Matchmaker Discussion Thread & MM Balance

[Fatal_Ramses]

26 minutes ago, ColonelPete said:

When you decide you lose, because of the lineup, then you will lose, because of your attitude.

There is no lineup that guarantees a loss.

Oh yes. "I decided" to lose. Taking out both of their DDs and one BB....Yes, I was not even trying to win.

[ColonelPete]

25 minutes ago, HP3070 said:

Yes it is. What's your point?

 

And the vast majority of players being between 40-60% is statisticly to be expected.

That random teams produce these results, without rigging.

25 minutes ago, HP3070 said:

....a mm that is focused on producing as much evenly chanced games as possible? 

That would be rigging and would put players even much closer to 50%.

5 minutes ago, Fatal_Ramses said:

Oh yes. "I decided" to lose. Taking out both of their DDs and one BB....Yes, I was not even trying to win.

What made you think your setup was without chance? You had a T-61, GC, Izmail, Pyotr....

That is far from weak.

[159Hunter]

7 hours ago, HP3070 said:

 The goal of wg's mm is not to produce a maximum number of "50/50 chance to win" games.

Wrong. The goal is to create games with equal line ups between classes, divisions and tiers (thus not looking at individual players and ships) as fast as possible.

7 hours ago, HP3070 said:

The goal is to actively balance the game so players are forced to 50% wr.

You do realize that this is exactly the same statement as before, yet you see the two as contradictions. Cause you know, always creating 50/50 win games will move people to a 50% WR.

 

PS: the median WR is 48% and has been for years. So even if your statements were true, WG is failing without compensating to achieve these 'goals'.

 

7 hours ago, HP3070 said:

If you can overcome a small number of "against you" games then your wr will increase slightly. If not it will drop slightly.

Slightly? I call 5 % up and down more than slight. A difference of 5% in player skill is huge gameplay wise.

7 hours ago, HP3070 said:

WG never obscured this. It's even present in the patent they have on their mm.

Source please.

 

 

[Fatal_Ramses]

8 hours ago, ColonelPete said:

What made you think your setup was without chance? You had a T-61, GC, Izmail, Pyotr....

That is far from weak.

Yes, I only look at the ships, not the players behind the helm.

[ColonelPete]

3 minutes ago, Fatal_Ramses said:

Yes, I only look at the ships, not the players behind the helm.

Once again, your team was not at a disadvantage, to the contrary. What are you imagining?

[COPlUM]

T61 top tier? You snuck a T9 ship in a T5 game, should have carried much harder.

[st_dasa]

Nice MM.

 

[GarrusBrutus]

1 hour ago, st_dasa said:

Nice MM.

 

Remember when BBs complained about HE spam? They've become the very thing they swore to destroy.

[159Hunter]

7 hours ago, GarrusBrutus said:

Remember when BBs complained about HE spam? They've become the very thing they swore to destroy.

''Can't beat them? Join them.''

And in this case WG made sure the players didn't  even need to change classes.

[steelfins_tuna]

well, my take on this might sound a little bit irrelevant to what people are trying to address regarding the MM, I think what is gonna make a huge difference is to involve a skill-based MM system, in addition to all the other points addressed in this thread. 

 

from what I have experienced in the past few days playing WOWS again, to me playing with a skilled CA player as a DD for example gave me very good experience and he was playing a Tech tree line, on the other hand, and I guess we can all relate to all the crappy thunderer players, which is a premium ship that they have no clue how to play it. and am not talking about farming HE damage without much contribution to win  the game, am talking about playing it properly as a battleship.

 

its really hard for me to play this game when u get matched with bot level players, no offense to novice players but its just not fair!

please consider doing this, every single MMO uses skill-based MMR except you guys which is very questionable!

 

[tocqueville8]

 

 

- 2 afk BBs

- 2 more that sort-of went for the cap, but then turned away and ran when they saw the afk ones

- "my" Moga began by reversing, then yoloed into the cap and was Dev-Struck by a Shima

- all caps lost by minute 4, there was simply no time to do anything: I tried to block the cap, got focused by GZ, Shima and Moskva

 

I'm not for "skill-balanced MM" (everyone would get 50% WR, yada yada yada...), but this is intolerable. It doesn't have to be 50-50, but these rolfstomps are absurd, and unfun for both sides.

[159Hunter]

1 hour ago, tocqueville8 said:

- 2 afk BBs

- 2 more that sort-of went for the cap, but then turned away and ran when they saw the afk ones

And there is nothing MM can do about this. Purely a player issue.

[DFens_666]

Just now, 159Hunter said:

And there is nothing MM can do about this. Purely a player issue.

 

Maybe self-fulfilling prophecy using MMM?

I think we need an anonymizer, so people dont know the WR from their teams or WR of a specific player so they can "focus that one Unicum"...

[159Hunter]

29 minutes ago, DFens_666 said:

 

Maybe self-fulfilling prophecy using MMM?

I think we need an anonymizer, so people dont know the WR from their teams or WR of a specific player so they can "focus that one Unicum"...

But but but...

My statshaming...

How can I get chat bans in such a world?

[tocqueville8]

12 hours ago, 159Hunter said:

And there is nothing MM can do about this. Purely a player issue.

I'm pretty sure there is at least *some* correlation between being afk, or running away at the first sign of trouble giving up half the map, and having poor WR.

[djole7]

Lets not speak about Ranked system.

Wg sould look to other games with ranked systems how it works.

This is just for grinding.

Nothing to do with ranked.

[tocqueville8]

On 3/12/2021 at 6:40 AM, 159Hunter said:

You do realize that this is exactly the same statement as before, yet you see the two as contradictions. Cause you know, always creating 50/50 win games will move people to a 50% WR.

Actually I was wondering...

 

If the two teams ware balanced *a little bit*, let's say enough to make sure the difference in avg WR difference couldn't be greater than 5%, would that lead to 50-50 odds, or simply to the same odds, but with fewer rolfstomps?

 

I mean, rolfstomps can happen even with perfectly balanced teams: DDs dying early for whatever reason, RNG when firing at turning cruisers, detonations, someone having a bad day...

 

I'm okay with losing my share of games (many due to my own mistakes, admittedly): but one-sided games are more than that, they're often a waste of time (and signals, special camos, etc.) for both parties. In the example I posted above, we were probably going to "lose from the start", and that's fine, but there's losing from the start in 12 minutes and losing from the start in 4 minutes...

 

I'm just wondering if clipping the most extreme cases of unbalanced MM wouldn't actually result in the same WRs, but much less frustration...

[DFens_666]

29 minutes ago, tocqueville8 said:

I'm pretty sure there is at least *some* correlation between being afk, or running away at the first sign of trouble giving up half the map, and having poor WR.

 

People can also have poor WR because they suicide after 3 minutes.

So unless MM would know, how each person behaves, "balancing" them between teams is not really possible. F.e. a division of Unicums is often able to carry their team, if they just stay alive long enough. But if half your team suicides to get zero'd out, then noone can change that.

[anonym_cwVecOS6ecVy]

On 2/3/2021 at 12:44 AM, Moinos said:

I'd like to mention a thing, that really annoys me:

 

When I'm playing in a Division with 2 friends (they have tier VII, I have tier VI), sometimes we get into a match with a lot of tier IX ships. I was effectively uptiered by 3 levels, which is very hard to deal with. And this happens regularly.

As far as I know, the MM takes the tier of the higher ship in your Division and finds a game accordingly. Would it be feasible to take the lower tier in a division for MM?

Wtf? You are creating a fail division and then expect WG to changes metrics, so your fail division buddies even profit from it by getting a more mild MM? Just do not create fail divisions and [edited]up your entire team's game, simple as that.

[tocqueville8]

Just now, DFens_666 said:

People can also have poor WR because they suicide after 3 minutes.

That would be our Mogador in the above example.

Still, if one does that regularly it should reflect in his WR.

 

I dunno, it seems reasonable to me that between the "balanced MM would make everyone's WR 50%" and the "MM can't know how each person will behave" positions, there's room for the "let's cut the most inane rolfstomps, but everyone would basically keep his WR" position...

[DFens_666]

3 minutes ago, tocqueville8 said:

Still, if one does that regularly it should reflect in his WR.

 

Games played / WR / survivability

No, you dont need to die always to get a really bad WR

 

Different persons, same WR, die much more often.

 

Different playstyles can have same outcome, but if you change the distribution, suddenly one playstyle can become more "viable" so to speak. That first Fletcher is probably always running to A1, but if he is on my team, atleast he cant gift points to the enemy. Would his playstyle result in some losses? sure, but i assume one of the other suiciding Fletchers will cost more games, because they simply die. Btw incidently, that guy with 75% survivability even deals more damage than the other 2.

[tocqueville8]

Just now, DFens_666 said:

Games played / WR / survivability

No, you dont need to die always to get a really bad WR

Clearly the 1st guy doesn't yolo, but maybe neither do the other ones, they just play the Fletcher more like a gunboat, or they go after DDs more instead of stalking and torping, or they're less careful around radar, but more proficient (relatively speaking) some other way.

 

I'm not arguing about balancing death rates, just WRs. In the example I posted, only 1 yoloed, 2 did the exact opposite, running away at the first sign of trouble (and 2 were afk). What's common to all of them isn't one specific strategy, it's the fact that it was an extreme one.

[Det_Haendlerkaertschen]

If the MM assigned players really randomly to the teams, and is only adhering to the published rules, everything should be fine. All parameters influencing the game (radar, player skill, stock ships,...) should be assigned randomly to the team (and therefore evenly) as should all other parameters Not affecting the gameplay (smoker vs non smoking player, old vs. young player etc.) Clinical trials are done that way, to ensure minimalisation of bias. High quality publications prove equal distribution of patient parameters between the two groups to be compared.

 

of course, statistics mean that especially in small groups, you cannot achieve total equality. I am aware of that. However, emotions tell me that what I experienced in WOWs matches over the last years, the system appears rigged to me: VERY often 5-10 losses in a row, unequal distribution of super unicum players in divisions (like BANCV in the red Team, xyzkriegamarinepotatoedivi in my Team). I am asking myself: can this constant experience really explained by statistics? Little white mouse  says: yes. She  says it always comes back to your own personal skill. The problem is that I fail trying to overcome my doubts.
 

Thus my conlusion is that the MM INTENTIONALLY AVOIDS real randomization to ensure asymmetrical games. Dunno what WG might want to achieve with that. And of course I have neither proof, nor is my knowledge about statistics substantial enough.

 

There you go.

Cheers and take care

Det_Haendlerkaertschen (game client still uninstalled due to opposition against Dead Eye)

[ColonelPete]

13 minutes ago, Det_Haendlerkaertschen said:

However, emotions tell me that what I experienced in WOWs matches over the last years, the system appears rigged to me: VERY often 5-10 losses in a row, unequal distribution of super unicum players in divisions (like BANCV in the red Team, xyzkriegamarinepotatoedivi in my Team). I am asking myself: can this constant experience really explained by statistics? Little white mouse  says: yes. She  says it always comes back to your own personal skill. The problem is that I fail trying to overcome my doubts.

 

Thus my conlusion is that the MM INTENTIONALLY AVOIDS real randomization to ensure asymmetrical games. Dunno what WG might want to achieve with that. And of course I have neither proof, nor is my knowledge about statistics substantial enough.

That is why you cannot judge the MM emotionally.

If you do that, the system is rigged for 99% of the players, which is obviously impossible.

[Dampfboot]

TVI getting constantly matched into TVIII matches, very fun&engaging^tm