156 replies to this topic

[Deathlike]

Dread Render, on 12 August 2016 - 11:16 AM, said:

I think the part that you are not getting is the W/L ratio is just a gauge of how well the MM is working...
NOT HOW GOOD YOU ARE. Your skill rating is for that, not the W/L ration...

Good players should be matched against other good players. which means their W/L ratio should be about 1to1
Bad players should be matched against other bad players. which means their W/L ratio should be about 1to1

again... the W/L of every player is really just a gauge of how well the MM is working for YOU.

You do realize there's no actual way a bad player will get close to a 1:1 K-D ratio if they do close to nada right?

You can't make potatoes shoot at the same target, and then expect them to stop a base cap in Assault. They won't do it... or they don't know how.

You're spouting unrealistic expectations of people.

[Dread Render]

Deathlike, on 12 August 2016 - 11:19 AM, said:

You do realize there's no actual way a bad player will get close to a 1:1 K-D ratio if they do close to nada right?

You can't make potatoes shoot at the same target, and then expect them to stop a base cap in Assault. They won't do it... or they don't know how.

You're spouting unrealistic expectations of people.

Who is talking about K/D ratio? not me or the OP.
Learn to read please.

[Mister Blastman]

Dread Render, on 12 August 2016 - 11:16 AM, said:

I think the part that you are not getting is the W/L ratio is just a gauge of how well the MM is working...
NOT HOW GOOD YOU ARE. Your skill rating is for that, not the W/L ration...

Good players should be matched against other good players. which means their W/L ratio should be about 1to1
Bad players should be matched against other bad players. which means their W/L ratio should be about 1to1

again... the W/L of every player is really just a gauge of how well the MM is working for YOU.

I think you are confusing should with how things really are.

Good players REALLY DO win more than average! This is fact. You can't dispute this. Not with how the game is structured right now.

[Mawai]

Mister Blastman, on 12 August 2016 - 10:58 AM, said:

If you had stats in college then you'd understand that an average is comprised of a set of numbers in a range that has a minimum and maximum value...

And right now the average win/loss ratio is 1.09 (most likely due to group queue disparity), the median is 1:1

-and-

You can get 1.09 with a set of numbers ranging from 0.2 all the way up to 2.2 with a undeclared mix between those two poles.

Does that make sense?

This means that there WILL be some players with an average win/loss ratio of 1.3, 1.4, 1.5, 1.6 and even 2.0 and this is not because the matchmaker is broken, it is because they are good at the game.

The statistics support this.

You will never have a matchmaker that fits everyone in at 1:1 and if you do somehow manage to create one, you will see a large exodus of the player population because frankly good players don't like to lose half the time and bad players like to think they should win more than half the time.



This for sure. If the system was weighted on a moving bell curve it would help weed out the bads and get rid of the "progress bar."

Just a couple of comments ...

1) An average or mean has nothing to do with minimum or maximum values ... it is simply the sum of the numbers divided by how many numbers there are. In this game W/L can go from 0.0 to Infinity (assuming you have zero losses ).

2) The average is not 1.09 due to group queue ... why? The reason is that for every person who wins in the group queue, there is a person who loses. Yes, there are some folks who can pad their stats because they play with well coordinated teams in the group queue ... individually they move up on the W/L scale. However, W/L (unlike PSR) is conservative so you get as many wins as losses tallied (with a few draws thrown in).

3) I completely agree that there will be folks with high W/L and folks with low W/L ... that is the basic nature of a distribution where folks have different skill levels.

4) "You will never have a matchmaker that fits everyone in at 1:1 and if you do somehow manage to create one, you will see a large exodus of the player population because frankly good players don't like to lose half the time and bad players like to think they should win more than half the time."

Totally wrong . The reason you don't see 1:1 for all players is that matches have to be made in a finite length of time and the player rating system is not perfect. In a perfect rating system there are always far fewer players at the top and bottom of the distribution than in the middle. The top players are often grouped with ones who are not so good while the worst are grouped with better players. These natural variations can shift the odds of winning. In addition, if the player is better than their rating they SHOULD have a W/L > 1. The problem with PSR in this regard is that it is capped. Folks reach the tier 1 cap and can't go any higher. They play lots of games against folks with similar ratings but the ratings don't reflect skill. Over time the pool of players at that rating grows and the actual skill broadens and folks at the top find the matches worse and they they can have a greater impact since they contribute more and can thus maintain a W/L > 1.

However, if you have a good rating system AND enough players at the top and bottom of the range then the W/L WILL tend toward 1.0 since you will be matched against players that are JUST AS GOOD as you are and who ALSO EXPECT to win more than 50% of the time. It is simple statistics ... if the matchmaker is working then W/L will tend to 1:1 with decent match quality.

P.S. Elo was a conservative rating system but was based solely on W/L (it resulted in a bell curve type distribution as expected) while PSR is non-conservative and is mostly based on damage done with a W/L bias (as a result, I expect that by this point the distribution is FAR from a bell curve).

[DAYLEET]

PORONOPAPOS, on 12 August 2016 - 06:03 AM, said:

Or maybe there is some kind of black list of players that MM puts them in the more likely to lose team ?

I know you cant create a black list yourself at least. You can block people and they will still be put in your team.

I block all the uninspired ******** who copy/paste the exact same sentence every single match. I do wait many matches before blocking them incase they also communicate useful info during match(never happens). I also keep getting put with people i report for intentional FF or Non Participation.

[Deathlike]

Dread Render, on 12 August 2016 - 11:21 AM, said:

Who is talking about K/D ratio? not me or the OP.
Learn to read please.

The same applies to W-L.

People who work at wanting to win their matches tend to win them.

People who don't even know how to steer their mech... probably won't win much.

When you're expecting people to have a 1:1 W-L ratio, the MM can't really guarantee that. It doesn't guarantee good players (frankly, there are not enough of them), nor does it guarantee good builds and too many other variables you realistically cannot control.

The best one can control is their own destiny. The better players steer how the match goes, and it leans in favor of the people willing to work together.

MM or no MM, you, regardless of the handicaps, are more or less responsible for the results of the match... even if there are days when you aren't doing much. Eventually, the game expects you to do something good... and that's again.. up to you.

On the other hand, the way PSR works guarantees that "not totally bad" players keep moving on up... for reasons unknown and undeserving.

Edited by Deathlike, 12 August 2016 - 11:30 AM.

[Dread Render]

Mister Blastman, on 12 August 2016 - 11:22 AM, said:

I think you are confusing should with how things really are.

Good players REALLY DO win more than average! This is fact. You can't dispute this. Not with how the game is structured right now.

No, I am not... Yes that is how things are. But that is an error.
Good players should be matched against other Good players and therefor have a W/L ratio of about 1to1.
SO They REALLY Should Not Win More than Average.
That would indicate the MM is not working.
but you are finally starting to see that. Good for you Big winner guy... you sure told me.

Edited by Dread Render, 12 August 2016 - 11:34 AM.

[Mister Blastman]

Dread Render, on 12 August 2016 - 11:30 AM, said:

No, I am not... Yes that is how things are. But that is an error.
Good players should be matched against other Good players and therefor have a W/L ratio of about 1to1.
SO They REALLY Should Not Win More than Average.
That would indicate the MM is not working.
but you are finally starting to see that. Good for you Big winner guy... you sure told me.

So we're talking wants and needs here.

You want everyone to win 1:1.

You think it is wrong that good players win more than 1:1.

You also stated somewhere else that good players can't change the outcome of their games...

Which is it?

[Mister Blastman]

Mawai, on 12 August 2016 - 11:28 AM, said:

Just a couple of comments ...

1) An average or mean has nothing to do with minimum or maximum values ... it is simply the sum of the numbers divided by how many numbers there are. In this game W/L can go from 0.0 to Infinity (assuming you have zero losses ).

2) The average is not 1.09 due to group queue ... why? The reason is that for every person who wins in the group queue, there is a person who loses. Yes, there are some folks who can pad their stats because they play with well coordinated teams in the group queue ... individually they move up on the W/L scale. However, W/L (unlike PSR) is conservative so you get as many wins as losses tallied (with a few draws thrown in).

To have an average you must have a set of numbers. For every set of numbers there is a minimum value and a maximum one, unless they are all the same number--which they aren't.

By your logic there should be a pure 1:1 average as for every win there is a loss. How come the average win/loss is 1.09 then?

PORONOPAPOS, on 12 August 2016 - 11:45 AM, said:

1- as I wrote before the problem in my main account is not that I lose (I can live with this). The problem is that from late June till now most of my defeats were stomps ! that never happened before. For me that indicates a problem in the way MM buids teams recently. I never asked or wanted to be in T1 witch I reached last December, then went full T1 around mid or late January. All this time and until June I kept playing with basically same people, my mediocre performance never before led to so many stomp defeats.So I don't understand what changed now. If my performance is the reason so many players droping with me end a match with less than 100 damage(in assaults) then why this didn't happen (at least so often) before June ?

Also consider the player population has dropped considerably since last December, so the release valves are being hit. When you're a higher tier, you get screwed harder by the matchmaker and more lower-tier bads are put on your team, thus it expects you to carry harder.

When release valves are blown, you start seeing tier 5s play vs. tier 1s (which they say shouldn't happen but forum reports say otherwise).

[Lostdragon]

Dread Render, on 12 August 2016 - 11:30 AM, said:

No, I am not... Yes that is how things are. But that is an error.
Good players should be matched against other Good players and therefor have a W/L ratio of about 1to1.
SO They REALLY Should Not Win More than Average.
That would indicate the MM is not working.
but you are finally starting to see that. Good for you Big winner guy... you sure told me.

Again, this is an unrealistic expectation. The top 10% of players can't only be matched with each other all the time because they are not always online at the same time and wait times would be horrible. The reality is there were only 34000 people who played 10+ matches last month, which is a relatively small population. You have to play with and against people of different skill levels to get acceptable wait times and distributions of mechs. The MM tries to make that fair by hitting a target PSR number for each team. A really good pilot may wind up on a team with several players of much lower PSR against a team of players that all have a PSR in the middle of those two. This is necessary for the system to work and release valves still get triggered all the time to facilitate timely match making.

What this means is that the good player is basically expected to carry harder because without him his team would on average be lower skill than the opponents. If he is truly as good as his rating he will indeed carry at least 50% of the time and maintain a w/l of 1 or greater. If he is truly an amazing player he will carry most of the time and have an exceptional w/l.

Matchmaking is a delicate balance between skill disparity and wait time. Compromises have to be made in that wider ranges of skills have to play together to make sure wait times are not too high, especially when populations are low like during the morning.

[Revis Volek]

Dread Render, on 12 August 2016 - 06:40 AM, said:

There is NO WAY you can prove this or even produce evidence for your Assumption.
If you are placed in a bad team, It does not matter how good you are, you cant carry an entire team.
You may feel you due and your ego loves to believe you do... BUT YOU DONT
The MM is broken and needs to be fixed.

And you have evidence to prove your counter argument?

NO you dont, but FYI you are extremely wrong. If this is the case please explain how I have managed to get myself up to almost 1000 more wins then losses?

I surely cant be credit to my team according to your hypothesis. Its the fact i have been carried over 800 times?

Get real bro and get gud.

Edited by Revis Volek, 12 August 2016 - 11:55 AM.

[Lostdragon]

Read this for a great dev post on MM in Overwatch, specifics are a bit different but general principles the same:

http://us.battle.net...45504371#post-3

[Roughneck45]

This thread is just... wow. So much crazy its hard to fathom.

Players have no impact on the outcome of the match? Really?

Like were all algorithms worshiping our god the MM striving for 1.0 ratios.

Stomps are the norm. Any game with a single life per round and a large pool of HP that cannot be healed are this way. Its just the nature of the beast. Once one domino goes down its much easier for the rest to fall.

[Dread Render]

Mister Blastman, on 12 August 2016 - 11:43 AM, said:

So we're talking wants and needs here.
You want everyone to win 1:1.
You think it is wrong that good players win more than 1:1.
You also stated somewhere else that good players can't change the outcome of their games...
Which is it?

All i am saying is if the MM was working correctly everyone would have a W/L ratio of close to 1:1 because they would be playing against people with the same skill level... great vs great, good vs good, okay vs okay, bad vs bad....
And that's what We Want isn't It? ... that's what I want.. that's what the OP wants. I think.
for the majority of players a close game is a good game,
I am complaining about stomps really...
If the MM was working well, there would not be so many.

[Lostdragon]

Dread Render, on 12 August 2016 - 12:00 PM, said:

All i am saying is if the MM was working correctly everyone would have a W/L ratio of close to 1:1 because they would be playing against people with the same skill level... great vs great, good vs good, okay vs okay, bad vs bad....
And that's what We Want isn't It? ... that's what I want.. that's what the OP wants. I think.
for the majority of players a close game is a good game,
I am complaining about stomps really...
If the MM was working well, there would not be so many.

Yet more unrealistic expectations. Stomps are inevitable in a pvp game with no respawns because of Lostdragon's Law of Weight of Fire which says the first team to get a kill is most likely to win the match and more than two unanswered kills at the start of the match will probably result in an overwhelming victory.

Comebacks are very rare in MWO because once you are losing by two mechs the enemy has a much easier time flanking you or just rolling over you en masse because they have significantly greater firepower than you.

Many matches come down to who successfully rotates the vanguard. A 12-0 match can be very close with all the mechs on the winning team beat to hell and about to pop. Knowing when to push, when to hold, and when to step up and take some fire off your teammates wins matches in MWO.

[Dread Render]

Okay I HAVE It!!!!
Imagine every game was 1vs1
If the MM was working your W/L ratio would be about 1to1 Right?
if it was higher... that means the MM is giving you opponents that are to easy.
if lower, the MM is giving you opponents that are more skilled than you are.
well... If coded well the MM should work the same in this game.
that's all I'm saying..."ALL"
I am NOT and Never said you cant effect the outcome of a game or there are not some very good players.

Don't you want even games? That's what I want and that's what the OP wants also.

[Kirkland Langue]

Dread Render, on 12 August 2016 - 12:00 PM, said:

I am complaining about stomps really...
If the MM was working well, there would not be so many.

Not true.
As a team loses mechs, it becomes easier and easier for the enemy to focus fire. If all players were equally skilled and all mechs were perfectly balanced, there would be far more stomps than there are currently. Once one team got any kind of advantage, only a disparity in skill or power would be able to overcome that advantage.

[Lostdragon]

Dread Render, on 12 August 2016 - 12:13 PM, said:

Okay I HAVE It!!!!
Imagine every game was 1vs1
If the MM was working your W/L ratio would be about 1to1 Right?
if it was higher... that means the MM is giving you opponents that are to easy.
if lower, the MM is giving you opponents that are more skilled than you are.
well... If coded well the MM should work the same in this game.
that's all I'm saying..."ALL"
I am NOT and Never said you cant effect the outcome of a game or there are not some very good players.

Don't you want even games? That's what I want and that's what the OP wants also.

Do you mind waiting an hour to play a 10 min match? If no, then this works, if yes then players of different skills have to play together.


Edit: clarity

Edited by Lostdragon, 12 August 2016 - 12:21 PM.

[Mystere]

Dread Render, on 12 August 2016 - 10:03 AM, said:

"griping about the MM is just a distraction from the real issues"
What Real issues. MM has the most effect on every player every game.
It should be the 1# issue.

Ahem! MM is not the #1 issue with MWO.

Hint:

[Kubernetes]

Dread Render, on 12 August 2016 - 11:30 AM, said:

No, I am not... Yes that is how things are. But that is an error.
Good players should be matched against other Good players and therefor have a W/L ratio of about 1to1.
SO They REALLY Should Not Win More than Average.
That would indicate the MM is not working.
but you are finally starting to see that. Good for you Big winner guy... you sure told me.

A match can be filled with 24 Tier 1s, but that doesn't mean all 24 are of equal skill. The MM only knows PSR, and PSR uses big buckets. You might have one monster and 23 good players. That one monster might do the work of 3-4 good players, which is why that monster likely has a yuuge W/L. It would be nice if PSR could rate players better, but it wouldn't make much difference with the MM because the population is too low to only match monsters with monsters.

Also to the OP, it could be the time of day you come on. I had you in a match yesterday in my T4 alt. There was at least one other T1 I recognized, but it was mostly populated by the T4-T5 people I'd been dropping with for hours. Seems like there were plenty of us, but not enough T1s to give you a proper match.