Matchmaking still flawed?

[BartonFink]

[The mods merged my post titled: "The broken matchmaking system is probably a feature, not a bug, I think" to here]

I started to think that the matchmaking process was broken once I reached Level 19, I tink, where it became common for me to lose. I had been playing Mon, I think and that stopped working for me. I then started playing Nor next and can't remember if that was at Level 19 or not. That faction stopped working too, so now I'm currently playing Nil and I'm at Level 17. Again, I've just been hammered often since Level 19. I go into round two with a two card advantage and comfortably lose in the last 2 or 3 plays of cards in the third round. Or my opponent has one card left to play, they're 16 or 20 plus points behind me or whatever it was and...they hammer me with their final play, comfortably winning.

I suppose that the devs would like me to spend a lot of money paying (for a chance to get better cards) to win. That's not something that I am interesting in doing. Grinding is an option, but when you lose so often, the game loses its appeal. It really feels like the matchmaking system is a troll designed to get players to spend real money in the game for the chance to get better cards. It does that with a broken matchmaking system based on your Level, not on your deck. Not sure if me being to play against Level 20 players would make me more likely to win. Not necessarily, I wouldn't think, as I'm sure that plenty of 'lesser' Level players can be script kiddies who just copy OP decks. I'm not that interested in being able to do that and it's not that important to me.

Being able to play against people who your deck is competitive with would be a start, but then, it would be hard for the devs to try and corral people into opting to pay for the chance to get better cards doing that.

 

[DRK3]

@BartonFink First, let me just make some corrections and guidelines for you:

You are not talking about Level, but Rank. In Ranked mode, it makes sense it is called rank. Im correcting you, just because i dont want to see this error propagated with other new players, that would start making the same error.

Second, use this terminology to address each faction, its what most people use and agree:
NR= Northern Realms
MO= Monsters
SK= Skellige
NG= Nilfgaard
SY= Syndicate
SC= Scoiatael

Regarding the metadecks/ netdecks (what you call broken decks, rightfully so), it is really frustrating i know. And the bad news is, the higher you reach on ranked mode, the more of those decks you'll see, and less variety there is.

The good news is that while NR and NG might be dominating now, next tuesday comes a new patch that might change everything, so dont go spending everything to craft a current powerful deck because its possible that just in a few days it will get severely nerfed.

 

[BartonFink]

@BartonFink First, let me just make some corrections and guidelines for you:

You are not talking about Level, but Rank. In Ranked mode, it makes sense it is called rank.

Regarding the metadecks/ netdecks (what you call broken decks, rightfully so), it is really frustrating i know. And the bad news is, the higher you reach on ranked mode, the more of those decks you'll see, and less variety there is.

The good news is that while NR and NG might be dominating now, next tuesday comes a new patch that might change everything, so dont go spending everything to craft a current powerful deck because its possible that just in a few days it will get severely nerfed.

When I play Classic mode, generally my opponents are at a similar Level to me, maybe one or two Levels difference (I am assuming that Level relates to completing mosaics though. Am I wrong on that? I luckily, perhaps, completed a mosaic today and reached 16. I say "luckily" because I've been complaining recently about the matchmaking system). I can't say that I remember how our 'Rank' varies though...maybe it's towards double figure but I can't say for sure.

I'm pretty sure that I'm seeing new cards in factions that I was familiar with and competitive against but which have started smashing me. Can't say that I've noticed them in the Deckbuilder though but I haven't really spent too much time in there recently.

In another thread of mine I suggest that it would be good to be able to matchmake for people who just want to play with the base set factions. I really don't want to have to take too much of an interest in new stuff going on to be competitive. In that other thread I did criticise how factions or decks swerve from one being OP to another one being OP. It's not ideal. It should be the kind of thing made balanced in testing, which I don't get the impression that they do.

 

[SigilFey]

I'm afraid it's just the reality of collectable card games everywhere. Finding the ideal balance is not "easy". Changes will be made, and of course, as new cards are released, it's highly likely that someone will figure out another "OP" deck. That is, after all, the whole point of deckbuilding.

I understand the frustration of encountering decks against which there's nothing the player can do. The matchmaking cannot always be ideal. Whenever you log on to an online game (any competitive online game), the system tries to find a match at your "rank" or "level". This, of course, will be limited by the number of people online, available, and not already in an active game. This is then set against a timer of some sort to ensure players are not sitting around for an extended amount of time waiting for a game (...and wait times can be long even with existing systems).

If the game can't find an opponent before the "timer runs out", it's just going to grab the next available opponent that's as close to your skill level as it can. Sometimes, this is going to be not-so-close. Because that's all that's available. The alternative is players sit there for 30+ minutes waiting for the game to find only opponents that are right in their skill and deck power range.

In short, there is no best answer to the situation. If you're encountering constant defeats, it's just bad luck. Murphy's Law: if something is possible, it will eventually happen. That means that someone out there is going to continuously roll snake-eyes. (I know this phenomenon well -- it's usually me. I once fought a battle in Risk as the defender with nearly 40 troops against, like, 8. I killed one or two units and got wiped out, even with defender's advantage and overwhelming numbers. It just happens.)

 

[BartonFink]

I coincidentally posted a reply a few minutes ago which illustrates what you're saying:

Is this faction the dud at the moment?

I'm not sure if I started getting regularly hammered by opponents once I reached Level 19 because it starts getting more competitive there or not. I think that I started out playing Mon. When that deck started losing a lot I switched to Nor. When that started losing a lot I switched to Nil. I'm...

forums.cdprojektred.com

However, I'm not convinced that the devs test their new cards in closed trials. Are you saying that they do? You would think that they have a lot of computer/simulation/AI power to just see how additions to factions and decks play out. Going on forum user comments, they just go from one extreme to another, making one faction OP, nerfing them, then making another faction OP. That shouldn't be the process. The process should be to not unbalance factions when new cards are added.

I've played Risk online myself. I've probably been dudded by opponents who get lucky with rolling 6s or whatever you need to beat someone with superior numbers to you. Even though it's been years since I played it, I can't remember thinking it particularly rigged.

MTG seemed particularly rigged. I'd like GWENT to not feel like that. I.e. whereas soon as you come up against a 'faction' or a deck there you want to quit because you know that you are going to lose. I would have liked to have played my Sliver deck there, but it was a slow deck and could easily lose to other decks. Being able to filter out fast decks or other decks would have been handy there. That was one of my suggestions in my thread here about how to improve the game.

 

[CallMeHoot]

So, this is a weird thing that has often crossed my mind, but I've always dismissed it as some kind of internal bias.

I've got it into my head that the matchmaking is looking at more than just MMR, but also deck composition and for whatever reason the game starts matching you against people running cards that counter your cards directly.

For example, I played a whole week using 1 or 2 decks, never seen a Lacerate played. Never seen a Werecat. A Peter Sars Gwynleve. The green NG artifact that destroys units of the same charge counter...never seen any of these cards. Not in weeks.

Start playing a Swarm deck and all of a sudden I'm running into Forktails, Werecats, etc, all this shit that just shits down a Glutsyworp/Swarm type deck. I mean, is it bias? Am I just unlucky?

EDIT :- Literally played against a guy who's got double Lacerate in the final round, I mean...who the fuck even runs 2 copies of Lacerate in a monster deck? Lmao!

 

[Draconifors]

I mean, is it bias? Am I just unlucky?

Yes, and yes.

Thread merged; this has been questioned since at least 2018. The REDs have stated that decks do not affect matchmaking. I think there's a link to such a statement in here somewhere.

 

[CallMeHoot]

Yes, and yes.

Thread merged; this has been questioned since at least 2018. The REDs have stated that decks do not affect matchmaking. I think there's a link to such a statement in here somewhere.

I dunno, man. I'm not sure I believe them. This thread is loaded with anecdotal evidence. I know it's a long way from actual proof but it's weird that a lot of people are having almost the exact same experience.

 

[Draconifors]

it's weird that a lot of people are having almost the exact same experience.

What would be even weirder -- and I know this has been pointed out multiple times already -- is the system being rigged only for some players. It would make no sense.

I also used to think there was something going on with the matchmaking (there's a post of mine in some old thread saying that), but I never did have any concrete evidence. The RED statement came after, and ever since that I've not bought the theories of decks being a factor in matchmaking.

 

[Balgar-bard-slayer]

What would be even weirder -- and I know this has been pointed out multiple times already -- is the system being rigged only for some players. It would make no sense.

I also used to think there was something going on with the matchmaking (there's a post of mine in some old thread saying that), but I never did have any concrete evidence. The RED statement came after, and ever since that I've not bought the theories of decks being a factor in matchmaking.

I think that is a separate thing. It's not about it being rigged against one player or another. But it's apparent that they are not being completely honest about how their algorithm works.

I'm pretty sure that the algorithm helped my MO insects deck to climb to Pro rank. At the same time it completely nullified my attempts to play a Witcher deck.

Case in point whenever I attempt to play my NG Witcher deck, I will get a sequence of back to back games sometimes three in a row against NG control decks. And this happens every time I try for a run of games using that deck. if you attempt to play a deck that runs a large number of Witcher cards, 70%+ of the games will be against NG. And I strongly suspect this is because it runs the most natural counter cards (locks and poison packages) and that has happened at every rank including Pro.

So I am confident to say matchmaking is based on what is in the decks.

Note I am not saying it is rigged against me as an individual player, but that archetype/decks are matched with decks that include specific counter cards.

Again I am sure the algorithm has favoured me too at times, but that is actually irrelevant to the argument. What is relevant is the example above due to its repeatability sticks out like a sore thumb.

Also, people want statistics and not anecdote, but RED make it very difficult to get statistics on decks, because you cannot track the win loss of specific decks, only factions statistics through playgwent.com. Also I've noted when I have posted statistics supporting criticism of the game my posts get deleted...just saying.

In summary it's pretty obvious that the matchmaking is examining decks and matchmaking is based on who runs which counter cards. There have been way too many reported examples for it to be dismissed anymore.

Here is a link to a pro player on YouTube who even comments how weird it is that he keeps seeing nothing but NG decks when he tries to play a Witcher deck on ranked:

 

[Draconifors]

So I am confident to say matchmaking is based on what is in the decks.

Except the REDs have stated the exact opposite in the past.

 

[Aris_Tarkus]

Over the last 20 games, my opponent breakdown is: NG:7, MO:5, NR:5, ST:2, SK:1

My decks have been MO and SK except for 1 match where I played SY. All of my MO/Witcher decks drew NG oponents. And in 1 of those I was hit with poison 12 times.

 

[Balgar-bard-slayer]

Except the REDs have stated the exact opposite in the past.

And you believe them? Why exactly?

I ask this because for it to be a valid counter argument would mean that I have to trust that games developers, especially those involved in Pay2Win online card games that contain loot boxes, have a history of behaving honestly and ethically.

And unfortunately they don't.

 

[Six-Sided-Prism]

Case in point whenever I attempt to play my NG Witcher deck, I will get a sequence of back to back games sometimes three in a row against NG control decks. And this happens every time I try for a run of games using that deck. if you attempt to play a deck that runs a large number of Witcher cards, 70%+ of the games will be against NG. And I strongly suspect this is because it runs the most natural counter cards (locks and poison packages) and that has happened at every rank including Pro.

I always thought these matchmaking complaints were a bit ridiculous, but I've actually noticed the exact same thing. I only ever encounter Nilfgaard when I play my Witcher deck, and if I switch to something else, it stops. It is a bit peculiar that this has happened to a bunch of people.

 

[Archbandit]

I woul

I think that is a separate thing. It's not about it being rigged against one player or another. But it's apparent that they are not


*snip*

I would tend to favour this argument based on my own experience, such as it is. In the last season I faced about 70% or more NG decks. I didn't mind: my win ratio went from about 5% to possibly even 50%, but it did get very boring. Sometimes I'd see an NG deck and just forfeit.

It may, also, be partly influenced by the level at which you play. As I'm crap and NG seems very popular at my level then there's likely to be a bias in there because of the numbers of each faction being played. Again, I found that the better players I played against (I can't think of any player who was especially bad, a few were very 'points on the board orientated', but they'll get out of that within a month) were either NR or MO, with the latter very often being inventive in their play. The worst deck I played was a removal deck that was excellent at culling my deck and quite frankly simply annoying. There are a lot of annoying mechanics in this game. A couple of times I faced SK, they were usually Gedyneth, ST were WOB/ME and the few SY players were usually good. I was playing the classic SY: "hit them all the time" deck. Also, probably, annoying as Hell to play against but better players and swarms had no problem with beating me usually.

 

[4RM3D]

So I am confident to say matchmaking is based on what is in the decks.

Your observations are not statistically significant because the sample size is too small and only focused on one particular (skill) level range. Thus you cannot say anything for sure, but you can still hold the opinion that matchmaking also looks at the deck archetype.

 

[Balgar-bard-slayer]

I acknowled what you are saying about sample size, but I suspect that is a hurdle that we will never overcome. I would have to ask what sample size would you require to start reaching anything like external validity of the claim?

Once you start factoring in variables like "the rank" on the assumption that they impact on the result we run into requiring yet more data.

However my personal experience is from playing over 200 games with this deck in the last month or so. I would say that is a large enough sample size from which conclusions can be drawn. I have played it from rank 25 to Pro, and I agree i don't have enough data at one individual rank, but the pattern across the ranks from those 200 or so games seemed consistent.

Don't get me wrong I'm an empiricist at heart, but at a minimum there is a common trend and consistency of various complaints about the game, that could be explained by the hypothesis that matchmaking is examining what is in each player's deck.

Take for example something that at first might seem unrelated: A thread started about a player complaining that Igni was ruining the game, because they had been running into it constantly at rank 2 and 3 (and it was clearly a deciding factor for them as to whether they won or lost). Whereas other players in the same thread at the same rank almost never encountered it.

This is easily explain by the above hypothesis as to why they were encountering Igni at such a high frequency.

It would be make sense for CDPR to matchmake in this manner too, whether they would be willing to admit to it or not, as it would achieve the goal of games being those of play and counter play, giving a feeling of skill through player interaction, rather than one side steam rolling the other because of an "unlucky" matchup. So given they pride themselves on the game being about skill (or at least the illusion of it) I'm highly doubtful that they would leave matchmaking up to chance.

Post automatically merged: Jun 30, 2020

Well this is going to get old real fast. Tonight's matchmaking brought to you by Gwent and Master Mirror.

I have literally played the against the same Wild Hunt Monster deck 6 times in a row.

 

[4RM3D]

I would have to ask what sample size would you require to start reaching anything like external validity of the claim?

That's part of the problem. No single user can ever provide statistical significant results. What you need is a large group of players working together and measuring the match-ups in the same way. That or having some sort of in-game tracker, which used to be possible in beta.

[...] Igni was ruining the game, because they had been running into it constantly at rank 2 and 3 (and it was clearly a deciding factor for them as to whether they won or lost). Whereas other players in the same thread at the same rank almost never encountered it.
This is easily explain by the above hypothesis as to why they were encountering Igni at such a high frequency.

No, not really. The user that didn't encounter Igni much could very well be running a deck where Igni doesn't get much or any value and a good opponent might mulligan Igni away, in that case. Incidentally, two users having opposite experiences doesn't tell us much without knowing all the circumstances. And even then it's only two players.

It would be make sense for CDPR to matchmake in this manner too [...]

No, it wouldn't. At least, not like that. The idea is to keep players engaged. This basically means that players need to win on a regular basis, but not constantly and vice versa. If CDPR wants to accommodate this, they shouldn't look at archetypes because it's far too complex. Instead, they only need to look at the average win (or loose) streak. Simply put, when players keep losing (on average) they should be matched against other players that keep losing.

I have literally played the against the same Wild Hunt Monster deck 6 times in a row.

The Wild Hunt archetype looks like fun and thus a lot of players want to experiment with it.

 

[Balgar-bard-slayer]

That's part of the problem. No single user can ever provide statistical significant results. What you need is a large group of players working together and measuring the match-ups in the same way. That or having some sort of in-game tracker, which used to be possible in beta.

I agree, but there are other research methodologies that exist. Large empirical studies in laboratory conditions are one way of drawing conclusions. But as you say yourself it's probably not feasible so maybe there can be a mixed method approach.

No, not really. The user that didn't encounter Igni much could very well be running a deck where Igni doesn't get much or any value and a good opponent might mulligan Igni away, in that case. Incidentally, two users having opposite experiences doesn't tell us much without knowing all the circumstances. And even then it's only two players.

You are reaching here. You seem to be building a scenario that doesn't reflect the content of what I posted. I know from reading the thread (as admin you probably read it too) it was more than two users, and at least one of them was running a deck at the same rank that row stacked tall units, but who was not encountering igni. But I can't comment on the make-up of the decks they were facing.

No, it wouldn't. At least, not like that. The idea is to keep players engaged. This basically means that players need to win on a regular basis, but not constantly and vice versa. If CDPR wants to accommodate this, they shouldn't look at archetypes because it's far too complex. Instead, they only need to look at the average win (or loose) streak. Simply put, when players keep losing (on average) they should be matched against other players that keep losing.

I agree that I would like matchmaking to worked like the ideal you describe. And in a mode like Arena it probably does.

But we might disagree on how players are kept engaged. For me the objective is to encourage users to buy new cards. A way of doing this would be by designing a system that sets up games with a back and forth, whereby if you lose to the same superior counterplay cards you would need to get cards to counteract the patterns you are being shown. Thereby boosting sales.

That is not to say that they could not set things to give a minimum win rate to a losing player. But it's well known (or at least its there in the research) that you don't achieve player retention by simply giving them opportunities to win regularly. So whilst I like your theory, I don't think it is based on what research into the field tells us nor how large games developers and publishers seem to view P2W Vs F2P games (from my experience working with them anyway).

The Wild Hunt archetype looks like fun and thus a lot of players want to experiment with it.

MO decks playing the same new cards 6 times in a row though? Yeah...you haven't convinced me.

Remember as well that it was only a minority of the player base that was able to actually access Master Mirror cards when I posted this due to the shop being locked down within the first hour of launch. And yet they made up not only the majority, but in fact 100% of the opponents I faced in the hours after launch.

 

[Archbandit]

In the casual part this time I've faced mostly ST, but not by much. That, I guess, is likely to the new challenge and my having chosen SK. So far this set of games has been far less skewed other than I'm playing, on the whole. either better or even much better players. And the dog/Roach combos are a bit boring.