Loading...
  OR  Zero-K Name:    Password:   

Map Selection

4 posts, 242 views
Post comment
Filter:    Player:  
sort
5 months ago
Currently, the map suggestions after a match seem to be randomly (?) chosen from a rather small pool of maps, which combined with some of those maps being more liked than others, seemingly often results in high repetition among the voted for maps in team lobbies.

I'd propose expanding the pool of maps the suggestions are chosen from (or reducing bias towards generally popular ones, if the pool is already larger than it appears to be?), in addition to blacklisting the last few maps the lobby played on from being suggested automatically (alternatively, they could be offered once directly afterwards, in case a rematch is desired). That blacklist could also be a "soft limit", eg. in the form of P = P0 / (1+c*e^-d), with c being a scaling constant and d the number of games since it was last active.

Another "trick" for reduced repetition would be to try to group maps that are similarly likely to be voted for together - this might reduce the occurrence of sets of "choices" that have an obvious winner and 3 "protest vote" options.

More complex systems that took into account historical/stated preferences of the players and adjusted map frequency to match might also be interesting, but might be more difficult to implement as well as difficult to understand for new players.

Toughts?
+0 / -0
5 months ago
Grouping maps that are similarly popular could help balance out the choices, preventing one map from dominating the vote. While more complex systems based on player preferences might offer even more customization, they could be confusing for newer players. Keeping things simple while still improving
+0 / -0
There are fewer maps which many people like than you might think - so increasing the breadth of maps which appear in votes might just increase the number of votes in which all four maps are disliked.
+0 / -0
5 months ago
Is this an actual "like", or an inferred one from votes? Because votes could confuse eg. familiarity for preference...
+0 / -0