EXPLAINED: The REAL reason for the leaver penalty in QP

The real reason for the quickplay leaver penalty is not to ‘make casual more competitive’ - it is part of what amounts to basically a psychological experiment to test the limits of how much match-manipulation the players are willing to accept. The goal is to better determine where people set the bar for acceptable match quality. And ultimately to find the optimal balance between queue times and match quality.

The goal of the matchmaker and match-balancer in quickplay should be to provide matches that are fun and appropriately challenging for all participants, regardless of skill level, solo/group queue, or platform (PC vs console.) If the player is over- or under-challenged, they will not have fun. A player who is not having fun will over time become dissatisfied, and likely not stick with the game or be inclined to spend any money. Even worse, a player who is repeatedly having a bad time will just close the game and potentially even uninstall.

The rate of leaving the game due to dissatisfaction is called ‘churn’ and the company has incentive to reduce churn as much as possible. The company wants to keep people interested in playing so that they ultimately spend more money. They constantly fine tune and adjust their sales approach (skin prices / bundle structures) to optimize revenue; and constantly influence player-enjoyment via matchmaking to minimize churn.

It can be difficult to quantify churn in casual play, because users have many reasons for stopping their play session that may or may not be related to dissatisfaction. Without a leaver penalty, players are free to leave the match at any time for any reason. Without a leaver penalty, leaving a game provides no data to the company about why the player decided to abandon a given match. The system cannot distinguish between someone leaving because they were upset about the match or because their Grandma just fell down the stairs and needs an ambulance.

Putting leaver penalties in place incentivizes players to stay until the end of the match, which they will in most cases do if they are having fun or are at least not totally miserable (or having an emergency.) From the perspective of the data analysis team, the new ‘leave game’ button now essentially amounts to an “I’m not willing to put up with this” button. People will only click the button when they are totally frustrated and outraged. The distinction between “player is having fun” and “player is not having fun” becomes more clear and easy to detect. They use the leaver rate to figure out where people draw the line with the matchmaking, and then make adjustments so there is less churn.

The adjustments they are currently working on right now seem to be matchmaker and match-balancer (DDA) related. I don’t think it’s a coincidence that we are seeing lots of forum posts about both ‘broken matchmaker’ and ‘tons of leavers’ - The leavers are a consequence of the poor adjustments to the matchmaking and poor tuning to match-balancing that, together, result in unfun matches. Because players are completely blind to the parameters which are utilized during matchmaking and fully unaware when the matches are being actively balanced, they assume (placing misguided trust into the company) that the match is assembled in a way that provides a fair chance of an equal outcome. The reality in fact is that the match is highly structured by the system to result in a specific and bespoke psychological outcome for each and every player individually.

Let me elaborate a bit on the Matchmaker and Match-balancer mechanics I mentioned.

The arrangement of players into teams is what most players consider to be the job of the matchmaker. This is a concept most players are probably familiar with at a basic level. Although as I already mentioned, they probably naively assume that the matchmaker’s goal is to create two equally matched teams based on skill. As described in the patents, the matchmaker will place ‘marquis’ players on a team with new players, so that the new players see all of the cosmetics and are motivated to spend their money. The matchmaker is putting more emphasis on product placement than it is on fair matchmaking. The patents are easily searchable on Google if you’re interested in a more thorough explanation or some further insight into the capabilities and techniques at work.

The system that the player base is probably much less aware of is the match-balancing system. Because the quickplay lobbies are crossplay enabled, the company needs to have a method of putting both controller players and mouse&keyboard players on a level playing field within the same match. This is done in many potential ways.

  • One technique involves adjusting the hit registration tolerance. For players on controller who have an inherent challenge in aiming, (or lower skilled players who have bad aim statistics) they are given a large tolerance to what registers as a hit. For players on PC who have good aim statistics, they are given a low tolerance for what registers as a hit.

  • Another technique involves what I will call ‘reverse aim assist’ which, as the name implies, works like aim assist in reverse, making it harder to aim at and focus fire your opponent. If you think of regular aim assist as being like a magnet that draws your crosshair in when you get close enough, reverse aim assist is like the opposition of two similar magnet poles; the crosshair is repelled more strongly when it is closer to the target. The end result is a bit more survivability for the disadvantaged player. They are not unkillable, but the skilled player has more of a challenge to secure the elim. This phenomenon is most evident when studying the crosshair movement and hit registration of weapons that require aim-tracking. This is what is responsible for what many players have described as the sensation of your aim feeling ‘floaty.’

These are just a few of the techniques available to the match balancer to create an advantage to one team or another (or to ostensibly create a fair matchup.) These match balancing parameters can be adjusted on the fly according to individual player performance in realtime. This allows for the system to balance matches regardless if you’re playing your best or having a bad day. Again, there are patents which can be found quite easily that detail these methods.

Together, the matchmaking and match-balancing systems amount to a manipulation engine that transparently (in theory) adjusts the matches in ways that maximize player engagement. This process is known in the industry as EOMM; Engagement Optimized Matchmaking. Because the player is not informed that any of this balancing is taking place, the company is basically gaslighting the players into attributing the outcome of the match to the performance of the team. The players are oblivious to these methods and can only be left to assume they just had a good or bad match for one of many reasons. Usually they just blame it on the supports or say ‘tank diff’ and move on. They don’t go back and watch the replay and realize that their team was basically themselves and 4 bots. The company is taking advantage of people’s aversion (or inability) of thinking too deeply about why things are the way they are.

The reason for the leaver penalty is to provide just enough friction to the leaving process that the players will sit through ‘tolerable’ bad matches. The penalty is not severe enough to prevent a player from leaving a truly bad match. It essentially amounts to a yes/no test to see how heavily-manipulated of a match the system can arrange without generating churn that cuts into the sales numbers. They want to determine exactly how far they can take the manipulation. Or a more optimistic take would be that they want to use the data to continue optimizing their matchmaking and match-balancing in order to produce great and fun games for everyone, regardless of skill level or state of mind.

The optimist in me dares to hope for the latter. But if that’s the case, the truth is they have a long way to go.

13 Likes

Yeah it’s something like that. Really it’s because they know all the other measures they’re planning on putting in to coerce players won’t work unless they force you to stay in games. Just look at Mauga.

5 Likes

or its been a complaint about the game for like 6 years…

29 Likes

They don’t need to force us to stay in games. They just need us to have a slight resistance to leaving. That way the ‘leave’ is more meaningful from a data science perspective. It serves as an indicator that the match is unacceptable.

I’m not sure what you mean, but thanks for chiming in.

1 Like

How do you propose they do that in a mode meant for quick access to short games where anyone of any skill level can play together without drastically increasing queue times?

The only thought I have is to make group stacks try to match with similar groul stscks (got a diamond and a silver together, it tries to match you into a game where the enemy has another duo with a similar rating distribution). The issue is that not only will this still increase queue times (I would accept this myself but know others won’t) but this won’t fix stomps either. Stomps can happen even with teams that are in theory equal to one another.

To keep it brief, I too would love better matched games. But I think it will take as a sacrifice (queue times) that many people do not want to make for QP. The possible parameters with who can play wither each other are far to wide to balance otherwise. And I think that’s okay to an extent, I like being able to play with my friends in a more casual setting when I literally cannot even queue with them for comp lol. I’d take longer queue times to try and find brtter matches but would not be shocked if the complaints persisted either.

People complained frequently about match making in OW1 as well is what they are saying.

make the game better ppl will want to play it?, just a thought… nobody wants to play stomp after stomp bud. with blatantly broken heroes and on bad maps with bad game modes… make the game better ppl will want to stay, that easy.

3 Likes

Players get to decide subjectively what is overly/underly challenging, objectively they have no choice, nor should they for that matter. While perception should exist on a gradient of stomp(friendly)/underly/fair(f)/fair(enemy)/overly/stomp(e) more often than not the perception often lumps together all of the favorable outcome matches as “fair” while looking at all the unfavorable outcome matches as “overly challenging/stomps” even if they were fair.

Players just don’t know how to take a loss, as well as hold their wins in such high esteem that it makes it harder to accept a loss. I’m speaking on my behalf too, I was in that mindset at one point.

My games are consistently fair, with stomps (from my perception of a stomp) being incredibly rare. I seldom ever feel cheated (not cheated against) out of a victory. Backfills turn around the flow of gameplay, the team just happens to pop off despite having a rough go of it up to that point, etc. And sometimes all that hard work just doesn’t pay off, there’s gotta be a winner and a loser after all.

It’s also entirely possible that the player’s experiencing such matches have actually skewed their experience by quitting so often, even if it’s not an intentional from a design standpoint.

2 Likes

I explain it in my post. It’s also explained in the patent documents. They adjust hit registration tolerance to make it easier to land shots for lower skilled players. Its basically as if they were shooting at a larger target. They implement several other balancing mechanics as well. It’s not just a basic matchup where they put similarly skilled players in the same lobby.

The end result is that they can reduce queue times by shipping ‘lower quality’ matches. Where the skill disparity between players is offset by the individual performance adjustments made to each player by the match balancer.

1 Like

No, it’s much more simple:

By joining a match, you’ve committed to a brief, average 7-10 minute game. If you leave, it’s entirely understandable.

If you make it a habit, you will get increasing penalties until you either modify your self-centered behavior, or go find a single player game.

11 Likes

Thanks for not reading my post, or the title. And contributing nothing of value to this discussion.

7 Likes

You laid out your flawed thesis in the first paragraph, everything else is unnecessary bloviating.

3 Likes

That doesn’t equate to a fair or balanced match though. While it being easier to land shots can absolutely help, that doesn’t fix bad habits or mistakes lower skilled players will make compared to better players. Blizz literally cannot fix bad skill habbits, and they make a huge difference.

You also run into the issue of players finding those very fixes you’re mentioning as frustrating. People already think matches are rigged, and will be increasingly upset if you add things like anti aim assist.

1 Like

Which is exactly why there are so many complaints of poor matchmaking, and so many leavers.

Yes, and the unfortunate part is that they have already implemented these fixes. They ARE frustrating. And they DO amount to rigging. I understand the reason they have implemented these adjustments, but in my opinion they need further refinement, and to be made more clear to the user that they are in play.

1 Like

What are you, some kind of sensible person or something? This is the OW forums, no room for that kinda stuff here.

2 Likes

Just apply occams razor. The simpler explanation is people disliked the mass leavers and backfill so they added penalties to keep people around in more matches.

As far as the matchmaking goes its basically impossible to guess how any individual player is going to play in any given round. MMR is just an average.

Its also nearly impossible to tell which players are going to work well together with 4 other random players with zero communication. If you have a team in quickplay thats able to effectively work towards the same short term goals (IE: We all want to kill the enemy Moira super bad) theyre going to do better then the scattershot team thats spreading fire between every enemy that they personally hate.

I’m also of the opinion that 5v5 may just be more prone to steamrolling too. Higher individual impact is cool but also means, in a team based shooter, that losing a player will have a larger impact as well. So the team that can co-ordinate better with no communication towards killing enemy tank/supports will have a pretty large advantage.

So these perfect matches people want will never happen.

4 Likes

Except blizz claims to be able to determine the outcome of a match with a certainty of 90%, just based on performance history alone.

OK so what about all the shills on this comment thread and others saying that “matches are perfectly balanced” … “I rarely if ever see stomps” … “my match quality is excellent”

For the record, I don’t want “perfect matches” - I want fun and reasonably balanced matches. I don’t want to win every time. I think a good start would be making it so I had fewer matches where I’m stuck with 3 crossplay kiddies on my team, just running in and dying off cooldown.

2 Likes

Okay, but then those aren’t really great fixes then to the question I asked earlier. What I’m getting at is you both want the games to be more equal but allow for the same parameters we currently have in QP, which is simply not feasible.

So I’ll ask again, how do you propose they create more fair matches without significantly increasing queue times?

To your overall premise of the thread, I think it’s much more sensible to think that people don’t like it when others leave and ruin match integrity and this was Blizz’s way of trying to address that. I’d rather have incentives for staying rather than punishments, but I don’t buy into the idea that it’s really any deeper than that.

you would think…but then they make their focus money over quality…

anyway, i was just saying to OP that we dont have to overthink this…leavers have been a complain since forever…maybe theyre finally listening to the complaints…

2 Likes

I know for damn sure I would. Their already implementing a system into comp where there’s accounting for solo players and groups of players right? That big change for season 9?

Put it in QP too, I say. To hell with it, if we’re going down this road where QP needs to be more serious and less casual with lockout penalties and the like, than the least they can do is make solo queuing even a tiny bit bearable for those of us who want to do it. I’m well aware it’s a team game, but plenty of people go it alone too.

1 Like

No one who complains about the matchmaker is willing to even acknowledge this point because they know there isn’t an answer.

Admitting that makes them realize how unreasonable their expectations are for quickplay and they can’t admit that.

1 Like