251 replies to this topic

[Mangonel TwoSix]

I keep seeing battle value pop up. That would be a horrible idea. In a perfect world BV would be great. But in reality there are too many moving parts. Is ecm on a fast mech worth more BV than ecm on a slower mech because they can get around the battle space faster where it is needed? Is tag worth more when put on an lrm boat? It goes on and on.

BV would just create anot her system to game. You.could.game BV hard in the table top by carefully selecting mechs and equipment. You could do the same here.

ELO is not terrible. It is terrible for below average players. It is obviously designed so that bad players loose more often. It is pretty simple.

A 12 to 2 match does not indicate an uneven match in regards to skill level. What happens is one side gets up a kill or two, and it snow balls. PGI has put this out there many times.

[Foust]

Christof Romulus, on 20 February 2015 - 10:42 AM, said:

What I don't think many people who are reading this question on the forums understand, and what Russ went into detail explaining is the following:

There are only so many MWO players.

This is not DOTA, this is not League - there are simply not 1m+ people waiting in the queue.

So every 'player choice' we have, divides the possible matches we can have as individuals. Put even more simply - just how fine do you want to split this hair, exactly?

If you have 1000 people looking for a match, 200 of them are 'universal' - able to join any queue, then you have 300 who want skirmish only, that's 500 possible for skirmish, right? Wrong, because there's 250 who are looking for conquest and 250 looking for assault. That 'universal' queue is now stretched mighty thin, because in addition to just those player choices, you have ELO to look at, where it's trying to find (and I do stress the word trying) an equally skilled opponent in all of those match ups.

Add in the ability to toggle maps, and now you no longer look at the above example, but you look at a queue of 200 'universal' players and that gets cut down to 50. 150 of the players who ARE willing to play any game mode, aren't willing to play River city, or Caustic Valley.

In addition, of the 300 who want to skirmish, 120 of them don't want to do it on Alpine.

The hair keeps getting more and more thin as you split it.

Exactly. Fantastic post.

Since I haven't seen it yet I am going to copy Markemp's post from reddit below.

Instead of taking the choices away, give incentives for using less choices. 10% c-bill bonus if you select any game mode. 10% c-bill bonus if you use the lowest represented weight class in the queue. Bump up those bonuses (i.e. make it a flexible, sliding scale) if you need to encourage people to give Any mode a try.

Edited by Foust, 20 February 2015 - 11:00 AM.

[Mawai]

To address the original question ...

Which is better ...

- fragmented queues, matches with the elements you want, larger spread in ELO or
- unified queues, no choice in game elements and better ELO match.

It is impossible to answer this since we (as players) have NO IDEA WHATSOEVER about how significant Elo values may or may not be to actual match outcomes.

Without knowing both the average Elo on each team, the standard deviation of Elo on each team and the recorded match outcomes (at a minimum)... I have no idea how the players can answer this question.

Russ and PGI HAVE the data ... LOOK at it.

Take a day of developer time (or contract it out to someone in the community ... many of whom might even do it for free) to write some quick simulations looking at how the Elo distribution would change with

a ) No queue fragmentation from choice
b ) Allowing full choice of modes and maps (note:re-writing the matchmaking code to allow map choice as well as mode choice is likely non-trivial and someone who selects one mode on an unpopular map could wait a very long time for a match).

Would these changes in expected Elo distribution have any significant effect on expected match outcome?

The typical difference in the group queue is 183 ... how often does the favoured side WIN? If it wins 51% of the time then obviously a 183 point average difference is INSIGNIFICANT.

But without that data the players can't express any sort of meaningful preference.

So PGI ... if you want people to express a useful opinion on this question ... get someone (Karl would be a good choice) to actually provide the numbers on the effect of Elo.

P.S. At the very least plots of Average Elo difference on team vs win percentage and another showing the effect of Elo spread are essential.

Edited by Mawai, 20 February 2015 - 11:05 AM.

[Willard Phule]

Zeece, on 20 February 2015 - 07:51 AM, said:

Do you want full choices of Mode and Maps at the Expense of Less Competitive Matches (because every choice shrinks the size of the pool of players that the Matchmaker can choose opponents from which means it will have to expand the ELO range to fill a match) or

Do you want the most Competitive(closet ELO) Matches but at the cost that you can no longer choose maps or modes?

Reference: Current Solo Queue ELO difference 38... Current Group Queue ELO difference 183

Frankly, I'm a little perplexed at the concept of a "competitive match" in the solo queue. It's the same as it's always been...the better you are, the more you have to carry.

Since there's absolutely no way that's ever going to change, then give us the choice of maps and modes...it's not like it's going to mess things up any worse than they already are.

[Chiron]

El Bandito, on 20 February 2015 - 07:54 AM, said:

I want 8v8 back for Solo-Q, for better hit reg, better FPS, and better matchmaking reasons. Group-Q can retain 12v12.

That is all.

CapperDeluxe, on 20 February 2015 - 08:10 AM, said:

I'll take 8v8 also for solo queue. Would also be nice to allow groups of only up to 2 players into the solo queue.

This. Have two queues, Normal and Ranked. Normal is 8v8, no groups over 2. Ranked is 12v12 and encouraged for groups, but solos/duo's can drop too? What do you guys think?

reddevil, on 20 February 2015 - 08:13 AM, said:

Back to the original question, choice is always important, but it doesn't have to be 0% choice or 100% choice. I would lean towards competitive matches being more important, but in a more 30/70 weight in favour of quality, competitive matches.

Great idea. Currently, I sometimes turn off assault and skirmish, leaving only conquest, since it's my favorite....you know what? I have to wait longer for a match. When time is a factor, I DONT DO THAT, and turn on A & S again. So a compromise between the two seems best.

[ZenFool]

I was astonished to find no lobby where you could pick maps and modes when I first started. I had always thought that was integral to any online game. Now I'm used to it and even enjoy the randomness.

I'm not sure if we can really have a "perfect" matchmaker that gives more 12-10 matches. The nature of the game makes it very difficult to come back after losing a mech or two. If there was some way to improve MM I'd be all for it, but I can't imagine anything on the backend that will make as much of a difference as comms has.

[BARBAR0SSA]

ELO doesn't affect CW at all so no I don't give a crap about it, I'd much rather see meaningful planetary battles with rewards and consequences that encourages solo and newbies to participate. Light scouting missions, heck put in an ELO Restriction mode for CW that only allows low elo players to participate, literally put in a 3 v 3 mode for low elo players only, to learn on and contributes to CW letting them make a difference.

I believe the main issue with new players is the "you sit out now" aspect of solo queue
having a persistent battle would alleviate a lot of the hatred, it sucks starting a match and dying within 2 minutes just to sit there and wait, CW respawn was a step in the right direction (yes the lore people hate it but whatever) you get to play a few times, except it's quite harsh as it's a linear battle corridor rather than an expansive, time sensitive environment where you can't sneak around because you HAVE to stay at omega.

Eg. Say there's a central planet that has 8 planets around it, if each faction controls 4 of those surrounding planets then attacking the central planet is a fair fight, However if one side owns 5 of the 8 surrounding planets then they get an extra 5 tonnes to their drop deck.

I'd love to see a council for each faction as well, voted in by faction players that have been with the same faction for 2 months or whatever, allow us to vote in 8 "clans" to control target planet attacks, build defenses(Future AI implementation), long attrition battles where it's a surprise attack on a base that doesn't go the typical build a 12 v 12. Surprise attack starts an "emergency" broadcast goes out allowing defenders to Hot Drop in to help with base defense(Which should have been built up by the leaders managing planetary income) yada yada more of a interactive CW, make it solo player friendly really just get people all playing it. Let the lower lower skilled players play with the higher skilled and learn from them and so forth.


Really we all had the dream of a persistent planet that we would have to form lances and galaxies, walk a distance towards different controlled cities and engage with limited rear supply lines perhaps getting the defenders to call in reinforcements leaving a smaller target lightly defended allowing a lance to sneak attack it. Or an ECM scout spotting this approaching horde of mechs and setting up an attack from the front and rear. Of course my lovely dream would require TTK to be drastically reduced in order to facilitate longer battles so the long march wasn't done just to have 10 seconds of fighting to lose a mech and respawn at supply camp.

Why don't I care about ELO MM crap:
I like the being out gunned or over gunned battles of CW, sometimes a group of random players can surprise everyone and it's so much more rewarding knowing you just beat the odds. I remember our team chat after complete pug team beat a 10 man on Canyon and everyone was just floored, adrenalin flowing and just that lovely "high" I'd get when I was a 16 year old playing Ultima Online fighting against the odds with a group of friends where team work would make a huge difference. You'd cancel an attack spell that would kill 1 guy, just to heal your friend who was fighting 2 v 1 knowing that if he went down you were most likely done for, both then unload on 1 target trying to change the battle in to 2 v 2.


TLDR; Rather new CW modes and features that encourage solos/newbies to participate in CW with smaller matches letting them learn the game, feel like they are doing something rather than MM doing the 12 v 12. CW allowing 4 mechs is much more fun from a new player perspective, you die you get 3 more chances to play and keep going rather than being told you sit on the sidelines for rest of match, yes you can quit and launch again but then you sit and wait for MM again rather than keep playing and most likely getting a kill eventually.

Edited by shad0w4life, 20 February 2015 - 11:17 AM.

[Aethon]

First, ditch ELO completely. It does not work in games like MWO, and serves only to lengthen queue times.

Then, have two queues: One pug queue, which includes groups UP TO FOUR PLAYERS, with an even number of groups per team. This would be a hell of a lot more balanced than 10-player teams against a smattering of 2-and-3-player groups. 12-player teams would only match against 12-player teams, as now.

This would do the following:
1. Eliminate one of the three non-CW queues, thereby focusing these players on the other two.
2. Not punish players who only have 1 or 2 friends to team with; the current system frequently pits them against 10-player team.
3. Pugs would still play against other pugs; their team's advantage would not change, as both teams would either have even numbers of groups, or no groups at all. No team would have the only group present.
4. Shorter queue times, due to larger playerbase per queue.





Want an even simpler answer?

Give us lobbies, like the MSN Zone during MW4. If players hate the other people in one lobby, they can join another. If the people in the lobby hate the map being played there, they can vote for a map change, or join a lobby playing the map they really want to play. No more getting stuck on River City with a Panther, or Alpine with a Wubgoyle.

To expand on that, if/when the playerbase becomes large enough to support it, add checkboxes for 3025 only, IS only, Clan only, Urbies only, whatever.

In short, put some community-building power back in the hands of the community. Give them a chance to shape the game, and they will enjoy it more, which means more recommendations to friends, larger community, more income for PGI, etc.

[Kristov Kerensky]

For the people advocating BV, do you know the difference between the current Elo system and using BV? NONE. You are simply substituting 1 bad system when it comes to teams with another that totally ignores skill of the players. So 2 high BV Mechs + 10 low BV Mechs get dropped against 12 medium BV Mechs, skill of the players having NO impact on the match up..yeah, and you think 12-0 stomps are fun now? It's an even easier system to game than Elo is, and works even worse for matching due to that.

Elo is not any good for TEAMS, never has been, never will be. PGI isn't the first game maker to try and defend their use of Elo, and just like all the others, it's a failed argument. "current solo que Elo is within 38 points"..sounds good right? Problem is, Elo is based on win/loss record, and with a team game, you can be so horribad that your teammates make jokes about you living as long as 30 seconds. And guess what, you can have a very high Elo despite being that horribad due to the TEAM being better than you ever thought of being.

Someone, ANYONE, explain how Elo works so damn well with that being the reality of it's system. It's a failure in a team game, pure and simple. Sadly, no one actually has a working system for ranking individual players in a team game when you can both play with that team OR with random strangers on a randomly assembled team, so we get stuck with Elo since no one has been able to figure out how to do that. Probably because it's impossible, too many factors at work.

Now, what Russ has actually asked, and he explained this on the Town Hall last night is simple..

We can have tighter Elo based matches IF they remove the options of game mode.
OR
We can have looser Elo based matchs IF they give us more options like map selection.

So, do we want tighter Elo matches which, in Russ's mind means more evenly skilled match ups?
OR
Do we want more options, like map selection, and LESS tight Elo based matches, again, Russ's mind this means less evenly skilled match ups?

I say scrap it all, give us lobbies already, we can sort the rest out, we only did it for over a decade with the previous MW titles, not like it's something we haven't had a massively successful time with.

[Evex]

When PGI took away player choice for a better match maker the public went ballistic. If I remember correctly change didn't even last for a full day before it was changed back to the current matchmaker. Most of the people I talked to during that change said it was the most fun they had in a long time when it came to this game.

While it might not be fun to play in a specific game mode for instance you might not like playing skirmish, or assault on a personal level. By competitive russ most likely means close matches, where you aren't necessarily rolled or the one doing the rolling. If I have to give up choosing a game mode for those sort of matches then so be it. I rather have fun enjoyable matches then deciding what game mode I'm going to play in. Most game modes boil down to kill the other team at some point any way.

[Slimspadey101]

I would me more than happy to forfeit choice of map if that's what PGI needs me to do in order to have solid queue times AND if I can know what map I will be dropping on before I choose my mech. All of the maps in this game are at least playable imo. River City is an exception since it was not designed for 12v12 play. That being said I would like to see the return of 8v8 in some way as well.

[CocoaJin]

Yes, it's a fundamental tenant of society,



"the needs of the many out weigh the needs of the narcissistic, short-sighted whines of gaming brats and man-child basement rats".

[Lordred]

Zeece, on 20 February 2015 - 07:51 AM, said:

Do you want full choices of Mode and Maps at the Expense of Less Competitive Matches (because every choice shrinks the size of the pool of players that the Matchmaker can choose opponents from which means it will have to expand the ELO range to fill a match)

or

Do you want the most Competitive(closet ELO) Matches but at the cost that you can no longer choose maps or modes?

I liked the system where it was a vote for a match, rather then "I want this mode only" I found the 'voteing' system better for the whole 4 hours we had it.



That said, I would also be on the bandwagon for Solo que to return to 8v8

[Codestar]

I would be happy to not be able to select maps if I could:-

Select camo/colours per map.
Loadout per map.
Modules per map.

I would play other gamemodes too if I could have a different setup per gamemode/map.

Some maps just don't work with some builds and most of the time I would have a poor game but sometimes there would be an awesome game in the mix.

Cheers

Codestar

[Deathlike]

Zeece, on 20 February 2015 - 07:51 AM, said:

Russ Asked me to Pose this Questions to the Community during last nights Townhall so that he could get honest and open feedback from y'all on it.

To give context to his question.

Do you want full choices of Mode and Maps at the Expense of Less Competitive Matches (because every choice shrinks the size of the pool of players that the Matchmaker can choose opponents from which means it will have to expand the ELO range to fill a match)

or

Do you want the most Competitive(closet ELO) Matches but at the cost that you can no longer choose maps or modes?

Reference: Current Solo Queue ELO difference 38... Current Group Queue ELO difference 183

It's not as simple as that sounds.

Personally, I'd like my skill/game matching to be close as possible, but that relies on getting on players on the same level of Elo as I am... and I know there's not enough people in that category to be with.

I know that other people want as many choices they can have... whether you believe that is personal preference (a mode sucks) or for personal gain (running sniper builds on brawling maps like Bog is usually a bad idea), people like their choices... and that will always still create lopsided games.

Either way, there's no net gain.

What ultimately needs to happen is to raise the skill floor from the new players through more/better/effective tutorials... basically "learning to play" instead of having to watch teammates or opponents seem like "test dummies" for the better players to feast on. That is the ONLY way to make people happier.

Winning against an ultra-competitive group can be a lost cause... but you getting really bad teammates are more likely to create problems long term... in that they either don't grow and get better... or that it puts back the onus on PGI to "fix something" that is a direct result of not bringing up more players to stick around because of that they need a lot of training to reach "basic competence" in a match w/o trying to denigrate them in the process.

[occusoj]

Zeece, on 20 February 2015 - 07:51 AM, said:

Russ Asked me to Pose this Questions to the Community during last nights Townhall so that he could get honest and open feedback from y'all on it.

So PGI does not even post on their own forums if they have a question to the community? Seriously?

Zeece, on 20 February 2015 - 07:51 AM, said:

Do you want full choices of Mode and Maps at the Expense of Less Competitive Matches (because every choice shrinks the size of the pool of players that the Matchmaker can choose opponents from which means it will have to expand the ELO range to fill a match)

I want fair matches. Do whatever it takes. If theres longer waiting to get into a match its fine with me and I play all maps and gamemodes.

The frequent stomping and well predictable result of a match makes it boring to play. No matter if its a win or loss. If one team gets significantly more "gifted" than the other it results in a bad game.

Besides that, there is no such thing as a "competitive" match in solo.

How about just tossing ELO for a couple of days? Beside weight(class) matching, assemble teams randomly and if time/resources permit add a simple lobby.

[Weeny Machine]

Kristov Kerensky, on 20 February 2015 - 11:18 AM, said:

For the people advocating BV, do you know the difference between the current Elo system and using BV? NONE.

I am sorry but...WHAT?!?!?!

A battlevalue is a fixed value which indicates how much "value" (damage dealing/taking, heat disipation etc capabilities) your mech brings to a battle. It doesn't take into account any skill. Maybe you want to look one up from the TT?

It also takes into account the Clan's sophisticated technology.

ELO doesn't take that into account. If I decide to play a subpar mech for the fun of it or nostalgia, my ELO still considers me to be on my "meta mech". This also goes for weight classes. Ja, the ELO adjust but before that happens you experience "50 Shades of Mech"

Also. ELO doesn't really work because of the limited playerpool
So sorry, you couldn't be "wronger" (ja, there's no superlative hehe)

Edited by Bush Hopper, 20 February 2015 - 11:38 AM.

[Lordred]

Bush Hopper, on 20 February 2015 - 11:36 AM, said:

I am sorry but...WHAT?!?!?!

A battlevalue is a fixed value which indicates how much "value" (damage dealing/taking, heat disipation etc capabilities) your mech brings to a battle. It doesn't take into account any skill.

It also takes into account the Clan's sophisticated technology.

ELO doesn't take that into account. If I decide to play a subpar mech for the fun of it or nostalgia, my ELO still considers me to be on my "meta mech". This also goes for weight classes. Ja, the ELO adjust but before that happens you experience "50 Shades of Mech"

So sorry, you couldn't be "wronger" (ja, there's no superlative hehe)

Only problem with BV, as that clan mechs are currently less powerful then the BV they would be rated for.

[Water Bear]

Mister Blastman, on 20 February 2015 - 07:54 AM, said:

12-2 drops are competitive?

Forced stacking of terribads on your team is competitive?

Assuming that an Elo difference of 80 is a small number...the terribads are all of your team including you, not the other 11 people.

Numerically speaking, of course.

Edit: My answer to the question posted by OP is yes, I would rather have less choice and more balanced death matches.

Edited by Water Bear, 20 February 2015 - 11:39 AM.

[AssaultPig]

I don't think 'stomps' can really be fixed in the pubqueue as the game is currently designed

it's just too easy to quickly lose 2-3 mechs early, and that puts their team at a gigantic, usually insurmountable disadvantage. Sometimes it's not even really those pilots fault; walk around the wrong corner into line of fire of a couple enemy heavies suddenly you've taken a ton of damage. This problem is exacerbated on skirmish/assualt, where there's no incentive for teams to spread out. I suspect that even in games that are perfectly matched (i.e. even tonnage, no huge elo disparity between players, etc) that stomps are still quite common.

This is why I'd like to see dropdecks included in pubqueue games; 'larger' teams means that losing a mech or two to circumstance is less impactful.

I understand why they don't want to implement map exclusion, but I'd like to see exclusion voting similar to what they tried with game mode voting. i.e. all players pick a map to exclude, and at the start of the match those 'votes' are tallied and the top votegetter or two is removed from the random pool.

battlevalue would solve nothing; we're not going to get asymmetric games, so BV matching wouldn't accomplish anything that tonnage matching doesn't anyway. It could also easily make it more difficult to get a game in an under-represented chassis, since the MM would have a more difficult time 'matching' a chassis that few people are driving.

Edited by AssaultPig, 20 February 2015 - 11:48 AM.