team: "randomKey_1"
With this structure you can query for information about the team itself or another query to search for any team member or members or information about them.