27 replies to this topic

[Remover of Obstacles]

Paul made some comments about PSR and Elo and that all Tier 1 players have the same ELO.
https://mwomercs.com...-session-apr-8/

Is anyone else concerned and confused why the matchmaking system isn't being addressed before pushing it to Faction Play? I think it has been pointed out many times that the current system allows players below the 50th percentile in measurable skill based metrics to be in Tier 1.

I am all for Faction Play getting some love and less choke points, but I do not understand the logic here.

[Athom83]

Possibly a hard reset back to tier 5 for everyone, but give a bonus to your next 20 matches or so for ELO/rank earning/loss.

[Curccu]

No need to fix matchmaking just tier system

[HammerMaster]

Zero Sum PSR.
DONE.

[Eatit]

What should Tier be based on?

Jarl's %?

1-20% are Tier 5
21-40% are Tier 4
41-60% are Tier 3
61-80% are Tier 2
81-99% are Tier 1

In this scenario the MM would pit 41% players against 99% players as it matches 2 tiers in either direction. Tier 3 players would have it the worst as they could be pitted against all other tiers.

Tier 2 players at 80% would be pitted against 21% players and so on. This is worse than it is now I think.

Where do we draw the line? The player base is so small that if we say that Tier 1 can only play Tier 1 then wait times for Tier 1 players will be astronomical.

I don't like the current situation any more than anyone else but I would like to see something constructive posted on the forums. Take some time to come up with a better solution and propose that.

I don't have any idea of how it can be done better within the constraints of the small player base. At least without creating long wait times.

[MechaBattler]

At least it's an attempt at some quality control of matches. But in any case the playerbase is too small. Not enough people to separate everyone up in their proper skill levels. Better than nothing. At least it's taking into account group sizes and what not.

Edited by MechaBattler, 10 April 2019 - 10:19 AM.

[Nesutizale]

MechaBattler, on 10 April 2019 - 10:19 AM, said:

At least it's an attempt at some quality control of matches. But in any case the playerbase is too small. Not enough people to separate everyone up in their proper skill levels. Better than nothing. At least it's taking into account group sizes and what not.

Just wanted to ask if we should be concerned about MM at all if the playerbase is so small that you can be happy to find 24 people in 30min.

[Eatit]

Here are some thoughts that might help some of you to come up with different ideas.

What is the lifespan of a MWO player?
1. Plays for a few months gets bored and moves on?
2. Plays for a few years never gets bored and never moves on?
3. Plays on and off for a few months at a time?

A ranking system that is built for player type 1 won't work for a system that is created for player type 2 or 3. The same is true for all types.

How then do we decide what to base the system on? Let's say that player type 1 is the majority of players. Lets base it on them. But what about player type 2? Haven't they spent the most time/money on the game? Shouldn't it be based on them? Aren't they the "Best" fans of the game?

All of these are subjective and based on opinion. Some will like it others won't.


How the MM has to create a match. Let's say there are 100 people in the que.

Pick the first 24 that meet the criteria outlined for weight distribution. Like 3 assault 3 heavy 3 medium 3 light. Now get those 24 from the players that have been waiting the longest.

Now the MM has to take the 24 players it got from that sort and sort them by "Skill"

What happens if there aren't enough of the right skill in the right size mech?
What happens if there aren't enough of the right size mechs?

Should we open the match up to more of a certain size of mech say we have 16 assaults 4 heavies 2 mediums and 2 lights.
How fun is that match going to be for the medium mech pilots?

Should we open the match up to more of a certain skill? 16 tier 1, 4 tier 2, 2 tier 3, and 2 tier 4.
How fun is that match going to be for the tier 4 pilots?

This isn't as simple as most people seem to think it is.

Now let's talk about the coding involved. Every system in the game that touches the "skill" value is going to have to be changed. Every screen that references the "skill" value is going to have to be changed.

Do you think that PGI is going to invest that kind of time/money in a 6 year old game with a small player base? That would be fiscal suicide for them. They are a business after all. They have employees that have to be paid. Overhead costs that have to be met.

I say be thankful that this free to play game is here for us to play. Don't be so critical of things you don't really understand. This is a game and should be fun. If you don't find it fun then move on to a game you do find fun. Don't bang your head against a wall that isn't going to budge just because you think it should or really want it to.

TL;DR Get a grip!

[Sixpack]

I don't think that it would have any kind of effect due to the limited playernumber for FW, the short countdowns and the long matchtimes leading to the MM just grabbing the first 12 people it can get on every side to get a match going in the first place.

[HammerMaster]

Eatit, on 10 April 2019 - 10:01 AM, said:

What should Tier be based on?

Jarl's %?

1-20% are Tier 5
21-40% are Tier 4
41-60% are Tier 3
61-80% are Tier 2
81-99% are Tier 1

In this scenario the MM would pit 41% players against 99% players as it matches 2 tiers in either direction. Tier 3 players would have it the worst as they could be pitted against all other tiers.

Tier 2 players at 80% would be pitted against 21% players and so on. This is worse than it is now I think.

Where do we draw the line? The player base is so small that if we say that Tier 1 can only play Tier 1 then wait times for Tier 1 players will be astronomical.

I don't like the current situation any more than anyone else but I would like to see something constructive posted on the forums. Take some time to come up with a better solution and propose that.

I don't have any idea of how it can be done better within the constraints of the small player base. At least without creating long wait times.

6 or 7 tiers to make it more granular.

[Vellron2005]

I see three solutions to this problem:

1) Recalculate PSR tiers, based on average matchscore in the last 100 matches played.

2) Set up a new FP-only event that tracks average matchscore, and use THAT as a basis for your ranking in FP.

3) Completely scrap PSR tiers, everyone back to T5, and then calculate ELO from the next 100 matches's average matchscore.

Edited by Vellron2005, 11 April 2019 - 12:09 AM.

[PhoenixFire55]

Remover of Obstacles, on 10 April 2019 - 09:43 AM, said:

Is anyone else concerned and confused why the matchmaking system isn't being addressed before pushing it to Faction Play? I think it has been pointed out many times that the current system allows players below the 50th percentile in measurable skill based metrics to be in Tier 1.

I am all for Faction Play getting some love and less choke points, but I do not understand the logic here.

PGI? Logic? ... Pfft.

The reason they push this half-azzed-non-effort update is because they promised us a FP update. And there you go ... We Listen to the Communuty^TM and Deliver on our Promises^TM. Who cares if the update is once again broken on realease and thus entirely meaningless? Nobody does.

Minimally Viable Product^TM.

[PhoenixFire55]

MechaBattler, on 10 April 2019 - 10:19 AM, said:

At least it's an attempt at some quality control of matches. But in any case the playerbase is too small. Not enough people to separate everyone up in their proper skill levels. Better than nothing. At least it's taking into account group sizes and what not.

10k active players. More than enough to separate by skill and then some. Hell, even 100 people is enough.

Bottom line. STOP. MAKING. LAME. EXCUSES. ABOUT. LOW. POPULATION.

Even 1k active players is enough population for a proper MM. The amount of players you need for a balanced 12v12 match drawing solo players is 24. Yes ... 24.

[Eatit]

TL;DR Skill is not being used to create matches. It's only being used to determine the order that solos get added.

I don't think I've ever seen 1k people que'd up for faction play in the history of this game. I don't know where you're getting your numbers from but I'd check your source again.

I think the best you would see would be 100 players in que.

That's 50 per side in an ideal world but never happens. Usually one side has more.

Currently FW does not use PSR or ELO for match making. The new system will only use those numbers for solo players giving the highest PSR/ELO preference for matching. What is the difference if there are only low PSR/ELO solo players in the que? They will still get matched. The new system matches up all the teams that will fit starting with the largest groups and then fills gaps with solo players starting with the highest PSR/ELO.

There is no real skill matching being done at all. This is really a mute argument. "Skill" is only being used to determine which solo players get matched first. That's it, nothing else. Players in groups don't have their skill checked at all. An all tier 1,12 man will be pitted against an all tier 5, 12 man if that's all there is in the que. The MM isn't even going to look at their "skill", not even once.

[Khobai]

the tier system isnt elo and was never a replacement for elo.

the sole purpose of the tier system was to keep completely inexperienced players from playing experienced players. it was never intended to do more than that. people dont seem to get that.

if PGI adds ELO they should just remove the tier system. there is no need to have tiers anymore if everyone has an ELO score.

switching from tiers to ELO would also basically reset everyone.

Edited by Khobai, 11 April 2019 - 01:25 AM.

[PhoenixFire55]

Khobai, on 11 April 2019 - 01:23 AM, said:

the tier system isnt elo and was never a replacement for elo.

the sole purpose of the tier system was to keep completely inexperienced players from playing experienced players. it was never intended to do more than that. people dont seem to get that.

Hmmmm ... Yeah. Makes sense to call a value determining tiers Player Skill Rating like they did then.

Oh wait ...

[Vellron2005]

Eatit, on 11 April 2019 - 01:02 AM, said:

TL;DR Skill is not being used to create matches. It's only being used to determine the order that solos get added.

There is no real skill matching being done at all. This is really a mute argument. "Skill" is only being used to determine which solo players get matched first. That's it, nothing else. Players in groups don't have their skill checked at all. An all tier 1,12 man will be pitted against an all tier 5, 12 man if that's all there is in the que. The MM isn't even going to look at their "skill", not even once.

I thought they were gonna look at individual players in the group, and then based on their skill assessment (be it PSR or ELO) assign an ELO to the group, then match that against similar ELO groups?

That is how they should do it.. cose' if it's 12-man of T5s Vs 12-man of T1s, then the MM is pointless..

[GotShotALot]

I recall reading a post in the FW section a couple years back, about a way to calculate a 'battle value' type number for a mech in MWO. (Different than TT battle values but that kinda concept) It was pretty straightforward. For instance, the current matchmaker seems to do a decent job on selecting roughly equivalent 'tonnages' for each side - but fails miserably at balancing the 'quality' of each side taking into account the builds, success rate, and quality of the players.

If PGI calculated a 'BV' for each mech which at least took some measure of these stats into account, and then balance for 'total BV' on each side rather than tonnage, it might give better results.

Unfortunately, as others have pointed out... this may have saved FW if it was done 4 years ago when everyone but the "my 12-man must have the ability to fight noob pugs or MWO will fail!" crowd was suggesting it. Seems odd to put in a matchmaker after 80% of the players have moved on.

Edit: Went digging through my old MW bookmarks, I think this was the post I remembered, although I was pretty sure it was in the Faction Play section. Not sure how to quote an old post from different thread so here's the link:

https://mwomercs.com...ost__p__6099489

Edited by GotShotALot, 11 April 2019 - 03:48 AM.

[Alan Hicks]

FP seems almost a dead corpse being tried to be revived, no good if it's going to be unequal matches again.

If they are going to keep the tier system, they should first modify it with a complete overhaul, it's already a mess at QP.

A real working MM which arranges balanced teams is something PGI has yet to achieve.

[Monkey Lover]

The mm in cw is mostly about matching groups with groups. The psr is only really there so you dont put a new player with a group match.

Over all nothing is going to work if you dont have large group of players.