47 replies to this topic

[AjerWerklWerkl]

People keep referring to the MM putting together teams with very high ELOs in order to balance our very low ELOs. So, a team would range from like 2000 down to 1200.

But I would expect that because there are a decent # of people always trying to play, it's a lot more likely that MM would be putting together teams that are much more closely bunched in terms of ELO... So I would expect a team to be more like 1200 - 1600, and that they would be matched with another team whose players are similarly ranged 1200 - 1600.

I would think the case where the ELO disparity is high on each time would come up only when the number of people online is very small, no?

It would be nice if the devs could clarify and potentially put the topic to rest.

Thanks!

Edited by AjerWerklWerkl, 09 November 2014 - 02:02 PM.

[StraferX]

If you see me in game you can count that your ELO must be very high to balance out mine. Seems like I always get placed with nubs that have no ecm and scatter all over the map while the other team has 8 LRM boat masters with pin point corelok.

All kidding aside, is ELO even real?

[Durant Carlyle]

The matchmaker has a base range. However, the longer you are waiting to be connected to a match, the larger that range becomes, as it just looks for anybody and everybody to fill out the teams.

[Redshift2k5]

Also it's not just about population total, many players/groups will only play in one mode (or two modes with one exclusion).

It can be the case that there is a good Elo match for your team/group, but you are in Assault only and the good match is set to Skirmish only! In this case the team will never meet, even if all other criteria would otherwise be a good, even match. They tried making mode a flexible choice but it did not sit well with the community.

Playing with all modes selected can offer better, faster matching because you are reducing hard separation.

[Glythe]

This game had the most number of satisfied players when there was no Elo.

They implemented a 1 vs 1 score keeping method and used that to judge 12 vs 12 confrontations. It doesn't work well.

[MischiefSC]

Glythe, on 09 November 2014 - 03:41 PM, said:

This game had the most number of satisfied players when there was no Elo.

They implemented a 1 vs 1 score keeping method and used that to judge 12 vs 12 confrontations. It doesn't work well.

No, apparently you liked it when it was all farming noobs and that's great for you. Hey, many of us miss winning 18 out of 20 matches every time we played in a twisted, should-probably-just-be-playing-single-player-games sort of way.

Elo works just fine for its job. That's why there are people with high Elo and low Elo scores in the game. if Elo didn't work then you wouldn't have good players with high Elo and bad players with low Elo...

Or are you saying that the best players in the game have really low, or average, or randomly determined Elo?

Please present the math to back up both your assertions that A) more people enjoyed pre-Elo play than post-Elo play and B ) Elo doesn't work in a multiplayer game.

I'm especially eager to see the mathematical support for B ), as it is effectively disproving the entire field of statistical analysis.

Not understanding something is not the same as something not being able to be understood. Not getting how something works is not the same as something not working.

[Abivard]

Glythe, on 09 November 2014 - 03:41 PM, said:

This game had the most number of satisfied players when there was no Elo.

They implemented a 1 vs 1 score keeping method and used that to judge 12 vs 12 confrontations. It doesn't work well.

Elo was designed for chess,

In chess there is only one variable, the player.

In MWO there is only one constant, the player.

Now why would anyone think it would still work... magic maybe?

[Iskareot]

Electric Light Orchestra

But yeah for real... Jeff Lynne is the bomb

[LordKnightFandragon]

Id have to vote the biggest issue with elo is how it slaps us with an arbitrary elo rating, despite what our own actual skill level is in game....

If they used like avg dmg, WR, avg xp, KD and other things and gave us a more accurate elo, matches might be a little better.....as each player in the match would be more equal......

current elo is like slapping a buck private with the rank and responsibility of General of the Army and then him going out and getting it all killed cuz he really has no idea what hes doing......but hey, your out of boot camp now, heres your rank!!

elo should be 0 for our first 25 games, then it uses w/e our stats are after that to assign us a proper skill level based on how we played.....then we either stay the same or suck worse or improve afterthat.....and we get matched up accordingly.

[Iskareot]

Actually you hit it on the head.... sadly pitting a bad or low level player with a good or experienced one will never work well. It is actually damaging if you ask me.

[Vasili]

All I know is over the past weeks in general (wider problems going back months) the only games I've seen barring a few rare treats are 12-3 one side or another (mostly against my side, it seems) and it's driving me crazy.

What kind of a matchmaker lets every match be you get curbstomped or curbstomp someone else?

:/

[Roland]

AjerWerklWerkl, on 09 November 2014 - 02:00 PM, said:

But I would expect that because there are a decent # of people always trying to play

I think we have found your error.

[RiggsIron]

Just started playing again after a year.

Lose about 90% of matches.

People scatter. Die one by one. Dont back up team. Body block you when trying to get out of a narrow box while taking fire from multiple mechs. etc etc.

Even random chance should mean once in a while I get the GOOD team? more than 1 in 10 games?

Just finished played for 4 hours. going through about 6 mechs trying to get basics done.

3 of the 6 mechs STILL didnt get the first win of the day bonus. Four. Hours.

Started playing again mostly from watching NGNG videos. Games looks fun.
...so why do I keep playing 9 out of 10 games with glue sniffers?

And holy crap I forgot how NON-fun playing this game is until you have elite done. Its a completely different game.

[PhoenixFire55]

Just yesterday (Sunday, lots of people playing) during the event (yeah, LOTS of people playing) we had a "highly competitive" 6-man going. We had a couple matches in a row where MM put another "highly competitive" 6-man on our team while filling the other team with 12 "casuals". But thats just how MWO MM works, it is built to give you, me and everyone a W/L ratio of 1.0 and in order to do so it creates matches like these, when one team is supposed to lose no matter what just because they are by far less skilled / dedicated (whatever you want to call it), and another team is supposed to win. MM could have easily put the two 6-mans into opposite teams but it didn't. Matches in question ended up 12:1 on Alpine and 12:0 on Mining. Balanced? Yes, because PGI can always point you to some mythical stats that show that most players have a W/L ratio of around 1.0. Statistics is such a nice thing after all, you can twist it however you like ...

[Willard Phule]

Sigh.

Look back through all the patch notes, man.

The matchmaker was re-written at the time of the clan release. That's when they went to a "team average" Elo base as opposed to individual scores as it was prior.

It works well for the group queue, not so much for the solo queue. In the solo queue, if you've got a high Elo, you can expect to have some low ranked players on your team so the team's average will be the same as the other team's.

It is what it is, man. I feel your pain. It's been like this since the beginning in one form or other. It simply doesn't matter how good you are, what your Elo is or anything like that. If you're going to play in the solo queue, you're going to end up with new players.

I have a hunch that it's designed this way on purpose. PGI can't provide any meaningful new player experience..no tutorials or anything like that...so they throw them into general population and let us train them. For free, since we don't get a "trainer" bonus or anything like that.

[Egomane]

Abivard, on 09 November 2014 - 05:29 PM, said:

Elo was designed for chess,

Elo was initially designed for chess. Since then Elo has been adjusted for many, many tasks, including group matchups.

There are various Elo formulas, adjusted for the tasks they are used for. It is not one formula to fit them all.

Repeating that this system should not work, because it was initially invented for chess, just shows, that you do not understand the concept of it at all. Ballpens where invented for space missions, they still work in a lot of different environments and for multiple tasks.

Willard Phule, on 10 November 2014 - 04:48 AM, said:

Look back through all the patch notes, man.

The matchmaker was re-written at the time of the clan release. That's when they went to a "team average" Elo base as opposed to individual scores as it was prior.

They did not change it. This was always the case, but you still misunderstand the implemented system completly.

If a group drops, in the group queue, the average Elo of the group is used, to match this group with other groups of similar Elo values. In the single player queue, every player is his own group and the matchmaker will try to match him with players with similar values.

The matchmaker will, for two minutes, try to create a match with all playing instances (24 players in the single player queue or a varying number in the group queue) having a +/- 50 Elo difference between each other. Only if he doesn't manage to do that, will he open some release vents and increase that range. 100 after two minutes and 1200 after five.

It's not: We drop twelve random single players in a group, average the Elo and then match it against another group of twelve random solos where we did the same.

That's how Karl Berg, the one who wrote the matchmaker, described how it works.

Edited by Egomane, 10 November 2014 - 05:52 AM.

[LordKnightFandragon]

PhoenixFire55, on 10 November 2014 - 03:25 AM, said:

Just yesterday (Sunday, lots of people playing) during the event (yeah, LOTS of people playing) we had a "highly competitive" 6-man going. We had a couple matches in a row where MM put another "highly competitive" 6-man on our team while filling the other team with 12 "casuals". But thats just how MWO MM works, it is built to give you, me and everyone a W/L ratio of 1.0 and in order to do so it creates matches like these, when one team is supposed to lose no matter what just because they are by far less skilled / dedicated (whatever you want to call it), and another team is supposed to win. MM could have easily put the two 6-mans into opposite teams but it didn't. Matches in question ended up 12:1 on Alpine and 12:0 on Mining. Balanced? Yes, because PGI can always point you to some mythical stats that show that most players have a W/L ratio of around 1.0. Statistics is such a nice thing after all, you can twist it however you like ...

so it picks which players it will permanently ensure stay below 50%, those it will keep at 50%, those it will let excel and those it will ensure are way below avg in WR.......

cuz balance.

[Joseph Mallan]

Iskareot, on 09 November 2014 - 06:45 PM, said:

Actually you hit it on the head.... sadly pitting a bad or low level player with a good or experienced one will never work well. It is actually damaging if you ask me.

How do you think a bad player learns what good looks like?

[AjerWerklWerkl]

Joseph Mallan, on 10 November 2014 - 05:41 AM, said:

How do you think a bad player learns what good looks like?

Well, just like any other game or sport, you should generally be playing with people of similar (and ideally, slightly higher) skill level... you LEARN by watching videos on YouTube, or spectating the best folks at your level after you die.

Creating wide-disparity-in-ELO matchups is not the solution to learning.

Egomane: thank you for your informative response! That is exactly the point I was trying to bring to light when this thread started. I hope everyone reads what you said!!!

[Joseph Mallan]

Egomane, on 10 November 2014 - 04:58 AM, said:

Elo was initially designed for chess. Since then Elo has been adjusted for many, many tasks, including group matchups.

There are various Elo formulas, adjusted for the tasks they are used for. It is not one formula to fit them all.

Repeating that this system should not work, because it was initially invented for chess, just shows, that you do not understand the concept of it at all. Ballpens where invented for space missions, they still work in a lot of different environments and for multiple tasks.



They did not change it. This was always the case, but you still misunderstand the implemented system completly.

If a group drops, in the group queue, the average Elo of the group is used, to match this group with other groups of similar Elo values. In the single player queue, every player is his own group and the matchmaker will try to match him with players with similar values.

The matchmaker will, for two minutes, try to create a match with all playing instances (24 players in the single player queue or a varying number in the group queue) having a +/- 50 Elo difference between each other. Only if he doesn't manage to do that, will he open some release vents and increase that range. 100 after two minutes and 1200 after five.

It's not: We drop twelve random single players in a group, average the Elo and then match it against another group of twelve random solos where we did the same.

That's how Karl Berg, the one who wrote the matchmaker, described how it works.

Does that group Elo match up factor attrition into the equation??? Cause THAT is what throws Elo out the window. With every death on the field, Elo changes in this game.