Starcraft is talked about as one game, but in fact it is six: three symmetric and three asymmetric. These six have things in common, but the differences are quite large – at a glance it's hard to tell that, say, a TvZ and a PvP are even the same game.
On ladder, players can restrict the possibilities to three of these by selecting a race, but they also have the option to select random. In this case they can wind up in any of the six games, and for the three asymmetric ones they can wind up on either side, meaning there are nine matchups they need to learn. In game, though, they will find out which matchup they are in, provided their opponent is not also random. Their opponent, meanwhile, does not start with that information and is forced to play without knowing which of the three possible games they are playing until they have a chance to discover in-game what race they are facing, creating an apparent imbalance. This thread’s purpose is to discuss that imbalance, and whether it should be changed in Legacy of the Void.
Random Imbalance
Balance is when more skilled players will beat less skilled players the majority of the time, and equally skilled players have equal chance of winning at the start of the game. Imbalance, then, is when a less skilled player still starts with a greater chance of winning, and between equally skilled players one is favored over the other from the start of the match. On this definition, does random advantage constitute imbalance?
If Starcraft is considered as one game, the question is difficult – the non-random player has missing information, but the random player must learn to play all the other matchups, so their skill in any one matchup will have less depth. Considering Starcraft as six games, though, the question is unequivocal. If a less skilled PvPer has a greater chance of winning than his more skilled opponent because he chose random, that PvP is imbalanced.
Edit: Since there's been some confusion on this point, let me rephrase like this: in a balanced game, if a player plays better than their opponent, they should win the majority of the time. If there's a game in which one player has to play better just to have an equal chance of winning, that would be called an imbalanced game. Random advantage creates exactly this situation: the non-random player has to either perform particularly well, or rely on his opponent making mistakes in order to retain an equal chance of winning with his opponent. The random player, meanwhile, can play worse than his opponent and still win more often than not.
Why Now?
Random advantage has usually been fairly small. It only affected matchups where a significant deviation would occur early in the game, before you could scout your opponent, and Starcraft 2 has generally had a significant downtime at the beginning of each match where players are just building workers and starting their basic production, so it was usually possible to send an early worker scout and play more or less normally. There have been a few exceptions: ZvZ deviates very early from ZvT and ZvP builds; PvZ has at some times in SC2 history depended heavily on FFE, while FFE in PvP or PvT was not doable; and in HotS TvP, many early deviations are needed to defend blink and oracle rushes.
The advantage becomes much larger in LotV because early game downtime has been reduced. Now nearly every matchup is likely to have deviations that would occur before the opponent can be scouted. The metagame isn’t very resolved yet, and many changes to the game are still in store, so it is difficult to tell exactly how strong random advantage will be – but it seems likely that some matchups will have deviations in the first minute of the game that would only make sense against that race. Consider an all-in like this and imagine trying to hold it without even knowing your opponent is Protoss from the beginning of the game. Then consider that whatever early deviations Terran needs to take to defend this build are necessary in every game versus a random player, whether or not they even rolled Protoss.
The Solution
The solution here is trivial: simply display the opponent’s race to the non-random player, just as in any other game. Many random players announce their race at the beginning of the game anyway, simply because they don’t want to use the crutch; the difficulty with this solution is that even if all random players did so, there is no way for their opponent to know whether they are telling the truth. With this adjustment that issue no longer exists.
This change might simply become necessary as a few matchups might be virtually unwinnable without some deviation in the first minute that, in any other matchup, would be a huge disadvantage. This happened in WoL and HotS even with the early game downtime, and LotV exacerbates the issue dramatically. Other solutions to the problem would be much more involved, such as removing any early-game threat that would require such a deviation from the game or limiting the map pool to small, 2 player maps where scouting can occur early enough to not be an issue.
Separate MMRs for each race
This is a suggestion that came later in this thread, but I thought should be added to the OP. As the system exists now, your MMR is the same whatever race you choose at the beginning of the game. This is a bit counterintuitive, since if a Terran player decides to pick Zerg one game, they will surely be a lot worse at the game. The result is that if you want to learn a new race, you need to:
a) buy a new account on which you main that new race
b) repeatedly leave games to bomb your MMR down to a level where you can play competitively with that race
c) lose a lot of games trying and failing with the new race while you try to figure things out (and probably bombing your MMR as a result anyway)
This subject perhaps merits its own thread, but I bring it up here because the situation is even more difficult if you play random. With one overall MMR for all three races, you will inevitably have some uneven arrangements. If, say, your Protoss is strongest, followed by your Terran, followed by your Zerg, then the games where you roll Protoss your opponents will be far too easy, and the games where you roll Zerg will be far too hard (even cushioned slightly by random advantage). Only for your Terran games will the MMR have achieved something like what it is intended to achieve: matching similarly skilled opponents together so the games are close.
The solution is easy enough: Give each player not one MMR, but three. For most players who only play one race, the off-race MMRs are insignificant, but for anyone seriously trying to learn an off-race they can do so without dropping their ranking down to a level where the games will be close. There is always some risk that players will abuse this to keep one race at low levels and try to give low-level players a hard time, although this isn't a very serious concern with the solution for a few reasons:
-This happens a lot anyway, and Blizzard has never seemed very interested in stopping it. Most commonly players do this to farm portraits, for which they would want to play random to get the associated portraits; in this system, you would have to bomb all three race's MMRs to do that.
-Blizzard could easily monitor this if they so desired. Any player who was maintaining a very high MMR with one race while keeping a very low one with another race would stand out as someone to keep an eye on, and if they showed other problematic signs (lots of very short games packed together, indicating that they were bombing their MMR intentionally, for instance), Blizzard could easily take appropriate action against such players.
-This might not be such a bad thing anyway. If I main Terran, but once in a while I like to do a bunch of dumb cannon rushes in silver or bronze, why is that bad? I'm cheesing like any other player can, and if those players want to learn to beat my dumb cheeses, they can get free wins off me and go about their business. I'm not violating ToS or anything, so why shouldn't I be able to play that way if it's fun for me?
FAQ
-"If Random is so overpowered, why don't pros play it?"
+ Show Spoiler +
I don't mean to claim that Random is overpowered in the sense of WoL infestor/broodlord. Just that by selecting random, your opponent automatically receives a handicap that you don't have, which gives you an advantage. For professional players, there are a lot of ways you can spend your time to get an advantage. You can study the latest top Korean VODs to learn the latest strategies for a given matchup. You can study an upcoming opponent's replays to find potential weaknesses to exploit. You can practice your builds with teammates to learn how to respond to obscure strategies an opponent could throw at you. All of these will give you an advantage when you play, but you don't have time to do all of them all the time. With that in mind, learning to play all three races equally well has less pay-off than other ways to gain an advantage, particularly since unlike those other advantages, this one comes down to a coin flip. In some matchups on certain maps it might be worth getting the random advantage, but pro players would prefer something more reliable.
These forums get a lot of threads based on the premise of such-and-such strategy being too strong and needing a nerf. This is not one of those threads. The point is not that learning all three races to get an information advantage is too strong a strategy; the point is that it doesn't make sense in the first place for someone to get an advantage in the game just for choosing to roll dice for their race instead of picking one. No other game I can think of gives players an in-game advantage for choosing to randomly roll their race or character, and I don't see why Starcraft should be different.
These forums get a lot of threads based on the premise of such-and-such strategy being too strong and needing a nerf. This is not one of those threads. The point is not that learning all three races to get an information advantage is too strong a strategy; the point is that it doesn't make sense in the first place for someone to get an advantage in the game just for choosing to roll dice for their race instead of picking one. No other game I can think of gives players an in-game advantage for choosing to randomly roll their race or character, and I don't see why Starcraft should be different.
-"Why don't you just send one of your starting workers? You have twelve now, problem solved!"
+ Show Spoiler +
First of all, this answer only works when the starting worker will still get there soon enough for you to do the deviations necessary to remain competitive in a matchup. Against a random opponent you have to make the choice of whether to go gas or gasless in the first minute or two before you scout them blind. On a four player map you might easily get a couple minutes into the game without knowing what race to prepare for, which is a lot of time for decisions to occur.
Second, even with maps and matchups which make the random advantage small, it still violates a fairly basic principle of competition, that ideally both players should start on equal footing. That means that if I play better than my opponent, I should win more often than not. Random advantage violates this principle, and the violation can be easily remedied by the above solution.
Second, even with maps and matchups which make the random advantage small, it still violates a fairly basic principle of competition, that ideally both players should start on equal footing. That means that if I play better than my opponent, I should win more often than not. Random advantage violates this principle, and the violation can be easily remedied by the above solution.
-"Random players are awesome for learning all nine matchups, don't they deserve some advantage for it?"
+ Show Spoiler +
Even if we accept that premise, why can't we have a conversation about what that advantage should be? The present system generally forces the non-random player to make some coin-flip decisions early game, so when they guess the right matchup they're on equal footing but when they guess wrong they're behind. Meanwhile the random player is insulated from any race-specific cheese, encouraging sloppiness in their builds because they never learn to defend those cheeses.
Why not give them a 5% discount on their units? or let units build a couple percent faster? or let them start the game ten seconds earlier? Or give them a 210 supply cap? All of those changes would achieve the goal of giving random players an advantage for their trouble, while avoiding the coin flips and sloppiness associated with the current system.
Why not give them a 5% discount on their units? or let units build a couple percent faster? or let them start the game ten seconds earlier? Or give them a 210 supply cap? All of those changes would achieve the goal of giving random players an advantage for their trouble, while avoiding the coin flips and sloppiness associated with the current system.
Poll: Do you play random? Do you think race info should be displayed?
I play random and I think both players' races should be displayed at the start. (31)
10%
I play random and I think the random player's race shouldn't be displayed at the start. (65)
21%
I don't play random and I think both player's races should be displayed at the start. (93)
30%
I don't play random and I think the random player's race shouldn't be displayed at the start. (112)
36%
I play random and I think neither player's race should be displayed at the start. (3)
1%
I don't play random and I think neither player's race should be displayed at the start. (6)
2%
310 total votes
I play random and I think the random player's race shouldn't be displayed at the start. (65)
I don't play random and I think both player's races should be displayed at the start. (93)
I don't play random and I think the random player's race shouldn't be displayed at the start. (112)
I play random and I think neither player's race should be displayed at the start. (3)
I don't play random and I think neither player's race should be displayed at the start. (6)
310 total votes
Your vote: Do you play random? Do you think race info should be displayed?
(Vote): I play random and I think both players' races should be displayed at the start.
(Vote): I play random and I think the random player's race shouldn't be displayed at the start.
(Vote): I don't play random and I think both player's races should be displayed at the start.
(Vote): I don't play random and I think the random player's race shouldn't be displayed at the start.
(Vote): I play random and I think neither player's race should be displayed at the start.
(Vote): I don't play random and I think neither player's race should be displayed at the start.
Edit: added a poll
Edit: added FAQ
Edit: Added a section about separating each player's MMR into different MMRs for each race.