134 replies to this topic

[Mechteric]

Captain Stiffy, on 22 October 2013 - 10:34 PM, said:

I'm starting to think that 8v8 was way better

It was ~50% more forgiving of mistakes, so in that sense it was "easier". I think the game mode of assault is the problem, in that it wasn't meant for 12v12. Conquest on the other hand is better, but only on larger maps. Forest Colony and River City are simply too small for 12's.

My opinion is they should modify the max number of players based on game mode at the least, not sure what to do about the maps though.

[Felio]

I thought they said they tightened the Elo range so you didn't have a couple of awesome players balancing out a bunch of bads.

[Kunae]

Felio, on 23 October 2013 - 06:28 AM, said:

I thought they said they tightened the Elo range so you didn't have a couple of awesome players balancing out a bunch of bads.

Tell that to the new players in trial-mechs, on my side in matches.

[PEEFsmash]

This would be absolutely hilarious. It would be the most viscous pugstomps of all time. I'm betting that I would have a 98% win rate.
Top-players know that Elo is working...we play against eachother every single game. There is never a game anymore that I don't know someone in, and most of the time I know most of the players on both teams. Then, I will match up with them again and again and again all night, and fail to find match if they stop queueing.

People who think Elo is flawed/isn't working are just mid level players in the "soup" where players aren't very consistent. That's unfortunate, but turning off Elo would just mean top premades just absolutely crush the common folk 12-0 every match, instead of the current situation where top premades face eachother all night long. The latter is preferential for all parties involved, trust me.

Edited by PEEFsmash, 23 October 2013 - 07:06 AM.

[mike29tw]

PEEFsmash, on 23 October 2013 - 07:06 AM, said:

This would be absolutely hilarious. It would be the most viscous pugstomps of all time. I'm betting that I would have a 98% win rate.
Top-players know that Elo is working...we play against eachother every single game. There is never a game anymore that I don't know someone in, and most of the time I know most of the players on both teams. Then, I will match up with them again and again and again all night, and fail to find match if they stop queueing.

People who think Elo is flawed/isn't working are just mid level players in the "soup" where players aren't very consistent. That's unfortunate, but turning off Elo would just mean top premades just absolutely crush the common folk 12-0 every match, instead of the current situation where top premades face eachother all night long. The latter is preferential for all parties involved, trust me.

For top players yes it would be hilarious. For the rest of us, nope it will be like every day, steamrolling/steamrolled every other match and occasionally a 10/12 close game.

No noticeable difference whatsoever.

I'm starting to think a tiered system will be better for us now.

[Kunae]

PEEFsmash, on 23 October 2013 - 07:06 AM, said:

This would be absolutely hilarious. It would be the most viscous pugstomps of all time. I'm betting that I would have a 98% win rate.
Top-players know that Elo is working...we play against eachother every single game. There is never a game anymore that I don't know someone in, and most of the time I know most of the players on both teams. Then, I will match up with them again and again and again all night, and fail to find match if they stop queueing.

People who think Elo is flawed/isn't working are just mid level players in the "soup" where players aren't very consistent. That's unfortunate, but turning off Elo would just mean top premades just absolutely crush the common folk 12-0 every match, instead of the current situation where top premades face eachother all night long. The latter is preferential for all parties involved, trust me.

You are over-stating this, as others have.

It would not change hardly anything, other than make the chance of a newbie on your team random, rather than by design.

[Feetwet]

I am not a top tier player but I do pretty well. If what peef and the rest of these "top tier" players are saying is true, I want out of the "soup". Last night I was ON, just tearing stuff up. Probably my best night in a while. Lost nearly every match. At contact with the enemy, half of the team would evaporate...puff...dead (fortunately some would fire their weapons). I don't care how humbling an experience it would be...make me the worst player in the top tier...please.

S

[Joseph Mallan]

PEEFsmash, on 23 October 2013 - 07:06 AM, said:

This would be absolutely hilarious. It would be the most viscous pugstomps of all time. I'm betting that I would have a 98% win rate.
Top-players know that Elo is working...we play against eachother every single game. There is never a game anymore that I don't know someone in, and most of the time I know most of the players on both teams. Then, I will match up with them again and again and again all night, and fail to find match if they stop queueing.

People who think Elo is flawed/isn't working are just mid level players in the "soup" where players aren't very consistent. That's unfortunate, but turning off Elo would just mean top premades just absolutely crush the common folk 12-0 every match, instead of the current situation where top premades face eachother all night long. The latter is preferential for all parties involved, trust me.

So you just admitted you are not, in fact, a top tier player! Cause if you are not in the top 2% You are not one of the best. You are just really good.

[FerretGR]

Kunae, on 22 October 2013 - 05:36 AM, said:

Assertion: The Elo system is very flawed, polluted with stats from both grouping and solo, and with their mechanic of balancing high-Elo with a proportion of very low Elo, in matches, to "even things out", it hurts the game more than it helps.

I would suggest that PGI turns off Elo for one week or patch-cycle so everyone can see how it plays. Just use a rough weight-class matching, and let it go at that.

It is my belief that the resultant matches will be the same or better than ones using Elo. I would just like to be able to test this theory.

You're wrong about how Elo matchmaking works (high-Elo and low-Elo are not combined in some mix/average Elo, but rather, a target Elo is set and players are pulled in who are close to that Elo). Any problems you see are because of small pools of players and a variety of other things and not because of Elo matchmaking failing. If you're seeing lots of newbies, it's likely because there aren't enough players close enough to your Elo at the time you're playing to consistently fill the teams. That's not Elo's fault, that's the nature of the playerbase, and randomizing matchmaking won't change that.

Is Elo matchmaking perfect? No. Is it better than random? Absolutely. You complain about 7000+ game players being thrown in with beginners... how is randomizing the pot going to improve this? Right now, perhaps there's a chance that it'll happen. Randomizing will simply increase that chance. How does that make anything better? Use logic: by what mechanism will randomizing matchmaking plunk less newbies and more experienced players in your team?

Anyone who says "it can't be worse than it is now"... are you new to the game? It was hella worse before Elo. I've had a much improved experience since Elo.
Look, if you don't think Elo is working, make a new account and hop in a trial and proceed to haand everyone their *** for 25 or 30 games. Trust me: the beginner Elo bracket is different than the one you're sitting in, Kunae.

[Mister Blastman]

Ghogiel, on 22 October 2013 - 07:01 AM, said:

I'm cool with it. I'll just go back to having 10 KDRs and W/L like back before Elo.

I miss those fanfare days of gallivanting around in my Dragon Chariot, lashing my electro-laser whip upon the scrubbery and peppering their foreheads with my holy srm-sprinkler. The peasants were kept in check as my knighthood rose in glory. I adorned my steed with numerous bloody heads that I gleefully stacked atop my antennae; oh my, it was so gory!

Sigh. I must lament these days gone bye. The fanfare. The revelry. The screams of horror and blood-curtling smash of metal and burning hulks. Oh the smell of charred carbon and wilting fluids. The limbs. The ashes. The crushed cockpits and strung-out bodies within. Such sadness befalls my great chariot, evermore stashed away in my stables, never to be used again.

Edited by Mister Blastman, 23 October 2013 - 08:24 AM.

[Murphy7]

I really do think the best of Matchmaker was when there was weight-class balancing, as attrocious as that could be when your sides' three Awesomes were matched against their three Altas-DDC's. Mod had a thread with the weekly drop composition, so you could attempt to ensure your 8 person premade got dropped against another 8 persone premade following the weekly composition.

Part of what may have helped that work was that our mech selection was far more limited.

ELO is apparently an individual's rating determined by their W/L ratio in a team game, which makes me questions the veractiy of ANY individual's ELO score.

[Jman5]

Kunae, on 22 October 2013 - 05:36 AM, said:

Assertion: The Elo system is very flawed, polluted with stats from both grouping and solo, and with their mechanic of balancing high-Elo with a proportion of very low Elo, in matches, to "even things out", it hurts the game more than it helps.

I would suggest that PGI turns off Elo for one week or patch-cycle so everyone can see how it plays. Just use a rough weight-class matching, and let it go at that.

It is my belief that the resultant matches will be the same or better than ones using Elo. I would just like to be able to test this theory.

That is NOT how it works! You don't have a system bringing in scrublords to balance out the high elo players. It just does not work that way. The way matchmaker works is this:

1. One player or group hits launch and the matchmaker creates a game. It uses that initial game creator to act as the base for the Elo it's searching for.
2. Using that player or group's (averaged) Elo as the base. It then begins searching for players AS CLOSE TO THEIR ELO AS POSSIBLE.
3. While it is searching it is also doing two other things. It is see-sawing premades of similar Elo into the two different teams so one team doesn't have all lone wolves. It is also trying to get weight somewhat close. However it's not as tight as it used to be here because it's favoring speed and splitting up premades over weight consistency. Premades that run either super heavy or super light have a tendency to mess up the weight balance.
4. As time passes and the matchmaker struggles to find people at or near that base ELO number it slowly expands and loosens it. (but only to a point)
5. After both teams are assembled, the matchmaker compares the two team's combined Elo. If one team has a higher Elo number than the system projects them to win.
6. If the projected winners win, Nobodies Elo moves much. If the projected winners lose, than the underdogs have their Elo go up and the losers' Elo goes down.

The matchmaker's biggest problem is weight mismatches, which happens mostly when premades run really heavy or really light. Once Min/Max weight caps get introduced for premades, this problem is going to largely disappear.

Edited by Jman5, 23 October 2013 - 08:40 AM.

[Kunae]

FerretGR, on 23 October 2013 - 08:06 AM, said:

You're wrong about how Elo matchmaking works (high-Elo and low-Elo are not combined in some mix/average Elo, but rather, a target Elo is set and players are pulled in who are close to that Elo). Any problems you see are because of small pools of players and a variety of other things and not because of Elo matchmaking failing. If you're seeing lots of newbies, it's likely because there aren't enough players close enough to your Elo at the time you're playing to consistently fill the teams. That's not Elo's fault, that's the nature of the playerbase, and randomizing matchmaking won't change that.

No, I am absolutely correct about how matchmaking works, you are just misunderstanding. Within the ranges that are set, which grows as the "search" runs, it will balance a side that has a high Elo player/s on it with the bottom of the ever widening range, to reach the "average target" that it's looking for.

FerretGR, on 23 October 2013 - 08:06 AM, said:

Is Elo matchmaking perfect? No. Is it better than random? Absolutely. You complain about 7000+ game players being thrown in with beginners... how is randomizing the pot going to improve this? Right now, perhaps there's a chance that it'll happen. Randomizing will simply increase that chance. How does that make anything better? Use logic: by what mechanism will randomizing matchmaking plunk less newbies and more experienced players in your team?

Right now it is adding those low Elo players, by design. I would prefer the chance to be random, rather than mandatory.

Also, there are newer players stuck in "Elo hell", who are not bad players. They are just punished by having their contributions not be able to significantly alter a battle. They also have minimal chance to learn how to be better, and to meet better players to learn from, if they're never exposed to them in any numbers.

FerretGR, on 23 October 2013 - 08:06 AM, said:

Anyone who says "it can't be worse than it is now"... are you new to the game? It was hella worse before Elo. I've had a much improved experience since Elo.
Look, if you don't think Elo is working, make a new account and hop in a trial and proceed to haand everyone their *** for 25 or 30 games. Trust me: the beginner Elo bracket is different than the one you're sitting in, Kunae.

The "pre Elo days" suffered from strict weight-class matching, which allowed groups to fugger up what they'd be facing. As an example, you could take 4 Jenners, and be matched up against 4 solo random commandos and ravens. They would almost have to have loose, no, or minimal weight-class matching for this to work, or it could be gamed as it was before.

I also never said I was experiencing the beginner bracket. It's merely that players in or just out of that area are being dropped into the same matches as people playing at a moderately high level for over a year.

Just to reiterate and clarify, I'd just like to see them let us test this for a week or two, so we and they could work off real information rather than speculation.

[Roland]

PEEFsmash, on 23 October 2013 - 07:06 AM, said:

This would be absolutely hilarious. It would be the most viscous pugstomps of all time. I'm betting that I would have a 98% win rate.
Top-players know that Elo is working...we play against eachother every single game. There is never a game anymore that I don't know someone in, and most of the time I know most of the players on both teams. Then, I will match up with them again and again and again all night, and fail to find match if they stop queueing.

People who think Elo is flawed/isn't working are just mid level players in the "soup" where players aren't very consistent. That's unfortunate, but turning off Elo would just mean top premades just absolutely crush the common folk 12-0 every match, instead of the current situation where top premades face eachother all night long. The latter is preferential for all parties involved, trust me.

Dude, I get matched against your guys constantly. I think you're mistaken in your assessment about who it's not working for. I think that in the middle of the Elo, it probably works better, since there are more folks who it can choose from.

Up at the top, I suspect that there aren't enough players in that bracket to fill up 24 players at any given time.

So then it becomes a roll of the dice, in whoever the matchmaker ends up picking to fill in the rest of the slots. And it's still picking folks who seemingly have never played before... folks who never actually hit R and target mechs, or other various things that makes them clearly very new players.

It's not that Elo isn't doing anything.. because you're right, it's clearly TRYING to put folks together that are as closely matched on skill as it can manage. But there don't seem to be enough players to actually do that.

And when it does that matching, but ends up with a huge tonnage disparity, then it's simply unfair.

Again, if you have to choose between matching on skill, or matching on tonnage, THEN YOU SHOULD CHOOSE TONNAGE. Because honestly, even if you matched the skill perfectly, but not tonnage.. then what? Then one team is basically gonna lose just because they were outtonned. That's not even remotely fair, if your skill matched that of the other team, but you lost simply because they had more firepower and armor.

[Kunae]

Jman5, on 23 October 2013 - 08:37 AM, said:

That is NOT how it works! You don't have a system bringing in scrublords to balance out the high elo players. It just does not work that way. The way matchmaker works is this:

1. One player or group hits launch and the matchmaker creates a game. It uses that initial game creator to act as the base for the Elo it's searching for.
2. Using that player or group's (averaged) Elo as the base. It then begins searching for players AS CLOSE TO THEIR ELO AS POSSIBLE.
3. While it is searching it is also doing two other things. It is see-sawing premades of similar Elo into the two different teams so one team doesn't have all lone wolves. It is also trying to get weight somewhat close. However it's not as tight as it used to be here because it's favoring speed and splitting up premades over weight consistency. Premades that run either super heavy or super light have a tendency to mess up the weight balance.
4. As time passes and the matchmaker struggles to find people at or near that base ELO number it slowly expands and loosens it. (but only to a point)

While "perfect unicorn world" may act that way, that's not how it's working in matches. Both of your underlined "points" are wrong. The range is much wider than you'd imagine.

And as for #3? It's not see-sawing anything. It's placing one on a side and then trying to fit one on the other side, and then balancing with solo's Elo's to make it all work out. Weight matching is barely a factor at all.

The broken Elo system is king.

[tuffy963]

to OP:
Seems like a reasonable assertion. Could be true.

Unfortunately, your solution is probably not a good fit for where the game is today. This would have been more reasonable for a game carrying the "Beta" tag, but might produce to disruptive an experience in a production environment.

[John MatriX82]

I'd like to support OP's idea. I can't stand to watch the "searching" icon for half an hour (some days ago I stood there for nearly 45' minutes!!!) before giving up because the game can't find a spot for any class of the mechs I own, so I give up until I can't get grouped.

Once grouped the wait time is still humongous, but at least sooner or later I can make a damn match (which most of the times completely is off the weights because of the huge search time because elo opens up the weight and elo tolerance).

I wonder who has an active premium account how can be happy with such a system and given I still have my two months of Founder PT, I guess I'll basically never activate them whatsoever nor I plan to ever invest in PT.

[LordVanquish]

Well, believe it or not, the Elo IS working to a certain extent. It does try its best to not match up top tier players with bottom tier players since maybe the Launch patch.

Proof of this:
We sometime run something called a "chaos drop," Everyone in the room hits ready at the same time solo. We get placed in matches randomly with random teams. The thing is, a few of us are always matched together, and a few are never with us and matched together in their game.
It is also pretty clear that some of the players are more top-tier and the others are lower-tiered.

So it IS working to an extent. However, the matchmaker will broaden the Elo range of the players the longer the game has been searching as well. Yes, sometime I still get ROFL stomps either one way or the other, it happens even at perfect Elo match making.

Sometimes I get idiots on my teams, but like PEEFsmash says, I almost always recognize some or most of the players now in my bracket, and these are the people that I run into a lot in 12v12 premades as well.

Having said all that, I wouldn't be opposed to test out your idea. The only issue with this is to balance out the 4man premades and those PUG drop-syncing scum as well. If i can make a good 4-man premade and have the rest of the teams are random, we will stomp much much more than we will lose.

[TexAce]

First educate yourself how the MM works at the moment, before trying to make up theories....

[Sarsaparilla Kid]

Bagheera, on 22 October 2013 - 10:41 AM, said:

Yeah. They really, really need to change that setup and start matching players who are closer together. Sure, if there were actually a voice client for Randoms and the experienced players could communicate effectively with the new players, it might work - but it would really be better for people to play "in-bracket," whatever that would mean. Maybe that's not possible for <Reasons>, so basing the matchmaker for Randoms on mech rather than player might be the way to go.

Frankly, the first dividing line for the matchmaker should be "Does mech have any L2 tech installed, Y/N?" Any single instance of DHS, Endo, Ferro, Artemis, etc and the player gets bumped into the "Yes" bracket. It's a simple split that would prevent trial, stock, or otherwise not "optimal" builds getting matched against/with builds bristling with L2 tech, extra mobility/firepower/armor, likely mastered with Mr.7000 matches behind the mouse.

I agree with an approach like the above...there needs to be some way at first to divide the player base by what technology they are using, or even how many matches they've played, before taking a look at their ELO rating, so that new players running trial mechs that start off with a mid-range ELO don't end up dragging everyone else down, including those that have played thousands of matches, have mastered mechs, but still manage only a 1:1 win/loss ratio for failing to carry a team hard enough. Currently, it's like watching a miniature version of today's economy, where the rich keep getting richer and the poor and middle-class keep getting poorer. If there was a more efficient way to communicate with randoms ingame, that would help speed up the learning curve for everyone, as a team is only as strong as its weakest links.

Drawing a parallel to level-based MMORPGs, you generally cannot group with someone who is more than, for example, 10 levels above you, because they have better gear (technology) and stats (due to better technology) and can take on bigger/badder enemies, and that's something they've earned over time, in a linear progression that moves forward with each win, but not backward with each loss. Actually, I believe at the Launch Event, they mentioned a leveling system of some kind up to level 60 or thereabouts (starting to sound more like a MMORPG) to be introduced with Community Warfare, so it will be interesting to see if one's level determines with whom they are matched, as either an alternate to or in conjunction with the ELO system.