Overwatch has 21 maps in the primary modes and appears to put them into relatively even use.
The thing is, the data I’ve seen strongly suggests that players do not like all of the maps evenly.
For example, the last major forum survey had 32% of the game liking Kings Row the most vs 1% for Volskaya Industries.
Polls on Reddit have had similar results:
I would suggest putting a survey in game to check and see if the results in game match up with what players have responded to surveys with. I’d recommenced asking players to rate map types, maps, and to describe what they like/dislike about maps if they want.
If it’s incorrect and just represents a vocal minority then it would be helpful to tell players.
For example, you have players that are dodging maps partially in the belief that nobody wants to play them and that they are doing people a favor. If that isn’t the case, players should know.
If on the other hand players really do near universally dislike a map the developers should know.
This data can also potentially be used to improve the quality of matches for most players if the data shows significant gaps in map preferences.
The matchmaker could attempt to select a map type then a map once it has the twelve players for the match based their responses to a survey.
I would suggest the following steps:
-
Check to see if 9 of the 12 in the match have rated the same map type highest. If they have automatically select that map type for the match and jump to step 4 otherwise go to step 2.
-
Check to see if 9 of the 12 in the match have rated the same map type the lowest if they have then remove that map type from being a possibility that can be selected.
-
Based on the player’s map type ratings for the remaining map types assign a % chance for the map type with a minimum of % for each option(to ensure that even unpopular content will occasionally be played if someone likes it. I’d recommend 1/2 of the normal value to start with) then average the results for all the players and roll the dice to see which map type is selected. If the map type selected was one of the two less likely options mark the selection as minority.
-
Once the map type is selected the matchmaker will then attempt to select a map. If 9 of the 12 people have rated one map in the type highest then that map will be selected and the match will start otherwise, it will proceed to the next step.
-
If 9 of the 12 players have rated one map in the type the lowest then that map will be removed from selection in the next step.
-
Based on the player’s map type ratings for the remaining maps assign a % chance for the map type with a minimum of % for each option(to ensure that even unpopular content will occasionally be played if someone likes it. I’d recommend 1/2 of the normal value to start with). If the map type selected was a minority count the each person that selected that option as three people (the purpose for this it to try and make the people that like unpopular map types happier at least some of the time) and then average the results for the players and then roll the dice to see what is selected.
TLDR:
- Public stats suggest people do not like maps evenly.
- The game makes people play maps evenly.
- An ingame survey should be done to check if the public stats are correct.
- If the public stats are wrong players should be told.
- If the public stats are right, developers should consider it in future development.
- The survey can be used to potentially provide better matchmaking selection for most people.
- If 75% of the match likes one map type the most type the most send them to it.
- If 75% of the match likes one map type the least don’t send them to it.
- Otherwise, do a roll based off player’s averaged preferences.
- Do the same thing for map selection with the additional step of putting extra value on people that like unpopular maps.
Expected outcomes:
Players will play the popular maps somewhat more frequently than they do now and play the unpopular maps less frequently.
Expected problems:
Players in a six stack with identical preferences may produce repetitive matches if certain map types are too popular.
For example, let’s say I get a six stack that likes Hybrid Maps the most and likes King’s Row the most. If 50% of the other team like Hybrid maps and King’s Row the most that stack will be going to King’s Row. Even if less than 50% of the other team likes Hybrid/King’s Row you still will have a sizable chance of ending up on it.
Players may attempt to min max their survey response rather than being honest.
Let’s suppose I think the data is going to be used in matchmaking. King’s Row might be my second favorite map but, I might strategically say it’s #1 because I’d rather play it than every other map in the game besides my favorite and I think lots of other people will answer that way.
Players that dislike popular maps will be stuck on them more and some of the unpopular map types may select some of the more hated maps due to slanting the vote in the favor of those that like unpopular map types.
Right now, players are playing each map type around 25% of the time and each map around 5% of the time. It is very likely in my view that the more popular maps in the more popular modes will see an increase in the amount they are selected. I think seeing King’s Row 10% of the time would be awesome but, for some people that would be unappealing. I hate 2CP. It’s entirely possible that the people that like 2CP may like 2CP maps that are totally the opposite of what I like and the system is designed to count their vote higher. This may mean that while I probably will end up on 2CP less, I might end up on the 2CP map I hate the most more which may or may not be worth it.
This will add to server load and if people actually like things at similar levels this won’t accomplish much.
Even if the net benefit out weighs the damage people may feel the damage more strongly than they feel the benefit.
You might not remember that time where you played your favorite map two of the last five games but, you’ll sure be annoyed when you get annoyed when you do that with your least favorite.