Jump to content

The Last Match Maker Thread We Need


248 replies to this topic

#81 The6thMessenger

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Nova Captain
  • Nova Captain
  • 8,104 posts
  • LocationFrom a distance in an Urbie with a HAG, delivering righteous fury to heretics.

Posted 11 June 2019 - 02:21 AM

View PostPhoenixFire55, on 11 June 2019 - 01:50 AM, said:

Tbh its not that hard to prove that pretty much any MM is going to be better than what we currently have, but sure.


Honestly, that's my only take-away from the thread. That being said, it's probably because of the the title.

View PostPhoenixFire55, on 11 June 2019 - 01:50 AM, said:

So it is, lets put it politely, strange to expect W/L MM to fare any better in actual game. Because none of the problems that Elo MM had in MWO were at any point addressed by anyone, including the OP in this thread.



You mean the problem of auto-fellative, self-referencing problem?

Like I said before, that's pretty much what we have to process since we're going to calculate the outcome based on the scores we get. Are you arguing for a completely random match-maker instead? Because it sounds like that.

Sure you could possibly normalize win-rates of matches by that approach, but it's still not fun seal-clubbing one witless player, and getting seal-clubbed by another. We have to somehow match the upper scorers with the upper scorers, the mid-scorers versus mid-scorers, and low-scorers vs low-scorers. If we're too concerned of the inevitable auto-fellative data processing, we also won't get anywhere.

View PostPhoenixFire55, on 11 June 2019 - 01:50 AM, said:

P.S. Also please stop typing ELO, its not an abbrevation, its named after a person - Arpad Elo.


You know you're just making it worse because Psychological Reactance right?

#82 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 11 June 2019 - 02:40 AM

View PostThe6thMessenger, on 11 June 2019 - 02:21 AM, said:

You mean the problem of auto-fellative, self-referencing problem?

Its only one of many.

View PostThe6thMessenger, on 11 June 2019 - 02:21 AM, said:

Like I said before, that's pretty much what we have to process since we're going to calculate the outcome based on the scores we get.

Eh ... no. There are plenty of player performance metrics game can easily track.

View PostThe6thMessenger, on 11 June 2019 - 02:21 AM, said:

Are you arguing for a completely random match-maker instead? Because it sounds like that.

Eh? ... LOL. I'm not arguing for any kind of MM. I'm like 5 years past that.

View PostThe6thMessenger, on 11 June 2019 - 02:21 AM, said:

Sure you could possibly normalize win-rates of matches by that approach, but it's still not fun seal-clubbing one witless player, and getting seal-clubbed by another. We have to somehow match the upper scorers with the upper scorers, the mid-scorers versus mid-scorers, and low-scorers vs low-scorers. If we're too concerned of the inevitable auto-fellative data processing, we also won't get anywhere.

It is my firm belief that the MM we currently have already acts EXACTLY like a totally random MM. Because think about ... everyone from elite players to baddie-bad-baddersons can easily end up in T1. Hence a T1 player is basically a random skill player as it is.

Personally I couldn't care less at this point. Bottom line ... MM is just a placeholder that ticks the "we have a MM!" box and nothing else. It won't ever change. But this thread doesn't provide a solution either, even if there was any willingness to implement it.

#83 FRAGTAST1C

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Fighter
  • The Fighter
  • 2,919 posts
  • LocationIndia

Posted 11 June 2019 - 02:58 AM

View PostPhoenixFire55, on 11 June 2019 - 02:40 AM, said:

It is my firm belief that the MM we currently have already acts EXACTLY like a totally random MM. Because think about ... everyone from elite players to baddie-bad-baddersons can easily end up in T1. Hence a T1 player is basically a random skill player as it is.

Personally I couldn't care less at this point. Bottom line ... MM is just a placeholder that ticks the "we have a MM!" box and nothing else. It won't ever change. But this thread doesn't provide a solution either, even if there was any willingness to implement it.


When you consider what it was like during the Quake 3 days, servers were private and people chose to play on specific servers and it didn't matter whether they lost or won. What mattered was whether they had a good time or not. Now, if we take any Battle Royal games, you, a level 10 player can team up with your friend, a level 50 player, and play matches that are level 50, not lower. It doesn't matter there 'cause you're teaming up with friends and just playing for fun. It's entirely down to your performance and whatever joy you can mine from it.

What happens in MWO? There are far too many buckets splitting people up and causing a downright sewage system for fun. So what if people want to play together? Somehow the balance is more important than fun? Just get rid of all this stat and lore bs and make mechs somewhat equal. Then let people have fun.

If PGI's hell-bent on having a MM, then make it uniform and remove the buckets. Just split players based on skill. But there's this problem of having many crappy mechs as well. So, what is a MM going to do if a 99% player decided to use a Spider and someone who's 90% brings a Cyclops?

There are a lot to consider than simply placing WLR or KDR or MS or whatever metric PGI "must" use to balance the games. First make it so that friends can get a game together.

#84 PhoenixFire55

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 5,725 posts
  • LocationSt.Petersburg / Outreach

Posted 11 June 2019 - 03:16 AM

View PostFRAGTAST1C, on 11 June 2019 - 02:58 AM, said:

What happens in MWO? There are far too many buckets splitting people up and causing a downright sewage system for fun. So what if people want to play together? Somehow the balance is more important than fun? Just get rid of all this stat and lore bs and make mechs somewhat equal. Then let people have fun.

That is your definition of fun. Other people have other definition of fun. End of story. Most people who play MWO are the fans of BT franchise, its lore and its limitations. Remove that and literally nobody will play a Minimally ViableTM generic robot shooter.

#85 Khobai

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 23,969 posts

Posted 11 June 2019 - 04:16 AM

balance and fun arnt mutually exclusive things though

theyre two separate axis.

games can fall in one of four quadrants: not balanced and not fun, not balanced and fun, balanced and not fun, balanced and fun.

right now MWO is in the not balanced and not fun quadrant lol.


PGI needs to come up with a solution for casual groups to play in group queue without getting stomped all the time. Implementing a working matchmaker and allowing solo players in group queue to help keep the player count up is one possible solution.

Edited by Khobai, 11 June 2019 - 04:27 AM.


#86 The6thMessenger

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Nova Captain
  • Nova Captain
  • 8,104 posts
  • LocationFrom a distance in an Urbie with a HAG, delivering righteous fury to heretics.

Posted 11 June 2019 - 05:00 AM

View PostPhoenixFire55, on 11 June 2019 - 02:40 AM, said:

Eh ... no. There are plenty of player performance metrics game can easily track.


Which is/are?


View PostPhoenixFire55, on 11 June 2019 - 02:40 AM, said:

It is my firm belief that the MM we currently have already acts EXACTLY like a totally random MM. Because think about ... everyone from elite players to baddie-bad-baddersons can easily end up in T1. Hence a T1 player is basically a random skill player as it is.


Oookay.

View PostPhoenixFire55, on 11 June 2019 - 02:40 AM, said:

Personally I couldn't care less at this point. Bottom line ... MM is just a placeholder that ticks the "we have a MM!" box and nothing else. It won't ever change. But this thread doesn't provide a solution either, even if there was any willingness to implement it.


I don't know though, Nightbird's simulation is kind of convincing considering he actually put work on it.

I get that it seems like a Spherical Cow and all, but in the end when we're merely playing with statistics, that's what we got, it's all theoretical with nothing math because all we have to measure is math. Chances are, we'd be making more problems than answering solutions, but at least it's a step towards better MM than what we got right now.

#87 Der Geisterbaer

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 802 posts

Posted 11 June 2019 - 05:36 AM

View PostThe6thMessenger, on 11 June 2019 - 05:00 AM, said:

I don't know though, Nightbird's simulation is kind of convincing considering he actually put work on it.


Food for thought on that: Take a closer look at some of the premises made for the simulation of the current MM. Nightbird explicitly used a "hidden" normal distribution of skill as the MM supposedly cannot discern skill levels directly. But if you take a look at the global stats that the Jarl's List provides for "Average Match Score" you'll find that the graph derived from all data points on that appears to be a normal distribution as well. So the available stats do seem to reflect "skill" quite nicely and guess what? Average match score quite obviously is not the same as W/L ratio that Nightbird tries to present as "ultima ratio" in terms of suitable metrics.

I'll leave it up to you to take from that whatever you like but from a "scientific" standpoint Nightbird should certainly assess his own level of confirmation bias and other potential sources of systematic error in what he presented.

#88 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 11 June 2019 - 05:58 AM

For the people that keep saying MS this, MS that, you can easily see on Jarl's list that people with 300 avgMS with thousands of games can have 1WLR to 2WLR, whereas people with 200avgMS can have 0.5 to 1.5 WLR.

In some way, higher avgMS correlates to higher WLR, but because of the overlap you only say that on average 300MS players are better, as some 200MS players will certainly be better than 300MS players. That makes it a bad measuring stick to use.

Another reason that avgMS isn't a good indicator is that people can easily inflate or deflate it without changing their contribution to their team winning. Shoot an extra arm off instead of getting a kill and you boost your avgMS by 25pts. Swing a laser and scratch everyone on the other team whenever you can to get a 50avgMS boost, etc etc. Don't do these things and you can lower your MS intentionally without hurting your team in the least.

At the worst of it, some people are trying so hard to boost MS that they're probably hurting their team and their own WLR to do it. This is the folly of Jarl's list.

Trusting MS is the same as trusting PGI to correctly assign weights to actions in match that contribute to winning. i.e. when you trust MS you trust PGI to be doing something right.

Edited by Nightbird, 11 June 2019 - 06:37 AM.


#89 FRAGTAST1C

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Fighter
  • The Fighter
  • 2,919 posts
  • LocationIndia

Posted 11 June 2019 - 06:31 AM

View PostPhoenixFire55, on 11 June 2019 - 02:40 AM, said:

It is my firm belief that the MM we currently have already acts EXACTLY like a totally random MM. Because think about ... everyone from elite players to baddie-bad-baddersons can easily end up in T1. Hence a T1 player is basically a random skill player as it is.


Is there a word to describe skewed thinking, tap dancing and a form of confirmation bias together?

View PostPhoenixFire55, on 11 June 2019 - 03:16 AM, said:

That is your definition of fun. Other people have other definition of fun. End of story. Most people who play MWO are the fans of BT franchise, its lore and its limitations. Remove that and literally nobody will play a Minimally ViableTM generic robot shooter.


Right. I keep forgetting that it isn't PGI who's screwing up the game. It's the vast chunk of this community to blame as well.

#90 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 11 June 2019 - 07:51 AM

For the people saying, make the game fun. That's like saying you want to be successful, or famous, or funny. You need a way of measuring it, is it in wealth, or followers, or laughs?

In my simple model, I cut the number of stomps in half, for most people, the game would be more fun. I created 3x more balanced matches, so your actions actually matter. Today, you can carry hard and not change the fact your team is destined to lose, or AFK and not change the fact your team will win, in half of all games.


View PostDer Geisterbaer, on 11 June 2019 - 05:36 AM, said:


Food for thought on that: Take a closer look at some of the premises made for the simulation of the current MM. Nightbird explicitly used a "hidden" normal distribution of skill as the MM supposedly cannot discern skill levels directly. But if you take a look at the global stats that the Jarl's List provides for "Average Match Score" you'll find that the graph derived from all data points on that appears to be a normal distribution as well. So the available stats do seem to reflect "skill" quite nicely and guess what? Average match score quite obviously is not the same as W/L ratio that Nightbird tries to present as "ultima ratio" in terms of suitable metrics.

I'll leave it up to you to take from that whatever you like but from a "scientific" standpoint Nightbird should certainly assess his own level of confirmation bias and other potential sources of systematic error in what he presented.


Most things in the world, from people's height, weight, to IQ are normally distributed so it is the default distribution assumed for everything. Here, have a rag, mop up the grey matter escaping through your nose.

#91 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 11 June 2019 - 08:03 AM

View PostThe6thMessenger, on 10 June 2019 - 11:48 PM, said:

I've read the article, and it says which is which is better at predicting games. Such as the ELO was over reacting it even becomes inaccurate at very long seasons, and lowering the K-Value to 0.5 and the Beta to 400 would make it predict better than WLR at a certain degree.

I didn't see anything that says WLR and ELO is basically the same, it just said that they could predict the win-rate using both, and it was to be used to get the Brier score. They come close in terms of predicting, but having near-similar result is a different matter when it comes to the mechanics.

It's basically saying that an atomic bomb is basically the same as an average lump of TNT, because both go boom.


This is a great summary, you understood it. I will add in one final thought that should tie this article and MWO together. In sports, stats are done on a seasonal basis. You can tweak Elo formulas between seasons and even if it blows up, the damage is contained to a season. For MWO, you cannot say, oh ****, the K value is wrong, pull the plug and start over. (which ofc is what PGI realized after trying Elo) Compared to Elo, WLR is robust, there is no parameters to tweak, and it will never blow up. That is why for a game like MWO where stats are accumulated forever, you want what is robust over that is fragile.

So to change your analogy a little, Elo will always go boom, WLR will never, and that is the difference between the two.

Edited by Nightbird, 11 June 2019 - 08:04 AM.


#92 Der Geisterbaer

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 802 posts

Posted 11 June 2019 - 08:04 AM

View PostNightbird, on 11 June 2019 - 07:51 AM, said:

Most things in the world, from people's height, weight, to IQ are normally distributed so it is the default distribution assumed for everything. Here, have a rag, mop up the grey matter escaping through your nose.


I guess you going ad hominem there is a testimony to your ability of reflecting over your own confirmation bias and dealing with systematic errors?! ~laugh~

#93 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 11 June 2019 - 08:05 AM

View PostDer Geisterbaer, on 11 June 2019 - 08:04 AM, said:


I guess you going ad hominem there is a testimony to your ability of reflecting over your own confirmation bias and dealing with systematic errors?! ~laugh~


I'm taking the high road, I responded to a idiotic and rude post with a smart and rude post.

#94 Der Geisterbaer

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 802 posts

Posted 11 June 2019 - 08:12 AM

View PostNightbird, on 11 June 2019 - 08:05 AM, said:

I'm taking the high road, I responded to a idiotic and rude post with a smart and rude post.


~laugh~ Where exactly was my post "rude" in any sense of the word?

No, you're not taking the "high road" there, you're just showing yourself to be the exact same type of person that you accuse others of being. In other words: pot meet kettle.

#95 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 11 June 2019 - 08:19 AM

View PostDer Geisterbaer, on 11 June 2019 - 08:12 AM, said:


~laugh~ Where exactly was my post "rude" in any sense of the word?

No, you're not taking the "high road" there, you're just showing yourself to be the exact same type of person that you accuse others of being. In other words: pot meet kettle.


The rude part is obvious, so I'll point out the idiotic part again to rub it in. I used the normal distribution for my hidden skill factor, Dr. DG here see that match score on Jarl's list is normally distributed as well, and proclaims, both are normally distributed, therefore MS is a good representative of skill! I point out that height and weight are also normally distributed, so they must represent MWO skill as well per DG

Edited by Nightbird, 11 June 2019 - 08:21 AM.


#96 Der Geisterbaer

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 802 posts

Posted 11 June 2019 - 08:38 AM

View PostNightbird, on 11 June 2019 - 08:19 AM, said:

The rude part is obvious,


Really? Humor me, instead of making just the claim ;)

View PostNightbird, on 11 June 2019 - 08:19 AM, said:

so I'll point out the idiotic part again to rub it in.


I guess you'll trybut won't succeed.

View PostNightbird, on 11 June 2019 - 08:19 AM, said:

I used the normal distribution for my hidden skill factor,


I do seem to recall that I mentioned that myself ;)

View PostNightbird, on 11 June 2019 - 08:19 AM, said:

Dr. DG here see that match score on Jarl's list is normally distributed as well,


Which happens to be nothing but an observation. One which I even left up to personal interpretation (and for good reason). Now the question is, why are you trying to make further ad hominem there, by referring to an academic title that I haven't even remotely claimed? Remember, the only person so far to make professional qualification a center piece of his claims was you.

View PostNightbird, on 11 June 2019 - 08:19 AM, said:

and proclaims, both are normally distributed, therefore MS is a good representative of skill!


Now if only I truly had made the claim that (average) MS was a good representative for skill ~laugh~
I guess it's not just confirmation bias that you have to remind yourself of but a whole plethora of different fallacies besides the ad hominem fallacy that you are currently engaging in.

View PostNightbird, on 11 June 2019 - 08:19 AM, said:

I point out that height and weight are also normally distributed, so they must represent MWO skill as well per DG


Try to identify all the fallacies you're committing there. Hint: straw man is just one of them.

#97 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 11 June 2019 - 08:43 AM

View PostDer Geisterbaer, on 11 June 2019 - 08:38 AM, said:

I have no idea what I am talking about


I know right!

#98 LTC Kilgore

    Member

  • PipPipPipPipPip
  • 142 posts

Posted 11 June 2019 - 09:00 AM

This thread, although well intended and strikes true, a new MM is needed; is a waste of time. PGI is not going to invest resources into MWO anymore. Thought that was made pretty clear in Russ' dev update.

#99 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 11 June 2019 - 09:03 AM

View PostLTC Kilgore, on 11 June 2019 - 09:00 AM, said:

This thread, although well intended and strikes true, a new MM is needed; is a waste of time. PGI is not going to invest resources into MWO anymore. Thought that was made pretty clear in Russ' dev update.


Thanks and no worries

View PostNightbird, on 09 June 2019 - 10:05 AM, said:

Improving quality of matches reduces player frustration and player attrition. If PGI implemented this from day 1 the revenue for the life of this game would have doubled if not more.


View PostNightbird, on 09 June 2019 - 12:39 PM, said:

This thread was written for my own edification. Both for PGI not making the choices that maximize revenue and growth, and for the community for complaining that PGI doesn't listen. They do listen, they do UNFORTUNATELY listen, to opinions not backed up with hard science and when it fails, the parts of the community that suggested something different clamor they weren't listening to them and the cycle repeats.

This is for all the face palms I did PGI....


#100 Nightbird

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The God of Death
  • The God of Death
  • 7,518 posts

Posted 11 June 2019 - 02:51 PM

View PostFeral Clown, on 10 June 2019 - 10:39 PM, said:

I was kinda shocked that average match score wasn't the main thing that a tweaked matchmaker should be based on. It makes sense though considering a vet such as myself can average over 300 match score and not necessarily influence the outcome of a game and why I don't reach a wlr of 2.0.


I've found that farming wins and farming match score are opposing goals. When I want to farm match score, I take high dps dakka mechs and spray mechs down from arm to arm for maximum damage before killing them. You can do this also with LRMs and ATMs. When I want to farm wins, I take laser vomit and drill through the CT without wasting damage on anything else.





12 user(s) are reading this topic

0 members, 12 guests, 0 anonymous users