Matchmaking Factors - Just Shiptype or Skill level as well ?

[gexstar]

I'd like to know what factors play a role in the current matchmaking system. Shiptype and -tier obviously, but what about player skills ? I dare say that player skill HAS TO BE accounted for as well. There is a reason pretty much every other MOBA is doing it. Does anybody know more about it and can shed some light on this ?

[Kastike]

I'd like to know what factors play a role in the current matchmaking system. Shiptype and -tier obviously, but what about player skills ? I dare say that player skill HAS TO BE accounted for as well. There is a reason pretty much every other MOBA is doing it. Does anybody know more about it and can shed some light on this ?

 

MM is random, outside of ship classes of course.

[Hedgehog1963]

Player skills are not factored in MM.  They won't and should not be in a RANDOM battle.

[gexstar]

Player skills are not factored in MM.  They won't and should not be in a RANDOM battle.

 

And who says that the random part is about player skills ? I understood it as random players playing in a game, as in no teams or only small teams (divisions) are allowed. So we wont get a 12 player team vs 12 randoms. So dont act like its written down somewhere the randomness is supposed to be about skill level.

 

But now that i know, i dont get why people are still talking about ship balancing and whatnot. In random battles, the skill level distribution will be the biggest influence by far. No wonder there are so many landslide wins/losses. 

[Kastike]

 

And who says that the random part is about player skills ? I understood it as random players playing in a game, as in no teams or only small teams (divisions) are allowed. So we wont get a 12 player team vs 12 randoms. So dont act like its written down somewhere the randomness is supposed to be about skill level, when its actually not.

 

But now that i know, i dont get why people are still talking about ship balancing and whatnot. In random battles, the skill level distribution will be the biggest influence by far. No wonder there are so many landslide wins/losses. 

 

I've seen plenty of other PvP games boasting about skill based MM, yet I've never seen anyone actually succeeding in it. MM is a pet peeve of any game's forum.

 

Oh, "and who says that the random part isn't about player skills?"... See that, it is rather fruitless approach.

[gexstar]

 

I've seen plenty of other PvP games boasting about skill based MM, yet I've never seen anyone actually succeeding in it. MM is a pet peeve of any game's forum.

 

Oh, "and who says that the random part isn't about player skills?"... See that, it is rather fruitless approach.

 

It is very hard to make it completely fair. I dare say pretty much impossible. But you can definitely balance the odds a little. And in all MOBAs i have played so far, and i've played a shitload, they use some kind of skill based matchmaking. And i've experienced less heavily onesided matchups than i see while playing WoWs. So it wont make things perfect, its not some kind of secret wonderbox, but it is an improvement nonetheless.

 

The second part was really meant as a question. Where is it stated that the randomness in "Random Battle" is meant to be about the players skill level ? I don't see it written down or noted anywhere. Mabe i just dont know where to look.

 

[Quetak]

I believe that player's WR is counted somehow so game try to hold you around 50%.

Ofc good players will be above it, when three such players are in division they have high impact on game and then they can reach 60%+ winrate(or even 70%+).

And on oposite bad players will be below it.

If its totaly random there must be someone with 100% WR for lets say first few hunderd battles (someone always win lotery).

If there isnt anyone with 99,xx% or 0,xx% winrate in 200-300 battles from maybe milion players worldwide it statisticaly cannot be totaly random.

Some hiden mechanism must be there I think, but who knows? WG?

[Hedgehog1963]

 

And who says that the random part is about player skills ? I understood it as random players playing in a game, as in no teams or only small teams (divisions) are allowed. So we wont get a 12 player team vs 12 randoms. So dont act like its written down somewhere the randomness is supposed to be about skill level.

 

But now that i know, i dont get why people are still talking about ship balancing and whatnot. In random battles, the skill level distribution will be the biggest influence by far. No wonder there are so many landslide wins/losses. 

 

I never said randomness is about skill level,  I said they're not factored in.  Big difference.

[lup3s]

 

I never said randomness is about skill level,  I said they're not factored in.  Big difference.

 

imo skill level should be factored in the MM

[gexstar]

 

I never said randomness is about skill level,  I said they're not factored in.  Big difference.

 

Ok. The way you wrote it down though, it sure looked like you wanted to say exactly that. With random written in huge, bold latteres and whatnot. Nonetheless, if they are not factored in, guess what they are... Random.

 

May i ask why you believe they shouldn't be a factor though ? I think it might actually improve the matchmaking system and i dont see any reasons against it to be honest. It can't any worse than random, can it ?

 

 

I believe that player's WR is counted somehow so game try to hold you around 50%.

Ofc good players will be above it, when three such players are in division they have high impact on game and then they can reach 60%+ winrate(or even 70%+).

And on oposite bad players will be below it.

If its totaly random there must be someone with 100% WR for lets say first few hunderd battles (someone always win lotery).

If there isnt anyone with 99,xx% or 0,xx% winrate in 200-300 battles from maybe milion players worldwide it statisticaly cannot be totaly random.

Some hiden mechanism must be there I think, but who knows? WG? 

 

I think you drew a wrong conclusion there. The matchmaking beeing random doesnt mean that the supposedly bad team wont win from time to time. Ingame RNG, Map, ship setup, player mood and lots of other factors weigh in before a winner is determined. The matchmaking should just try to give everyone the same opportunities. But the actual data is probably still interesting and there probably are some oddballs who have very high/low winrates anyway.

 

[crazy5982]

 

It is very hard to make it completely fair. I dare say pretty much impossible. But you can definitely balance the odds a little. And in all MOBAs i have played so far, and i've played a shitload, they use some kind of skill based matchmaking. And i've experienced less heavily onesided matchups than i see while playing WoWs. So it wont make things perfect, its not some kind of secret wonderbox, but it is an improvement nonetheless.

 

The second part was really meant as a question. Where is it stated that the randomness in "Random Battle" is meant to be about the players skill level ? I don't see it written down or noted anywhere. Mabe i just dont know where to look.

 

 

I'd just like to point out that WoWs is not a MOBA, as such shouldn't be compared to one.

 

My view is that skill based MM shouldn't be a thing in WoWs, if you want to win more, (now some people don't like hearing this) play better.

 

There is only one constant in all the battles you play. You. Thus if you want a better win rate play better, don't always rely on your team to carry you to a win, but support them when it's needed.

[snipershot]

If youre doing well you get bad rng. if youre doing badly you get good rng. In wot the devs said the rng is applied at the start of the game, ie your team has a bad game or good game. A player with +25% on his rng will see citadels and detonations, a player with -25% will see full bb salvoes straddling the target, lower damage hits etc. 

How wg define is said to be if you play x games and win, the games will get progressively more difficult to win.

 

Lastly this is mostly speculation and urban myth, but without wg to explain the intricacies thats all there is.

[Hedgehog1963]

 

 

 

May i ask why you believe they shouldn't be a factor though ? I think it might actually improve the matchmaking system and i dont see any reasons against it to be honest. It can't any worse than random, can it ?

 

 

 

 

Because everyone's WR would end up the same.  49%

[gexstar]

 

I'd just like to point out that WoWs is not a MOBA, as such shouldn't be compared to one.

 

My view is that skill based MM shouldn't be a thing in WoWs, if you want to win more, (now some people don't like hearing this) play better.

 

There is only one constant in all the battles you play. You. Thus if you want a better win rate play better, don't always rely on your team to carry you to a win, but support them when it's needed.

 

It's not ? What is your definition of a Multiplayer Online Battle Arena ? Two teams fighting for victory by engaging sepcific objectives on a defined battlefield. People tend to see MOBAs to be just Dota clones, but thats a misconception. And i dont think it is about winning or losing. The winrates could be the same for everyone, i dont mind. As i sated several times, the problem that makes so many games less enjoyable are the heavily onesided games. Landslide victories or catastrophic losses. And i think the occasion of those games could be reduced by introducing some kind of matchmaking that takes the players into account as well. 

 

 Because everyone's WR would end up the same.  49% 

 

I disagree. Look at the many games that use a skill based matchmaking. Better players will still have better winrates, but the games themselves might be more enjoyable. 

 

Thanks for all your feedback though. At least we got a discussion going

 

[praetor_jax]

If youre doing well you get bad rng. if youre doing badly you get good rng. In wot the devs said the rng is applied at the start of the game, ie your team has a bad game or good game. A player with +25% on his rng will see citadels and detonations, a player with -25% will see full bb salvoes straddling the target, lower damage hits etc. 

How wg define is said to be if you play x games and win, the games will get progressively more difficult to win.

 

Lastly this is mostly speculation and urban myth, but without wg to explain the intricacies thats all there is.

Would you mind quoting a source on this?

If you can.

[GenerallKrizmuz]

 

I disagree. Look at the many games that use a skill based matchmaking.

 

name one

[Surmaaja]

name one

 

Counter Strike revolves around skill based MM. Many strategy games do too (ELO rating).

[GenerallKrizmuz]

 

Counter Strike revolves around skill based MM. Many strategy games do too (ELO rating).

League system

 

next

[Surmaaja]

League system

 

next

 

Huh? You can totally play matches with random teams without any sort of league. Well random as in random people with similar skill level as yours. And I'm talking about CSGO matchmaking here.

[gexstar]

name one

 

Starcraft 2, Overwatch, Counter Strike, LoL, Dota, Fifa, you name it. Most companies use some reflection of the players skill level to determine the matchmaking even for "normal" games. They dont just use it for their ranked/league games.

[Kastike]

 

Starcraft 2, Overwatch, Counter Strike, LoL, Dota, Fifa, you name it. Most companies use some reflection of the players skill level to determine the matchmaking even for "normal" games. They dont just use it for their ranked/league games.

 

Can't say anything about Overwatch or CS, so I'll skip those. But Fifa and starcraft don't apply, their "ladder" is 1vs1, determining ones skill in such setup is fairly simple. If LoL and Dota do have skill based MM, it is broken as hell, those few times I've tried those games, I've been matched against seasoned veterans and games have been one sided slug fests. I was even part of winning team at times, because some players were just so good that they could carry me.

 

On personal level I don't think that MM should be based on player skill, even if that would be possible. Matching up players with equal skill levels would destroy the queue times for the unicums and the less fortunate. This would be a problem only at early months, after that MM would start returning similar results as current does. Trying to balance teams in a way that they both would have equal amount of unicums, great, good, mediocre and bad players, would only serve to punish above average players and easen the burden on those who fall below the average. Do I think we should punish skill, no I don't. I'm a OK player and I only way I can get better is by getting my <donkey> handed to me in regular basis. I need to make mistakes to learn from them.

 

I believe that player's WR is counted somehow so game try to hold you around 50%.

Ofc good players will be above it, when three such players are in division they have high impact on game and then they can reach 60%+ winrate(or even 70%+).

And on oposite bad players will be below it.

If its totaly random there must be someone with 100% WR for lets say first few hunderd battles (someone always win lotery).

If there isnt anyone with 99,xx% or 0,xx% winrate in 200-300 battles from maybe milion players worldwide it statisticaly cannot be totaly random.

Some hiden mechanism must be there I think, but who knows? WG?

 

This is what is commonly known as flawed analogy. Having a player with 99,x WR after 200, while possible, is highly unlikely. Comparing it to lottery isn't working. Yes people do win in lottery against greater odds. Problem is that in lottery there are quite often a far greater number of players, who often play with more than just one combination and they do this weekly. With WoWs account, you'll have one try and that is it.

 

Yes there is a mechanism that makes WR hover around 50%, it is not hidden one though, you can read about it in here: https://en.wikipedia.org/wiki/Normal_distribution

 

[gexstar]

Can't say anything about Overwatch or CS, so I'll skip those. But Fifa and starcraft don't apply, their "ladder" is 1vs1, determining ones skill in such setup is fairly simple. If LoL and Dota do have skill based MM, it is broken as hell, those few times I've tried those games, I've been matched against seasoned veterans and games have been one sided slug fests. I was even part of winning team at times, because some players were just so good that they could carry me.

 

Fifa and Starcraft dont apply their ladder rankings to determine the matchup of usual games, but they do both have an internal system for normal games. Same with the other games. You cant see some kind of normal ranking, but it is there and is used. 

 

 On personal level I don't think that MM should be based on player skill, even if that would be possible. Matching up players with equal skill levels would destroy the queue times for the unicums and the less fortunate. This would be a problem only at early months, after that MM would start returning similar results as current does. Trying to balance teams in a way that they both would have equal amount of unicums, great, good, mediocre and bad players, would only serve to punish above average players and easen the burden on those who fall below the average. Do I think we should punish skill, no I don't. I'm a OK player and I only way I can get better is by getting my <donkey> handed to me in regular basis. I need to make mistakes to learn from them.

 

It wouldnt really destroy anything. The queue times might get longer, thats true. The goal here would be that there are players of equal skill in the entire game. One valid point might be the low number of people that play this game. But in the end, wargaming can decide at which point someone has waited long enough and loosen the matchmaking restrictions. These algorithms are nothing special, everyone uses them. Heck, if wargaming gave me access to their current code, i'd write them a new matchmaking algorithm... lol. We already have very accurate tracking of the players stats, which would have been the bigger chunk of work. I dont think it would punish anyone either. Statistically, everyone will probably have about the same winrate as before. Even super unicums get uberfailteams from time to time and cant carry them.

 

And this isn't about improving my or anyones winrate either. You keep saying that people should look to themselves to improve their play and i totally agree with that. This is about making the games balanced. You will still lose, but maybe less often after 5 minutes of utter lemming style failure. You will still win, but mabye less often by totally crushing everyone in sight in a matter of minutes. This will still happen, but probably not as much. Resulting in smoother and more enjoyable games for everyone. Theoretically... ;)

 

[Kastike]

 

Fifa and Starcraft dont apply their ladder rankings to determine the matchup of usual games, but they do both have an internal system for normal games. Same with the other games. You cant see some kind of normal ranking, but it is there and is used.

 

This has nothing to do with my point. My point being that both of those games are 1v1 thus they're very different from games, that have 12 random guys thrown together and told to start duking it out with another 12 random dudes.

 

These algorithms are nothing special, everyone uses them. Heck, if wargaming gave me access to their current code, i'd write them a new matchmaking algorithm... lol.

 

This honestly made me laugh a bit. You could of course post here an algorithm that could do the job, but I know you won't.

 

 

We already have very accurate tracking of the players stats, which would have been the bigger chunk of work. I dont think it would punish anyone either. Statistically, everyone will probably have about the same winrate as before. Even super unicums get uberfailteams from time to time and cant carry them.

 

And this isn't about improving my or anyones winrate either. You keep saying that people should look to themselves to improve their play and i totally agree with that. This is about making the games balanced. You will still lose, but maybe less often after 5 minutes of utter lemming style failure. You will still win, but mabye less often by totally crushing everyone in sight in a matter of minutes. This will still happen, but probably not as much. Resulting in smoother and more enjoyable games for everyone. Theoretically... ;)

 

 

Yes, WG is tracking player stats. And all those stats are measured against player base as a whole, thus they would not work in skill based MM environment. Just imagine a very bad player playing against a very bad player. After hundred thousand matches what would their WRs be? Then make an unicum duke it out with another unicum, what would their WRs be after same ordeal. I bet that all of 'em fine stats would be kinda equal in both cases.

 

There has been a metric <donkey> ton of these threads since the dawn of days, yet I've never seen anyone actually posting a working method, how to measure skill in such environment. I think that WG should just take the obvious path and say that MM is skill based, without actually doing anything - then we could go to the "MM is broken" threads on that subject. Every PvP game in existence has always people arguing how MM is broken and using different games to prove their point.

[gexstar]

Starcraft is a 1vs1 game ? Since when ? 2v2, 3v3, 4v4.

 

Creating such an algorithm takes some time, probably days or longer for a single person. And without knowing their code base, or if it would even be used at all, i wont sit down and put shitloads of work into it for shits and giggles. Would you ? If wargaming would be open to it, i would be happy to help. I've been a software engineer for years and working with optimization problems a lot. 

 

About the last point. Skill based matchmaking takes into account against who you are playing. Meaning if two teams of shitty people fight each other, neither will get a lot of points after winning. If you win against a supposedly better enemy, you will get more points and thus your rating will improve a lot more. But two failers fighting each other for eternity, their stats might ge good, but their machmaking rating would stay the same cause they would both be at 50% winrate. 

 

Finally i don't really see why so many people are AGAINST improving the matchmaking. Lots of arguments that aren't really thought through or are based on insufficient knowledge. No offense, just my subjective impression. Do you really think something like this will make it worse ? Why are you actually against it ? I dont get it. I understand there may be many threads about matchmaking already, but that doesnt mean we shouldnt talk about it, to the contrary. If you tell me right know that wargaming just doesnt give a crap, because they want to earn money and dont make a good/better game, that's fine. Thats something i can understand... But just arguing against something that improves a game you like is outside of my grasp.

 

 

[AgarwaenME]

If youre doing well you get bad rng. if youre doing badly you get good rng. In wot the devs said the rng is applied at the start of the game, ie your team has a bad game or good game. A player with +25% on his rng will see citadels and detonations, a player with -25% will see full bb salvoes straddling the target, lower damage hits etc. 

How wg define is said to be if you play x games and win, the games will get progressively more difficult to win.

 

Lastly this is mostly speculation and urban myth, but without wg to explain the intricacies thats all there is.

 

Stop

 

lying

 

or

 

Stop

 

repeating

 

other

 

peoples

 

lies

 

/There could be a tiny grain of truth that people just make ridiculous conclusions out of. It could very well be that the server generates a list of random values at the start of a battle, which it pulls values out of whenever a random value is needed, as this reduces the delay and server load when it runs its functions for hits, misses and whatever else. However as you have absolutely no way of knowing which players on which team will end up "using" which value then this is no way whatsoever gives either team an advantage.