How smart has the Matchmaking-System already become?

[von_Cronberg 258]

There are plenty of comments complaining that the MMS would not work as fine as the respective commenters would wish, in particular because it would not produce the always "balanced battles" they desire. In my opinion, these complaints generally are baseless, as two central problems for composing such always "balanced battles" are impossible to overcome: First the complexity problem of evaluating every player's effectiveness in a given context of map and team composition, let alone considering their current mood. Second and most important the problem of quitters, which desert their comrades at the start of a battle, thus obliterating any consideration the MMS algorithm had in composing the teams and leaving it struggling to fill the ranks with whoever becomes available for a battle assignment (and as far as I can see, the MMS does an amazing job in this, given the huge difficulties). Besides, I think that having to quickly adapt to very diverse circumstances and odds in every battle is not a problem but an enjoyable challenge, in this game as in life.

However, there are many ideas how to make the MMS smart and smarter. As I am not sure what has been already implemented, here are some points in the form of questions, for anyone to comment, and if possible also for admins/mods to provide definite answers.

(1) Does the MMS finally avoid having friends fight each other? My major personal complaint with the MMS used to be that it pits me against friends of mine, causing me to immediately leave such battles. In the early days of the MMS, this happened frequently. However, these days, it has become a very rare occurrence. When I meet friends in battle now, they are almost always in my own team. Is it correct to assume that the MMS now generally gives friendship relations between players the high priority they deserve?

(2) Why is there still no compensation for late assignments? A major complaint is that players often are assigned to battles in their late stages, an issue that is caused by the aforementioned problem of quitters. Such late assignments frequently are an annoyance for a variety of reasons. There have been many suggestions on how to mitigate such annoyance, like starting with an already charged overdrive. Does anyone know why this easy step to produce fairness and avoid anger among players is not simply implemented?

(3) Why does the MMS reassign players to battles they had deliberately left? I do only rarely quit battles the MMS assigns me to, certainly less than ten percent. However, when I do, I apparently have reasons to do so, and I apparently do not want to play that battle. However, when I afterwards then click for another assignment, I am quite often reassigned to the very battle I just left. One would think that avoiding such bizarre reassignment to the same battle would be the first and simplest thing to do. Why is it not yet done?

(4) Why are there no steps taken to encourage pride in being a finisher? The problem of quitters being the issue at the root of all problems with the MMS, and while I do understand that Tanki does not want to outright punish players for abandoning their comrades in a battle, and there are good reasons to not do such punishment, there would be simple and friendly ways of nudging players to finish, such as calculating every player's finishing quota and publishing it on their profile page. Why is this simple thing not yet done? 

Edit: Dear mods, I use this forum long enough to know that some of you may, upon reading this, feel the violent urge to "merge" this post into some other thread about the MMS, effectively burying it. I would like to kindly ask you to resist this violent urge. I never asked you such a thing before and will never ask you such a thing again, I never criticised the (in my opinion overactive) frequent merging of threads in this forum and never will do so in the future, but I ask you now to let this posting and thread live as it is.

Edited June 20, 2020 by von_Cronberg

[Azonyx 124]

I find MMS is actually very smart, in the same way a slot machine is configured to keep unsuspecting players coming back for the occasional jackpot. You lose 3 / 4 games but the one time you do dominate the game, you get a nice payout and a rush to play more. This is the reason why most of the time there are mega-buyers + super drone druggers on one team and mults / hopelessly undergeared f2ps on the other. Sometimes there will be that one tryhard carrier too, but the battle is already decided by the time the buyers all spawn with Crisis / Booster / Defender at 06:45.

So... Improve MMS? What is there to improve? So long as the buyers get their bullying fix, there is no lost profit. And as long as there is a gray technical area where devs can simply say, "just your luck," "not enough players of your kind online now," or god forbid "sometimes you have to lose too," there will be no compromise.

Keep in mind too that the devs are swamped with HTML5 development, Flash maintenance, Ares / Tesla and a myriad other shiny things nobody asked for, all with the goal of buyer retention at this point in the game's lifecycle. Before MMS gets any improvement (if at all, considering how well it works right now for the buyer$), all these updates have got to be finished first with an accompanying explosion in new player numbers. Sounds like a dream? Exactly.

[von_Cronberg 258]

10 minutes ago, Azonyx said:

You lose 3 / 4 games but the one time you do dominate the game, you get a nice payout and a rush to play more.

Statistically, every player's teams lose and win 50 percent of the battles they participate in. My personal feeling is that the wins are more. And even if not, what exactly is the problem of your team losing? Crystal payouts are almost the same in the MMS (other than in PRO battles), stars are even arguably easier to win on the losing side.

 

17 minutes ago, Azonyx said:

This is the reason why most of the time there are mega-buyers + super drone druggers on one team and mults / hopelessly undergeared f2ps on the other.

This is simply not true, as far as the MMS is concerned.

If anything, there is a quitters issue among some strong but personally pathetic players, too. Just yesterday I was assigned to a battle, clicked the teams' list, and was excited to see two of the strongest players I know, each of them in a different team. I expected an epic battle. However, within seconds, the dude in the other team quit. This is not the fault of the MMS, this is the fault of a misguided (and in my view, psychologically pathologic) attitude among some players, and which I honestly also read between the lines of your comment, that only battles where your team can "dominate the game" would be the real fun. This attitude is not uncommon among both some weaker as well as stronger players, and this attitude, and not the MMS, is the problem.

You can also observe this in PRO battles, where such players seek to flock to the team which they think will "dominate the game", creating a tendency towards a reinforcing cycle of the stronger team becoming ever stronger in the team composition process (and this is exactly the reason why I appreciate the existence of the MMS, which mitigates this problem). By the way, I spent money for a Heat Immunity augment for my Hunter, to be used with Vulcan and Incendiary Band augment, for one purpose: I use this gear, which is brutally overpowered in no-drug battles, to join the "dominated" side in PRO battles, seeking and frequently succeeding in turning the battle and frustrating those who had sought to "dominate the game". And I have fun with it.

 

45 minutes ago, Azonyx said:

Sometimes there will be that one tryhard carrier too, but the battle is already decided by the time the buyers all spawn with Crisis / Booster / Defender at 06:45.

I do for a long time mostly use a drone (Defender or less often a Saboteur, Booster or Driver) in MMS battles, and I assume that I fall under your definition of "buyer". Let me assure you, from the perspective of the dark side in your world, that I do usually find other "buyers" on both sides in MMS battles. And that I do often expierence skilled "non-buyer" players as the most effective adversaries. Half or at best a little less than half of the MMS battles I play, my team loses. I think that your view may be biased by the desire to "dominate the game" every time, and frustration when you do not.

 

57 minutes ago, Azonyx said:

area where devs can simply say, "just your luck," "not enough players of your kind online now," or god forbid "sometimes you have to lose too,"

Did you ever consider that all of this might just actually be true?

[Venem 615]

Whoa, this is a decently researched topic, and very well thought. @von_Cronberg, I must compliment the effort you have put in... I won't comment on the MMS for two reasons:

• It's still not completely developed, and I personally have seen it's negative side very often, though I admit that mostly quitters are the cause of frustration.
• I haven't researched this system much, and that might result in a controversial opinion.

 

[Picassoo 191]

Regardless how smart it is, the system we had in 2015 wass better for such a game, and would be more suitable now especially when the player base is so small (I wonder why )

[Gauss-Hornet 489]

Well, Hazel himself had said (in many old vlogs) that the new Matchmaking system was designed to make battles "exciting" instead of "boring and balanced" by having a player equally be on both sides of frequent blowout battles. 

See, I'm not completely against this ideology. Some blowouts would be fun, but I want their percentage to be less than 30%, not 99% as it currently stands. Also, I want each and every player to be placed, around 50% of the time, on a potential winning team, instead of continuously being paired with mults that don't contribute anything for the team.

 

If the game is going to have P2W, fine. But have the buyers split equally amongst the teams, instead of racking them up against a team of mults.

Edited June 20, 2020 by Gauss-Hornet

[DieselPlatinum 5,121]

1) minor problem.

2) agreed.

3) agreed.

4) so basically yet another useless stat in profiles so you can bash on the player for when you're losing an argument. Gotcha.

1 hour ago, von_Cronberg said:

If anything, there is a quitters issue among some strong but personally pathetic players, too. Just yesterday I was assigned to a battle, clicked the teams' list, and was excited to see two of the strongest players I know, each of them in a different team. I expected an epic battle. However, within seconds, the dude in the other team quit. This is not the fault of the MMS, this is the fault of a misguided (and in my view, psychologically pathologic) attitude among some players, and which I honestly also read between the lines of your comment, that only battles where your team can "dominate the game" would be the real fun. This attitude is not uncommon among both some weaker as well as stronger players, and this attitude, and not the MMS, is the problem.

If it's a battle where the player could dominate in, one reason they'd leave that battle is if the damage numbers aren't showing, which would indicate that they're lagging. And sometimes the only way to fix that is to leave the battle.

Another reason might be that they just lost internet connection for the time being.

[von_Cronberg 258]

1 hour ago, Picassoo said:

Regardless how smart it is, the system we had in 2015 wass better for such a game

50 minutes ago, Gauss-Hornet said:

Hazel himself had said (in many old vlogs) that the new Matchmaking system was designed to make battles "exciting" instead of "boring and balanced

In the stage pre-MMS, battles were everything, but not balanced. Those who actually played in those days might recall that the most important skills in winning battles (and being in the winning team was more important back then, when compared to today's MMS, because the battle fund system gave a huge advantage to the winning side, like in PRO battles then and now, but unlike today's MMS) were (a) having the name recognition and judgement to watch team composition processes and know which emerging team will most likely win, and (b) persistence in clicking the Join button for that team in quick succession for minutes, so that when some "noob or social justice warrior" joined the other team, your click for the presumable winners would be the first and you get in. I prefer the MMS over those days. And if you like choosing teams, play PRO battles (which I myself do about half of my Tanki time).

[Maf 32,300]

I agree with pretty much all your points, some of them very strongly.

First of all, thanks for appreciating the difficulty in establishing balance. The reality is that our world is extremely chaotic, so even in identical circumstances with all variables remaining the same, battles can end with drastically different results due to pure chance. One overdrive used a second too early or too late could give an momentary disadvantage that cascades into complete domination by the other team. Even if we assume that everyone has a fully upgraded garage and is equally skilled, the result of the battle will heavily depend on the equipment and protections the players will choose. It's impossible to predict.

That being said, I do sometimes think that the MM balancing system is either not working properly, or intentionally causing disbalance. I'm talking about cases with one team having myself and two guys with Veteran paint and top-100 efficiency, while the other team having a bunch of non-legends. Such games end very quickly and predictably, and it really makes you wonder how MM matched the players.

5 hours ago, von_Cronberg said:

(3) Why does the MMS reassign players to battles they had deliberately left?

Definitely agree with this. Devs must implement either a way to see in advance the kind of battle (i.e. map and team) you're being put into, and/or not put you back into the same game once you leave.

5 hours ago, von_Cronberg said:

(2) Why is there still no compensation for late assignments?

This would be a cool thing to have. My personal suggestion would be to have a new kind of container with basic rewards (supplies and crystals) that would have a chance of being rewarded if you stay till the end of a battle on the losing side. The container would only be given to active players on the losing team, and the longer you stayed in the battle, the higher your chance of getting this container.

Container would be similar to the one given for free in the mobile version, but perhaps with about 5x the amount of rewards (around 600 crystals and 20 supplies total).

[Gauss-Hornet 489]

13 minutes ago, von_Cronberg said:

 

I am not saying that the previous system was better, I just pointed out how the MMS has purposefully been rigged to provide unbalanced battles. 

[JohnMatrix 18]

Agree it's difficult to come up with an ideal algorithm.

1) May be an unpopular opinion, but I think friends lists can get abused & aggravates problems of power concentration.

4) Not a fan of the current profile page setup. Just makes people prone to judge others based of it.

Sometimes it may be better for a battle to just blow out & get over quick.

[wolverine848 13,097]

4 minutes ago, JohnMatrix said:

Agree it's difficult to come up with an ideal algorithm.

1) May be an unpopular opinion, but I think friends lists can get abused & aggravates problems of power concentration.

4) Not a fan of the current profile page setup. Just makes people prone to judge others based of it.

Sometimes it may be better for a battle to just blow out & get over quick.

1) Agree - there definitely are Groups vs non-groups.  Group tags would allow transparency but they wont implement

4) Agree.  The number of times players bring up k/d for comparison purposes is laughable.  They have no clue...

[DieselPlatinum 5,121]

40 minutes ago, wolverine848 said:

1) Agree - there definitely are Groups vs non-groups.  Group tags would allow transparency but they wont implement

4) Agree.  The number of times players bring up k/d for comparison purposes is laughable.  They have no clue...

Hmm, what if there are two or more groups in a team?

[Azonyx 124]

3 hours ago, von_Cronberg said:

Statistically, every player's teams lose and win 50 percent of the battles they participate in. My personal feeling is that the wins are more. 

Theoretically speaking, yes. Being a PvP game, someone has to lose in order for someone to win - it's a zero-sum equation. But since we're talking statistics here, the reality is that distribution of wins and losses won't be as smooth and linear as win > loss > win > loss > [...]. Especially when equipment plays a massive role in tilting the balance. From the perspective of a data analyst, it will look like typical random clusters. But as far as players are concerned, it's a game design problem when it happens too frequently to the point of annoyance.

3 hours ago, von_Cronberg said:

And even if not, what exactly is the problem of your team losing? Crystal payouts are almost the same in the MMS (other than in PRO battles), stars are even arguably easier to win on the losing side.

I will tell you the problem with a losing team. First look at this:

 

 (Hi there congospider)

You should be able to tell the Red Team lost due to the star distribution. Now look at the score differences and the battle funds awarded.

The more scores you get, the higher the payout. How are you going to score? By destroying, capping, or doing assist actions - a.k.a. "dominating." When you and your teammates dominate, it also has a nice side effect of making your team win.

Now here is a question. How high of a score can you get in a match where you and your teammates are getting spawn-killed every 10-20 seconds by Vulcan-Hornet-Defenders or EMP Gauss-Boosters? Still think payouts will be the same for you and your poor buddies that only managed to get <200 score by the time the dust settles? How will you feel when you can't do anything meaningful for 3 games in a row, let alone enjoy TO as a "fun gaming experience"?

As for stars, you can earn them all you want in the losing team. But then you have to ask yourself in the long term - would I rather earn 6 stars trying hard and getting bent the whole match, or earn 6 stars and more crystals minus the frustration?

3 hours ago, von_Cronberg said:

If anything, there is a quitters issue among some strong but personally pathetic players

[...]

This is not the fault of the MMS, this is the fault of a misguided (and in my view, psychologically pathologic) attitude among some players, and which I honestly also read between the lines of your comment, that only battles where your team can "dominate the game" would be the real fun. This attitude is not uncommon among both some weaker as well as stronger players, and this attitude, and not the MMS, is the problem.

So there is a quitters issue. Why do people quit MMS games in the first place?

To you, it's cowardly way to cycle into rigged matches to "dominate". What I see is a refusal to play by Tanki's cannon-fodder system of placing an obviously stacked death-squad against moving score tanks in the name of buyer retention. When team composition is already broken at match start, players can't be blamed 100% for turning down a sure beating. Who in their right mind even stays in a blowout match and says they enjoyed playing in intervals of <4 seconds inbetween respawns?

3 hours ago, von_Cronberg said:

Half or at best a little less than half of the MMS battles I play, my team loses. I think that your view may be biased by the desire to "dominate the game" every time, and frustration when you do not.

Here's the thing. MMS is designed to put players into unwinnable fights and guaranteed victories regardless of how geared or skilled they are, making it a pure gamble whether you get to do well or not. Granted in other games there are some battles that just can't be turned around, but here in TO the one-sidedness just gets plain disgusting and downright un-fun. You can try P2W'ing your way and it does help improve your chances... until you get put by yourself in a dysfunctional team against your same kind.

At the end of the day, you are entitled to believe whatever opinion you have of my attitude to the game. Although, you have to wonder if it's really just my personal bias when time and again we see people echoing the same sentiments against the travesty that is the slot machine MMS.

Edited June 20, 2020 by Azonyx

[von_Cronberg 258]

1 hour ago, Maf said:

That being said, I do sometimes think that the MM balancing system is either not working properly, or intentionally causing disbalance. I'm talking about cases with one team having myself and two guys with Veteran paint and top-100 efficiency, while the other team having a bunch of non-legends. Such games end very quickly and predictably, and it really makes you wonder how MM matched the players.

My impression is that this issue, which I see myself (being in one team with some esports champions, although I might be, while certainly not their equal, at least a more resilent adversary to them than anyone in the other team), is caused by the MMS being overly obsessed with pitching groups against groups. So when some esports champions play as a group of two or three, as they frequently do, the MMS is not primarily concerned with putting up the strongest players available against them, but rather primarily concerned with putting up another group available, which will in many or most cases be some group of random friends. Only then are single players distributed, and my impression is that in considering the further team composition, the different strength of the groups is not sufficiently taken into account.

So my suggestion to everyone is: Only play in groups if you are prepared to be considered, by the MMS, as a match for the best of the best. I do not often play in groups, but when I do, I always have this in mind.

[Maf 32,300]

3 minutes ago, von_Cronberg said:

MMS being overly obsessed with pitching groups against groups

Yes, I was going to add that in my message, but didn't. The guys were obviously in a group, so MM probably doesn't even consider their power and simply puts another group in the enemy team. Not the right way to do it, tbh.

[wolverine848 13,097]

1 hour ago, DieselPlatinum said:

Hmm, what if there are two or more groups in a team?

You can have different clan members on same team and each clan-tag is displayed.

But that wasn't my point.  It just has to be a generic tag "Group".  If you joined the battle as part of a group then, instead of Clan-tag showing it would be ["Group"] .

[von_Cronberg 258]

31 minutes ago, Maf said:

The guys were obviously in a group, so MM probably doesn't even consider their power and simply puts another group in the enemy team. Not the right way to do it, tbh.

Fully agree. However, I would like to point out that the forum community here cannot put the blame for this issue on the game developers. I remember that after the implementation of the MMS, this forum screamed murder and for weeks ranted against single players being pitched versus groups and demanded that groups shall always have to face other groups, so the developers only heeded the vox populi in this. I myself never understood those rants, a group of random friends are hardly more of an adversary than the same number of random single players.

[DieselPlatinum 5,121]

1 hour ago, von_Cronberg said:

Fully agree. However, I would like to point out that the forum community here cannot put the blame for this issue on the game developers. I remember that after the implementation of the MMS, this forum screamed murder and for weeks ranted against single players being pitched versus groups and demanded that groups shall always have to face other groups, so the developers only heeded the vox populi in this. I myself never understood those rants, a group of random friends are hardly more of an adversary than the same number of random single players.

Groups are far more likely to formulate coordinated attacks than random players are.

[Maf 32,300]

8 hours ago, DieselPlatinum said:

Groups are far more likely to formulate coordinated attacks than random players are.

Are they though? My experience playing in a group consists of doing what I always do (playing alone) but occasionally also getting distracted by chat. I don't think it's reasonable to assume that players in a group are significantly more coordinated than random players.

[ghost-guns 1,486]

It's garbage. There is nothing smart about it.

[JohnMatrix 18]

5 hours ago, Maf said:

Are they though? My experience playing in a group consists of doing what I always do (playing alone) but occasionally also getting distracted by chat. I don't think it's reasonable to assume that players in a group are significantly more coordinated than random players.

Let's just ignore soft power & popularity.

[NooNooHead 121]

14 hours ago, DieselPlatinum said:

Groups are far more likely to formulate coordinated attacks than random players are.

A week ago, I keep bumping into alklines and his group with my different accounts in different battle modes. I greet him/her and then bye bye. He is polite. Of the matches I stayed, I have yet to see his team lose or play a close match. A few days ago, I bumped into 2 players on the other side with the spectrum protection module. Clearly, another team but I stayed and we lost 30-5 in CP. That is with everyone on my side going to the points, no campers.

Edited June 21, 2020 by NooNooHead

[wolverine848 13,097]

6 hours ago, Maf said:

Are they though? My experience playing in a group consists of doing what I always do (playing alone) but occasionally also getting distracted by chat. I don't think it's reasonable to assume that players in a group are significantly more coordinated than random players.

Then you're doing it wrong.  Communication is huge in war.  Why the heck are u not communicating in your group?

At the very least a Group should have a plan - with responsibilities.  Else - what's the point?

[Maf 32,300]

1 hour ago, wolverine848 said:

Why the heck are u not communicating in your group?

I get invited and I join, because why the hell not? I don't know how other people do it, but I suspect that the majority is the same.