30 replies to this topic

[pjnt]

Hi all!

I used to play with a group called KaoS Legion when the game was released and took a 4 year break. Back at it and have about 400 games invested this time round to date.

Got a couple questions about the match maker, since it is really confusing. Also, sorry this will be a repeat post, but the search function doesn't seem to work on the web page, so I couldn't gather all the info on my own.

All comments are solo drops.

I understand the pilot tier system. However, I have seen many T5 pilots up against T1. I have read this is to cut down waiting times at the expense of balanced games.

I tracked my last 37 games since I felt there was a trend to build a sample size and found this: (for this it does not matter if I win or lose, it is the stats that matter)

GMT +1

00:00 - 18:00
12-2
12-2
12-3
12-8
12-4
12-4
12-7
12-2
12-3
12-6
12-9
12-6
12-2
12-6
12-2
12-2
12-11
12-0
12-2
12-2
12-4
12-5
12-3

AVG 12 - 2,9

18:00 - 00:00
12-6
12-3
12-3
12-6
12-4
12-5
12-5
12-9
12-6
12-5
12-5
12-5
12-3
12-9

AVG 12 - 5.8

So there is clearly a time to get better and more competitive games. a 12-3 game is a stomp and a 12-6 game is usually a good grind that tilted at some point where a team takes advantage and cleans up. I suppose this is due to the player population being low...? Thing is if the game can find 24 players, why are there so many blow outs? Every now and then things can go pear shaped and you lose, no problem, but the trend is

A couple pilots that work together make all the difference, from my experience. Best I come up with is all the best players are on in the evening. Thus the chance of having capable people on both sides is increased. It still leaves me confused though that the MM cannot seem to spit these players, even at low use times.

Feed back with proof would be great! Opinions I got piles of while dropping with some old friends, but even after 7 years, the reasoning behind it remains elusive to me.

At the end of the day, I find it very frustrating that in half my games, the outcome is looks set before I power up. I can get 4 kills, do over 1k damage and the score can be 12-4.

Are there 2 different match makers working? One at high traffic and it changes after the waiting times increase?

[sceii]

More skilled teams are, more stomps you will get.
People who know how to play can easily capitalize on small advantages.

[Clownwarlord]

MM doesn't really work.

[Zergling]

Near as I can tell, it tries to match players of a given PSR Tier against opponents of the same PSR Tier... and there may also be a team PSR averaging system at work.

[Bombast]

Generally speaking, MM builds matches by PSR tier, and teams by weight class.

[pjnt]

Clownwarlord, on 20 March 2018 - 04:42 AM, said:

MM doesn't really work.

Kinda got that bit, but it does seem to work for 6 hours each day The games are, for the most part, competitive and fun.

sceii, on 20 March 2018 - 04:34 AM, said:

More skilled teams are, more stomps you will get.
People who know how to play can easily capitalize on small advantages.

Thanks, but why the discrepancy in the time to play?

Zergling, on 20 March 2018 - 04:45 AM, said:

Near as I can tell, it tries to match players of a given PSR Tier against opponents of the same PSR Tier... and there may also be a team PSR averaging system at work.

Which would indicate, we don't really know what is happening. Fair enough. I can limit my play times to 18:00-0:00, at least then I will get good games, for the most part.



Bombast, on 20 March 2018 - 01:47 PM, said:


Generally speaking, MM builds matches by PSR tier, and teams by weight class.


So it gathers similarly tiered pilots into a pile, if there are not enough it goes further from the desired level, and then just carves that in half by tonnage? -stunned emoji-

The data available must allow for a more complex and accurate formula to create a balanced outcome. Using win/loss or kill/death numbers... under this a tier 1 pilot with a 10% win ration could be considered equal to a tier 1 pilot with a 90% win ratio, no?

Edited by xlentx, 20 March 2018 - 04:56 AM.

[UnofficialOperator]

xlentx, on 20 March 2018 - 04:33 AM, said:

I understand the pilot tier system. However, I have seen many T5 pilots up against T1. I have read this is to cut down waiting times at the expense of balanced games.

T5 playing against T1 is impossible. Please show proof.

xlentx, on 20 March 2018 - 04:33 AM, said:

So there is clearly a time to get better and more competitive games.

IMO, this is unfortunately true and server dependent.

xlentx, on 20 March 2018 - 04:33 AM, said:

So there is clearly a time to get better and more competitive games.

IMO, scores like 12-4 is simply 1 aggressive side winning over 1 passive side coupled with the snowballing effect of the 1st kill. Scores like 12-8 or 12-11 are simply 2 passive sides fighting mini battles spread over the map. Scores like these are actually indications that it was a bad match without coordination on both sides.

xlentx, on 20 March 2018 - 04:33 AM, said:

At the end of the day, I find it very frustrating that in half my games, the outcome is looks set before I power up. I can get 4 kills, do over 1k damage and the score can be 12-4.

Sorry I have to call this out man, judging from your wlr, kdr and matchscores, I don't think that is happening in half of your games.

xlentx, on 20 March 2018 - 04:33 AM, said:

Are there 2 different match makers working? One at high traffic and it changes after the waiting times increase?

According to one of the more recent patches, T1 should be fighting T2 at most. They will loosen the MM requirements to T3 I think after long wait times. T4 is fighting T5. As I said earlier, T5 vs T1 is not possible.

[EnochsBook]

Part of the responsibility falls to the MM in that it often fails to balance the best players between the two teams, but the "stomps" we all regularly see are mostly an effect of the game mechanics themselves.

In the average solo queue match it's not uncommon to see one team group up and just roll all over the enemy team as they attempt to nascar or otherwise split up. In that situation there's not much the other team can do: they will kill one or two of the first team's 'Mech that are leading the assault and get overrun very quickly.

[pjnt]

UnofficialOperator, on 20 March 2018 - 04:55 AM, said:

T5 playing against T1 is impossible. Please show proof.

We were sinq dropping with a new guy were are bringing into the game and were happy he was with us... I will grab a screenshot next time and post, np. Lets leave it until proven.

Proof https://mwomercs.com...r-how-it-works/ is the official guide on the PSR? states :

[color=red] it's hard to promise that Tier 1 players will never meet Tier 4 or Tier 5 players.[/color]

UnofficialOperator, on 20 March 2018 - 04:55 AM, said:

Sorry I have to call this out man, judging from your wlr, kdr and matchscores, I don't think that is happening in half of your games.

Call away man! Check my archived stats and you will see my play, they are good enough for most. Once I stopped, I handed over the game to my son who had fun for some time. My numbers are going up quick now that I am back in the seat and my kid has moved on to Overwatch


So it comes down to skill, and skill alone. Play at a time when the better players are on. Can do.

Thanks all for the info.

Edited by xlentx, 20 March 2018 - 05:36 AM.

[MrMadguy]

Explanations here. Everything else is self-evident: MM tries to match players with closest ratings according to 3/3/3/3, ±2 Tiers and may be equal tonnage rules, blah blah blah. Problem is - it doesn't work at all.

Edited by MrMadguy, 20 March 2018 - 05:28 AM.

[PocketYoda]

Zergling, on 20 March 2018 - 04:45 AM, said:

Near as I can tell, it tries to match players of a given PSR Tier against opponents of the same PSR Tier... and there may also be a team PSR averaging system at work.

Not from what i've asked in chat almost every match.. i always get a response of between tier 5 and tier 1 in the same team every match no fail, there is no way that many people are lying, and if they are why lie for no reason at all..

xlentx, on 20 March 2018 - 05:22 AM, said:

We were sinq dropping with a new guy were are bringing into the game and were happy he was with us... I will grab a screenshot next time and post, np. Lets leave it until proven.

Proof https://mwomercs.com...r-how-it-works/ is the official guide on the PSR? states :

[color=red]it's hard to promise that Tier 1 players will never meet Tier 4 or Tier 5 players.[/color]



Call away man! Check my archived stats and you will see my play, they are good enough for most. Once I stopped, I handed over the game to my son who had fun for some time. My numbers are going up quick now that I am back in the seat and my kid has moved on to Overwatch


So it comes down to skill, and skill alone. Play at a time when the better players are on. Can do.

Thanks all for the info.

I see tier 1s and tier 5s all the time in my teams don't worry about unofficial operator hes full of it.. He would not believe you even if you do prove it.. The tiers are all mixed these days and match maker is busted ***..

Edited by Samial, 20 March 2018 - 05:56 AM.

[pjnt]

Samial, on 20 March 2018 - 05:53 AM, said:

Not from what i've asked in chat almost every match.. i always get a response of between tier 5 and tier 1 in the same team every match no fail, there is no way that many people are lying, and if they are why lie for no reason at all..

Yup, all the tiers are mixed, my experience, too. I don't think they want it this way, but there isn't a large enough player base for the MM to generate games with any reasonable wait time that would be balanced.

In a few hours of gaming I will come across the same players all the time. In other similar games but with large player bases, I never saw anyone, ever again in solo drops.

From what I am reading everything is functioning properly, there just are not enough players for the system to gain traction and do what it was intended to do.

Luckily for MWO fans, MW5 will arrive soon and the players WILL love it and WANT to play more.... a couple Google searches later and there will be an large influx of players here. We will have to embrace them as a community and not STOMP them into another game... Hopefully, this is realized and this opportunity is not wasted.

[Ghogiel]

xlentx, on 20 March 2018 - 06:03 AM, said:

there will be an large influx of players here. We will have to embrace them as a community and not STOMP them into another game...

We won't have any choice in the matter.

CW: by design it is about farming solo noob players right out of the mode if not the game in the most lopsided, no match maker, stacked team vs solos you can make. Plus the way the meta game is designed sucks.

Group: More competitive than CW due to having a MM and not allowing solos, but the MM here doesn't even look at PSR so mostly sucks.

QP: this is where 90% of players will always exist. due to some remote semblance of competitive games, ie it has a MM, and the direly poor other modes.

solaris: well there is a MMer. and looks to be the best one MWO ever had. it is 1v1 and 2v2. And I'm not sure people in this game are into that. Plus the balance is going to be whacky with little room for experimentation in each tier after the metas filter down to the goyim. but fact is when thier face gets run through over and over and over again, a lot of people ain't going to be able to take that. So I think it will be niche and kinda boring after some time.


Also as for your original post, it's well known when to farm solo queue if you want bigger and better stats. GMT afternoon is pretty good for potato farming.

Edited by Ghogiel, 20 March 2018 - 07:03 AM.

[Duncan Aravain]

Yes, the mm is broken and, sadly enough, probably has always been broken since it relies on a larger player base and a flawed tier system in determining teams. It also has a nasty history were it tends to place the highest skilled players unequally; one side having 2 or more players that are " competitive level" skilled verses the other side. When the player base is so small that you can recognize at least 4-6 players per match and their skill level, you start to see how truely flawed the mm is at the present. The algorithm that mm uses needs to be modified, but PGI has both manpower and programing skill problems as well as other issues that need to be addressed. Certain times of the day highlights these issues, as you found out, and events will also bring out some unequal matchmaking despite the increased player pool. A guess would be the return of players who have not piloted for awhile having their temporarily rusty skills grandfathered in at a higher evaluated level. Whatever the reason is, you either have to understand and game the mm the best you can (i.e., certain times) or just sigh and keep hitting the launch button.

[El Bandito]

MM works better during prime time.

[pjnt]

Ghogiel, on 20 March 2018 - 06:57 AM, said:

We won't have any choice in the matter.

CW: by design it is about farming solo noob players right out of the mode if not the game in the most lopsided, no match maker, stacked team vs solos you can make. Plus the way the meta game is designed sucks.

Group: More competitive than CW due to having a MM and not allowing solos, but the MM here doesn't even look at PSR so mostly sucks.

QP: this is where 90% of players will always exist. due to some remote semblance of competitive games, ie it has a MM, and the direly poor other modes.

solaris: well there is a MMer. and looks to be the best one MWO ever had. it is 1v1 and 2v2. And I'm not sure people in this game are into that. Plus the balance is going to be whacky with little room for experimentation in each tier after the metas filter down to the goyim. but fact is when thier face gets run through over and over and over again, a lot of people ain't going to be able to take that. So I think it will be niche and kinda boring after some time.


Also as for your original post, it's well known when to farm solo queue if you want bigger and better stats. GMT afternoon is pretty good for potato farming.

Duncan Aravain, on 20 March 2018 - 07:04 AM, said:

Yes, the mm is broken and, sadly enough, probably has always been broken since it relies on a larger player base and a flawed tier system in determining teams. It also has a nasty history were it tends to place the highest skilled players unequally; one side having 2 or more players that are " competitive level" skilled verses the other side. When the player base is so small that you can recognize at least 4-6 players per match and their skill level, you start to see how truely flawed the mm is at the present. The algorithm that mm uses needs to be modified, but PGI has both manpower and programing skill problems as well as other issues that need to be addressed. Certain times of the day highlights these issues, as you found out, and events will also bring out some unequal matchmaking despite the increased player pool. A guess would be the return of players who have not piloted for awhile having their temporarily rusty skills grandfathered in at a higher evaluated level. Whatever the reason is, you either have to understand and game the mm the best you can (i.e., certain times) or just sigh and keep hitting the launch button.

Good honest answers. Cheers.

I didn't want to say this in the OP to avoid trolls and fanboyz, but what you are saying pretty much bullseyes my thoughts over the last week of trying to get back into the swing of things.

[yrrot]

To say that PGI is lacking programming ability to create a more robust matchmaker is nonsense. They already have it, they've done it, and the results were drastically longer matchmaking times--especially in off-peak times--and lots of other complaints. Solo players complained about being matched with groups. Small groups didn't like having to meet the 3/3/3/3 requirements to matchmake. Big groups didn't like limited group sizes.

The end result: a matchmaker that tries to only roughly approximate skill-based matching in solo queue for shorter wait times, and group queue that only uses tonnage limits and group sizes to match up people to fill teams. And...stomps still happened. FW matchmaking has been crazy since they initially rolled it out.

Solaris is kind of the ideal setting to use Elo based matching, especially since they are breaking each player down to elo ranks in each division.

[Jackal Noble]

It on average groups/stacks more better players on the other team.

Edited by JackalBeast, 20 March 2018 - 08:58 AM.

[Duncan Aravain]

yrrot, on 20 March 2018 - 08:33 AM, said:

To say that PGI is lacking programming ability to create a more robust matchmaker is nonsense. They already have it, they've done it, and the results were drastically longer matchmaking times--especially in off-peak times--and lots of other complaints. Solo players complained about being matched with groups. Small groups didn't like having to meet the 3/3/3/3 requirements to matchmake. Big groups didn't like limited group sizes.

The end result: a matchmaker that tries to only roughly approximate skill-based matching in solo queue for shorter wait times, and group queue that only uses tonnage limits and group sizes to match up people to fill teams. And...stomps still happened. FW matchmaking has been crazy since they initially rolled it out.

Solaris is kind of the ideal setting to use Elo based matching, especially since they are breaking each player down to elo ranks in each division.

While I agree that previous PGI attempts to improve mm have failed, their current incarnation has not succeeded,either. Tarogato's thread statistically examining some stomps yield some interesting statistics and strongly hinted that certain simple changes in the algorithm of the mm might yield significantly more evenly matched contests. Having the current mm depend so heavily on Tier ranking, an evaluation not based on actual skill atm, and mech tonnage classification (no BV) results in many lopsided matches. As so many other posters have noticed and posted in the forums, when you are frequently being able to judge the upcomming match's outcome just by looking at the team's roster in the dropship, you begin to wonder why PGI can't get their mm to be more acurate.
The frequency where the truely elite players available in a specific match are unequally devided between the two teams is astounding; one's views on how broken the mm is currently broken might be partially based on what side you end up on. A stomp is not an indication of a closely matched contest (although it may be true), but probably an indication of a failure in some part of the mm algorithm or assumptions. PGI holds the mm algorithm secret; its lack of ability everyone knows.

[Helsbane]

MM works on three basic elements. Blindfold, swivel chair, dartboard...