MM balancing - imcompetency or favouring russian players?

[TinBoxGunner 140]

 

Just had a battle with my alt acc of General.  MM Balancing sucked once more.  Enemy team had 4 legends, out team had none.  Enemy teams GS average was about 8800, our was about 6800.  Devs are either imcompetent or somehow favouring russian players with MM balancing.  Which way is it?

[von_Cronberg 258]

I think it is neither. To be fair, I am pretty impressed with the algorithm organising MM battles, and I play a lot of them. The issue of the MMS is a large number of players leaving battles they were assigned to, with the MMS then struggling to fill the ranks with players who click to join a battle in that second. The quitters are the problem.

[Maf 32,305]

What does this have to do with Russian players? You have no way to be sure where the enemy players are from.

Also, GS is just one of the factors taken into account in MM balancing. A player could have a low GS, but enough skills to be equivalent to one with higher GS, so their overall K/D, W/L ratio and other stats will affect their strength when being put into teams. And sometimes you just get unlucky and lose even while having an overall gear advantage.

[coconuttree 519]

51 minutes ago, von_Cronberg said:

 

I disagree. Many of the matches I play are lopsided from 7:00

Edited May 21, 2020 by coconuttree

[von_Cronberg 258]

I have played well over ten thousand battles on Tanki. Some where wins for my team and some were losses, in some I was more pleased with my personal performance and enjoyment and in some I was less pleased with it. In a number of them we massacred the enemy, and in a number of them we were massacred. But never, not once, not a single time in those well over ten thousand battles, not once was my team massacred as long as every player stayed in the game and persistently sought to hold their position.

[799169 48]

19 minutes ago, von_Cronberg said:

I have played well over ten thousand battles on Tanki. Some where wins for my team and some were losses, in some I was more pleased with my personal performance and enjoyment and in some I was less pleased with it. In a number of them we massacred the enemy, and in a number of them we were massacred. But never, not once, not a single time in those well over ten thousand battles, not once was my team massacred as long as every player stayed in the game and persistently sought to hold their position.

You must be very lucky. About 50% of battles I play are complete blowouts from the beginning. As I spawn, there are are 4 enemies in our base on full supplies.

Meanwhile, my team has nobody who has activated even a single supply. And my team is full of wasp twins or hornet shafts who never use supplies, and end up with 0 kills 13 deaths...

Some of the strong players in other teams should have been swapped with the weaker players in my teams, so that both teams have equal number of good players.

Its not about GS so much, just that there team splitting algorithm is so incompetent at making balanced battles.

Once it has selected 16 players for a battle, the matchmaker does a very bad job of splitting them up into 2 teams.

It seems to completely ignore the statistic "avg. supplies used per hour in last 20 MM battles" for each player, and this matters the most, along with GS, when analyzing team balance. There are heavy druggers, and there are players who don't use supplies at all.

In a pool of 16, if there are 4 heavy druggers, and 12 no supply users, their balancer often puts the druggers as 3:1 or 4:0, and only about 50% of the times, in 2:2 ratio.

The biggest problem with this game is the poor matchmaking system, and not p2w. Because many times that mega buyer will get 7 no supply users in his team, and lose the game.

 

Some other statistics they need to consider are win/loss ratio in last 20 MM games, score/hour in last 20 MM games, avg. damage dealt, amt. healed in last 20 MM games etc. ...

And they have so many analysts and programmers...

 

 

Edited May 21, 2020 by 799169

[von_Cronberg 258]

While the MMS can always be further optimized (and I just made a concrete suggestion about it), there are natural limits to it. The issue of quitters I mentioned above is one of them. The mood of players is another. For example, if I play a couple of MM battles with Hornet-Vulcan-Defender, dominating them with many hundred experience points as first in my team, and then feel in the mood to play a relaxed Hunter-Magnum-Saboteur or a relaxed Titan-Shaft without any drone, good for my K/D but bad for my battle score and probably also bad for my team, how should an MMS calculate my change of mood? There is one player I sometimes meet in MM battles (will not write the account name), who sometimes plays like the god of Tanki, masterfully deciding battles all by himself, and sometimes plays without using any supplies like an almost asleep little kid, it is always one of these extremes. When I see this account on my team list, I have no idea what I am in for, how would the MMS know?

[799169 48]

8 minutes ago, von_Cronberg said:

While the MMS can always be further optimized (and I just made a concrete suggestion about it), there are natural limits to it. The issue of quitters I mentioned above is one of them. The mood of players is another. For example, if I play a couple of MM battles with Hornet-Vulcan-Defender, dominating them with many hundred experience points as first in my team, and then feel in the mood to play a relaxed Hunter-Magnum-Saboteur or a relaxed Titan-Shaft without any drone, good for my K/D but bad for my battle score and probably also bad for my team, how should an MMS calculate my change of mood? There is one player I sometimes meet in MM battles (will not write the account name), who sometimes plays like the god of Tanki, masterfully deciding battles all by himself, and sometimes plays without using any supplies like an almost asleep little kid, it is always one of these extremes. When I see this account on my team list, I have no idea what I am in for, how would the MMS know?

Nobody knows what the price of a stock will be tomorrow. Yet there are people who can make 10x more money off the stock markets than others.

Its about prediction, and accuracy - gathering relevant information, and building good forecasting models, regression and analytics. Good analysts can get much higher percentage of their predictions to be true.

The guys who give the weather forecast in the newspaper get it right more than 90% of the times. The predictions of Tanki's matchmaking system are more like that of a drunk old man sitting at a bar who says "It would be nice if it rains tomorrow. Maybe it will rain tomorrow. Yes, I am convinced that it will rain tomorrow! I might as well an umbrella today to avoid getting wet."

No matchmaking system can ever be perfect, but there is significant improvement that can be done in this one, without using much brains. A lot of good ideas have been suggested in the forum. Even if half of them get accepted, the system will improve a lot in balancing battles.

Here's another statistic they could consider:

People who have a mission to "Finish ? number of battles" are less likely to play good, relative to the case where they do not have this mission. A good analyst can predict numbers like "they are likely to perform only upto 75% of their potential / usual battle score when they get this finish battle mission".

(This might explain the account you saw)

Thousands of players have played 100s of hours of this game. Millions of MM battles have been completed. Thats enough data to make a lot of better predictions, which their system does not currently do.

 

 

[NooNooHead 121]

1 hour ago, von_Cronberg said:

But never, not once, not a single time in those well over ten thousand battles, not once was my team massacred as long as every player stayed in the game and persistently sought to hold their position.

That is somewhat true. However, I am one of those who leave MM if I see too many rails, shafts and magnums camping for kills in the back corner of the map in MM (except for tdm or juggernaut).

[printing_now 59]

Gee... I wonder.

How can a game system favor Russian players?

[Guest]

5 hours ago, TinBoxGunner said:

 

Just had a battle with my alt acc of General.  MM Balancing sucked once more.  Enemy team had 4 legends, out team had none.  Enemy teams GS average was about 8800, our was about 6800.  Devs are either imcompetent or somehow favouring russian players with MM balancing.  Which way is it?

Working as intended because losses get you to play more.

[DieselPlatinum 5,121]

6 hours ago, von_Cronberg said:

The issue of the MMS is a large number of players leaving battles they were assigned to, with the MMS then struggling to fill the ranks with players who click to join a battle in that second. The quitters are the problem.

 

5 hours ago, von_Cronberg said:

But never, not once, not a single time in those well over ten thousand battles, not once was my team massacred as long as every player stayed in the game and persistently sought to hold their position.

 

4 hours ago, von_Cronberg said:

The issue of quitters I mentioned above is one of them.

There are many reasons why many people leave the battle that they dont like. They don't just do it without reason.

It would be best to fix the reasons that are making people not want to stay in that battle.

Good luck trying to fix people.

[wolverine848 13,097]

5 hours ago, von_Cronberg said:

I have played well over ten thousand battles on Tanki. Some where wins for my team and some were losses, in some I was more pleased with my personal performance and enjoyment and in some I was less pleased with it. In a number of them we massacred the enemy, and in a number of them we were massacred. But never, not once, not a single time in those well over ten thousand battles, not once was my team massacred as long as every player stayed in the game and persistently sought to hold their position.

The you are the luckiest player to have ever played the game.  It happens all the time. 

And 4-0 after 4 minutes... it really does not matter if a player leaves at that point.  Battle was already decided.

5 hours ago, Maf said:

What does this have to do with Russian players? You have no way to be sure where the enemy players are from.

Also, GS is just one of the factors taken into account in MM balancing. A player could have a low GS, but enough skills to be equivalent to one with higher GS, so their overall K/D, W/L ratio and other stats will affect their strength when being put into teams. And sometimes you just get unlucky and lose even while having an overall gear advantage.

What?   K/D and W/L? 

So I'm playing a CP battle and MM might "reward" me with a bunch of campers?  How does that help?  It's a recipe for a LOSS.

And where is that W/L ratio coming from?

[von_Cronberg 258]

2 hours ago, wolverine848 said:

The you are the luckiest player to have ever played the game.  It happens all the time. 

And 4-0 after 4 minutes... it really does not matter if a player leaves at that point.  Battle was already decided.

If your idea of "getting massacred" is losing a battle by 4-0, then you certainly are the luckiest player to have ever played the game. You must be so thrilled when you hear, late at night at the fireplace, others talk about experiences of hardly surviving respawn, being engaged by several enemy tanks every second, finding no place to avoid incoming enemy fire, watching the stupid kid with the Mammoth-Twins relocate from defending the enemy flag to your spawn zone. When you hear others talk about the kind of stuff that happens when teammates desert their position and quit a running battle, the stuff that does not happen as long as every teammate stays in the game and holds their position.

[wolverine848 13,097]

5 minutes ago, von_Cronberg said:

he stuff that does not happen as long as every teammate stays in the game and holds their position.

This is a pipe-dream unless you play with friends all the time. Co-operation among random players in team battles who often don't speak the same language is rare - at best.

And those 4-0 games usually includes all the other stuff you mentioned - that was basically implied.

 

Not sure what your solution to "deserters" is... if it's what TX tried = doomed to fail.  Players leave battles for a long list of reasons - many of them legitimate.  Until TO fixes MM (and a great many other things) players will leave battles early.

 

[von_Cronberg 258]

3 minutes ago, wolverine848 said:

Co-operation among random players in team battles who often don't speak the same language is rare

I do not talk about sophisticated cooperation. Them not deserting the running battle is good enough for me. And the concept of desertion is known across civilizations, languages and cultures, and the act of desertion universally despised.

I just played half a dozen SGE battles. All of them were wins of 5-0 for one of the teams. Only one of them included a massacre, because in all the others, the folks of the lesser scoring team did not desert and quit battle.

 

5 minutes ago, wolverine848 said:

Players leave battles for a long list of reasons - many of them legitimate.

Nope. The overwhelming reason for desertion is mere convenience, at the peril of one's teammates.

After this conversation, I do no longer believe that you are the luckiest player to have ever played the game. I do now believe that you simply habitually are the first to desert your comrades and quit a battle when the going gets tough.

[wolverine848 13,097]

1 minute ago, von_Cronberg said:

I do not talk about sophisticated cooperation. Them not deserting the running battle is good enough for me. And the concept of desertion is known across civilizations, languages and cultures, and the act of desertion universally despised.

I just played half a dozen SGE battles. All of them were wins of 5-0 for one of the teams. Only one of them included a massacre, because in all the others, the folks of the lesser scoring team did not desert and quit battle.

 

Nope. The overwhelming reason for desertion is mere convenience, at the peril of one's teammates.

After this conversation, I do no longer believe that you are the luckiest player to have ever played the game. I do now believe that you simply habitually are the first to desert your comrades and quit a battle when the going gets tough.

Who said anything about "sophisticated cooperation".  There is little to none in 85% of the battles.

DO YOU REGULARLY PLAY TO BATTLES WITH FRIENDS?

 

LOL - you know nothing about me.

 

[von_Cronberg 258]

6 minutes ago, wolverine848 said:

DO YOU REGULARLY PLAY TO BATTLES WITH FRIENDS?

I like to play PRO battles with friends. I am not a fan of playing MMS battles in groups. But how does that relate to our topic here?

 

7 minutes ago, wolverine848 said:

you know nothing about me.

True. But I recently made the suggestion to Tanki to calculate players' finishing quota and make the information public on their profiles. My own finishing quota certainly is above 90 percent, in both PRO and MMS battles. Now I am really curious to see yours, after the impression I got in this conversation.

[wolverine848 13,097]

5 minutes ago, von_Cronberg said:

I like to play PRO battles with friends. I am not a fan of playing MMS battles in groups. But how does that relate to our topic here?

 

True. But I recently made the suggestion to Tanki to calculate players' finishing quota and make the information public on their profiles. My own finishing quota certainly is above 90 percent, in both PRO and MMS battles. Now I am really curious to see yours, after the impression I got in this conversation.

When you play with friends the co-operation in battle goes up exponentially - and thus reaches a much better outcome far more often than not.  That's not obvious to you?

I stick it out in majority of battles.  But I will not hesitate to leave if MM places me in a battle > half over or a team is capping a flag before most of my team is loaded or I see multiple players doing nothing for the first couple minutes of battle.  We don't start chess matches down a piece (or 3) so why should I start an MM battle that way.  Will also leave if it's just the other team spawn-killing us and not even caring about capping - I'm playing to have fun - not pad the other teams score.

[JohnMatrix 18]

I'm curious about W/L ratio being an input to the MM algorithm. Is this a new stat? Is it public? Is quitting considered a loss?

[Maf 32,305]

4 hours ago, wolverine848 said:

What?   K/D and W/L? 

So I'm playing a CP battle and MM might "reward" me with a bunch of campers?  How does that help?  It's a recipe for a LOSS.

And where is that W/L ratio coming from?

I'm mostly guessing. As you know, devs didn't disclose the exact formula that matchmaking uses, but it's obviously more than just GS.

[JohnMatrix 18]

11 hours ago, TinBoxGunner said:

 

Just had a battle with my alt acc of General.  MM Balancing sucked once more.  Enemy team had 4 legends, out team had none.  Enemy teams GS average was about 8800, our was about 6800.  Devs are either imcompetent or somehow favouring russian players with MM balancing.  Which way is it?

I always assume there are groups playing vs randoms which would explain a lot of imbalances.

[von_Cronberg 258]

19 minutes ago, wolverine848 said:

I will not hesitate to leave if MM places me in a battle > half over or a team is capping a flag before most of my team is loaded or I see multiple players doing nothing for the first couple minutes of battle.

I do not consider any of these a legitimate reason to desert an MMS battle. One of the teams will lose the battle score anyway, and in statistically half of the cases it will be your team. Crystal rewards in MMS battles are almost the same for players of the losing team. When I get assigned to an MMS battle and see my own team as weak, my prime thought is that I will now most certainly earn six stars as the best scorer of the losing team.

[DieselPlatinum 5,121]

2 hours ago, von_Cronberg said:

The overwhelming reason for desertion is mere convenience, at the peril of one's teammates.

Nope, this is just mere assumption, not fact.

2 hours ago, von_Cronberg said:

I do now believe that you simply habitually are the first to desert your comrades and quit a battle when the going gets tough.

Hate to break it to you, but it's 2020 in a video game, not 1960 in reality.

2 hours ago, von_Cronberg said:

My own finishing quota certainly is above 90 percent

You're probably kissing ponytail boy's feet to be treated like such a king.

1 hour ago, von_Cronberg said:

I do not consider any of these a legitimate reason to desert an MMS battle.

What's one legitimate reason to stay in an MM battle where your team is getting spawn killed over and over again?

1 hour ago, von_Cronberg said:

My prime thought is that I will now most certainly earn six stars as the best scorer of the losing team.

Nice to know that you care about earning stars.

[wolverine848 13,097]

4 hours ago, von_Cronberg said:

I do not consider any of these a legitimate reason to desert an MMS battle. One of the teams will lose the battle score anyway, and in statistically half of the cases it will be your team. Crystal rewards in MMS battles are almost the same for players of the losing team. When I get assigned to an MMS battle and see my own team as weak, my prime thought is that I will now most certainly earn six stars as the best scorer of the losing team.

Who said anything about "weak"?  "Weak" and "non-participating" are not the same thing.

It's like you didn't actually read anything I posted.

Good thing you're not in a position to make important decisions here -  the game would close even sooner than it inevitably will.