Feel free to PM me if you want to team up. I work as a Data Scientist / Software Engineer. Maybe you can help me w/ the data, and we can align the modeling?
2 Likes
Iām not exaggerating. Ask anyone how much fun they were having in SC2 10 years ago in random team games. That was the first time Blizzard started matching random against premade. Before that WC3 had separate arranged team and premade team and no one complained about it.
Current system isnāt bad enough to stop me from playing completely, but is definitely built in such a way to discourage solo players from playing and is something that canāt be rationalized. You canāt disagree how this is one of the things that many people dislike and you canāt just turn the back on them and pretend like there is no room for further improvement.
Premades already enjoy too many perks and bonuses. The entire reward system is build around earning xp that earns you loot chests. Playing with a friend in premade gives you a boost. Winning more often gives another flat boost. Nothing stops little Andy from playing friends but most of the time he may prefer to play alone because playing with his friends results in boring and unchallenging games.
Now this i would like to see. Unfortunately you two would probably have to pay for data from heroes profile which isnāt too expensive but also not really worth it.
I almost exclusively played random team games in SC2 because I didnāt like solo (why I like HOTS). I did just fine for the most part, got to diamond in various seasons? The only really abusive stuff was from certain teams exploiting VERY specific rushes. Those were not a majority of my games. And over time, those were usually ironed out with balance fixes anyways. I lost some, but also won plenty against premades.
But yes, the unique MMR in WC3 is something they should definitely think about. I donāt think it can be a complete reset for every team, but a team MMR modifier that has the potential to push the team MMR higher the more games you consistently win together. Iām not an engineer, so I donāt know the complexity of implementing such ha system, Iām just not sure the Blizzard we have on HOTS today is capable of doing that.
MMR for team games is extremely challenging compared to 1v1. That said, I think the āsolo disadvantageā is only a one angle that shows issues with HOTS matchmaking. Another huge factor Iāve written about is smurfs (especially in Silver/Low Gold). If you think about it, when you solo queue, thereās only 4 possible slots for Smurfs on your team, and 5 on the other.
What if youāre the smurf?
1 Like
I think we know the answer to that 
The point is that for the vast majority of player who just want to ātinkerā with ranked play, none of this is going to be a significant impediment to their ability to compete on the ladder.
If it were up to me, ranked play would REQUIRE team rosters and there wouldnāt be individual rankings at all. Teams would be ranked.
Leave Solo Queue for Quick Match and an Un-ranked Draft mode.
You might as well hate very game out there that offers a solo queue option.
The take home from this is that you shouldnāt solo queue if you care about playing the game seriously.
Except this is a scenario where no argument would help some players change their mind and Iām not being stubborn despite appearing as such.
Whats the problem with increasing queue times for larger parties so they get matched against complete randoms less frequently? With all the group bonuses that premades enjoy, wouldnāt it be just enough to average out the XP reward over time with that of solo players?
Wait longer, better games for both sides, compensated with xp for time wasted in queue. Ostracizing solo players is how you lose even more players. This is not an MMO where you are locked out of content because you donāt wish to join a guild.
Just look at some stats how bad this becomes in ARAM for example where you donāt get matched according to any kind of MMR.
Please do post back if you ever get around to doing it. This is the kind of stuff that actually makes for interesting reading.
Of course 4 and 5 stacks are favored since most 4 or 5 stacks are IRL friends that play together a lot so they know how each person plays.
I am not sure what the OP is trying to do here ā this isnt breakthrough information.
Ummmmm not even going to give a full repsonse to this and I stopped reading your comment after thisā¦
OP is not proving anything but merely showing data that may come to a conclusion that solo players may not be as a huge disadvantage as they claim to be but the data shows some stacks do favour stacks vs solo players. But more data needs to be mined and collected to āproveā solo players are at a disadvantage in different game modes.
Of course you did because none of you can stand the idea of not being the lone wolf hero that carries the rest of the team to victory.
I am by using raw numbers. I donāt think youāre quite understanding the math, but thatās ok, statistics is a high level concept that requires knowledge of more advanced calculus to understand.
When I tally ALL modes where you can end up as a solo queue player, I have taken into account your chances of winning in EVERY situation where you end up as a solo queue player. It doesnāt really matter that I factor in winrates of 5 man groups because if you queue solo you wonāt be in a 5 man group. I just need to factor in when you FACE a 5 man group which I have.
You have too understand what hypothesis testing does. When I fail to reject the null (in the case that the winrate is 50%) it doesnāt mean that the null is true. It means that there isnāt sufficient evidence to support the ALTERNATE. The alternate could still be true, but your data is insufficient to conclude that confidently.
Sure, I would be happy to! The database is MASSIVE, much more that 8k games. There should be plenty of data to make a conclusion there!
I found a free dump of the HotsApi that HeroesProfile uses. All it will cost me is about 20GB of space and my time setting it up. I just need to brush up on my SQL. I never took databases so this is new territory for me, but an excellent thing to put on my resume.
I just canāt break it down by MMR. Iād need to write my own MMR calculator and it would take MONTHS to calculate for data going back to 2017.
I will be sure to. Iām wrapping up all of the time consuming aspects of my garden then Iām going to get the data onto my laptop in Linux and start building the SQL database.
Precisely.
3 Likes
Enough with the āitās advanced calculusā stuff. I do this professionally. Iām realizing though that weāre missing data from HotsProfile:
- We donāt know what the matchmaking queue looks like
You need this in order to calculate the likelihood a solo queuer ends up in a 5v2 or All Solo game for instance (each with their own winrates).
No you donāt. Your probability is already accounted for simply by just summing the data.
As a professional you should have noticed that here you are dealing with humans with emotions, not robots that can be fed data tables to help make sense of the world.
If you are happy with the current system a simple statement would have sufficed and no one would get mad at you, otherwise you are coming off as a crazy person in trying belittle other playersā concerns. Itās not a question of scales being perfectly in balance but the overall experience for players on both sides.
2 Likes
Thatās not true. That will calculate the probability of winning given you have at least one solo on the team. What we want to calculate is the probability of winning given that you are solo queuing. They are not the same thing.