64 replies to this topic

[Appogee]

CrockdaddyAoD, on 29 January 2015 - 07:50 AM, said:

PGI has mentioned in the past that the HIGHER ELO matches tend to go one way or the other. You likely in your low ELO got swept up into a match with many HIGH ELO players on each side.

In my experience, at least in the solo queue, you get only a couple of high Elo players on each side, and the rest are relatively low Elo players. By adding up the Elos and dividing by 12, the matchmaker stupidly thinks this makes for an "average" team.

Then the match is decided by which team's low skills players marched into the enemy guns and died first. The imbalance in firepower becomes very hard to recover from once one team is ahead 3 kills.

Edited by Appogee, 30 January 2015 - 12:58 AM.

[PhoenixFire55]

Meta 2013, on 29 January 2015 - 05:15 PM, said:

Reason: ELO is a system designed to keep you challenged by trying to drive your win loss ratio to 1. That does NOT work for a team. This is a system for individuals, and was not developed for large teams of players.

Oh it works perfectly fine even with groups. Half the matches it makes you an uber team of premades vs bunch of 2 man scrubs, other half you are on the team with scrubs yourself. Challenged? Yes. Half the time you go vs impossible odds. W/L ratio ~ 1. Again, yes. You roflstomp one match then get roflstomped the next match. As I said, working as intended^TM.

Meta 2013, on 29 January 2015 - 05:15 PM, said:

ELO does work, if you really look at your win loss record, and how you really play, the Elo system works, for win/ loss, that does not mean it provides good matches, just means it works for driving win/loss record.

And I wish more people will actually start to realize this.

Meta 2013, on 29 January 2015 - 05:15 PM, said:

I've asked before...and will again, we need some tags for folks who have reached a certain number of drops ... ie like 1000, 2500, 5000, 10,000, 25,000 ..etc so we can see the experience levels, and know why we probably got roll stomped.

Won't do any good as is. You can play 25.000 matches and still be a baddie or you can have a talent for it and play like a pro after 500 matches.

Meta 2013, on 29 January 2015 - 05:15 PM, said:

When you sit back and wait 3, 4 minutes for a match, and your in a light mech at 8% then you keep seeing the same folks in all the drops, you realize MM has such a little player base to choose from, its almost impossible to get a good match.

Its only true in higher Elo tier.

Meta 2013, on 29 January 2015 - 05:15 PM, said:

I play 2 different accounts ... one with a win loss ratio just a tad less than 1 to 1, the other account has several hundred wins more than losses, with each, class elo's have been calculated, one account is not remotely close to he other account for light class mech elo's.. yet I see the same players no matter which account I play.

The latter reason is probably 70% of the reason your getting roll stomps, not enough players to choose from. Once ina while you see a match with a few older players on ... for an event, and then suddenly you start getting good matches for awhile.

Unless we know how Elo is calculated you can't say anything for certain. For all we know your Elo scores might be identical. For starters it might actually take your performance into account, thus measuring your skill rather than your luck of getting decent teammates. And even if it doesn't, those several hundred more wins on one account could have been vs weaker than average opponents, while several hundreds less wins on the other were against a stronger opposition, thus leading to same Elo scores.

Meta 2013, on 29 January 2015 - 05:15 PM, said:

Lastly,not being able to choose the map you playing on, an blindly picking a mech,that you hope gets the right map match up, can also lead to those roll stomps.

Not necesserily. If a Team A which is supposed to be far better than Team B ends up with a brawler deck on Alpine it just might lead to a far closer game than expected. Its all really really random, and that is why it is frustrating, the average match quality (due to vast amount of random factors like this that MM doesn't account for) isn't any better/different from what we had with random matchmaker 3 years ago. Only difference is that it takes forever to find a match nowadays.

Meta 2013, on 29 January 2015 - 05:15 PM, said:

Speaking of that, why are we getting Terra Therma and the bog all day long again, with a few crimsons for good measure, been a week since I've seen alpine.

Because whiney babies don't like Alpine. Alpine takes guts to charge/hold the hill or brains to avoid it. Viridian is a new map with an increased chance to drop as derp-o-devs are gathering data about it (which they'll ignore later anyway like they always did, I mean, there is a spot on RiverCity you can get stuck in that they know about since 2012, and its still there ...). It is frustrating that you do seem to have "map of the day" very often, but if you check your stats in the long run all maps appear with more or less same frequency.

All in all ... I've said that before that if the teams are really closely matched like the devs claim MM does with like 95% chance, then there is no chance in the world that the match will end up with a 12-0 stampede. When a skilled team is matched vs a skilled team you don't see 'lance-charlie-assaults-got-left-behind-and-died-terribly' situations. If both teams are skilled teams then both teams will focus fire and even if one does so better than the other, when 12 mechs are shooting smth there is no chance they won't destroy at least 5-6 opposing mechs. And visa versa for closely matched not-skilled teams, your usual PUG derps will happen, but they'll happen in both teams.

Edited by PhoenixFire55, 30 January 2015 - 01:20 AM.

[PhoenixFire55]

xe N on, on 25 January 2015 - 01:46 AM, said:

I feel that hit detection is again quite bad. This could also explain the current imbalance.

Hit reg has always been awful. However, it doesn't matter in terms of balance. Chances one team is filled with guys who have excellent hit reg while other team has really poor hit reg are VERY slim. And even if that happened, there is a 50-50 chance that the "better" team will have worse hit reg compared to a "worse" team having better hit reg ... i.e. hit reg disbalance balancing out the initial MM disbalance. Visa versa match can become less balanced, however such hit reg disbalance between teams happens rarely and even if it does, its 50-50 chance to improve/worsen match balance. This simply can not explain the vast amount of matches that end up 12-0 / 12-1.

[PhoenixFire55]

Nightshade24, on 29 January 2015 - 03:10 AM, said:

I think the main thing that occurs in a 12 - 0 (in smaller group matches and pugs) that if someone dies very early and fast then the game has a disadvantage (11 verse 12 for eg... also dc's and afk's hurt as well0

First of all, for a 12v12 one missing person is practically nothing. Second, if somebody "dies very early" it only means he is a bad player, his Elo score the MM used should and probably does reflect that. That means that the remaining 11 players should be about equal to the other team 12 guys in terms of balance, and the "die early" guy is meaningless for game balance. Now, on the contrary, if we assume that one person dieing early does matter and the result of the match is affected by his early demise, then its either MM failing to balance a match using avaliable Elo scores including his, or his Elo score does not reflect his "contribution" to the team properly. Either way, the only conclusion here is that the Elo-MM doesn't work.

D/Cs and AFK'ers are a different thing. Losing a high Elo player due to a D/C will undoubtfully hurt the team chances. However, if a guy does have a high Elo it means that he rarely D/Cs (otherwise he wouldn't be able to have a high Elo score), thus the match where he D/Ced is simply one particular unfortunate match and have nothing to do with trying to explain 12-0 stomps that happen all the time. Also, nowadays you can rejoin the match and most of the time people are safely able to do so before the actual engagement starts. If on the other hand a guy is a frequent AFK'er, then his Elo score should reflect his contribution and be somewhere around zero, i.e. he is meaningless.

Nightshade24, on 29 January 2015 - 03:10 AM, said:

For instance you are in a cicada and woops, turned the wrong corner, 2 king crabs, an atlas, and a centurion found you. over 150 damage flying to your face in a split second.

Bad scout is bad.
A good scout always understands where the main enemy force is or where it potentially can be given their speed and elapsed time.
If they expected you and you didn't expect them its your fault. Bad player.
If they didn't expect you either, then they won't be able to insta-kill you. If they did, you are a bad player.

Nightshade24, on 29 January 2015 - 03:10 AM, said:

This can happen a second time or 2 at the same time (lets say it was a cicada and a locust that turned the bad corner or some other situation.).

If your Locust buddy also somehow managed to die while they were killing you he is even a worse player than you are.
(just in case, I don't mean you as in YOU, only using it as a reference to a hypothetical Cicada pilot).

Nightshade24, on 29 January 2015 - 03:10 AM, said:

There is many variables that can lead to easy kills, which lead to a disadvantage and a easy win, most of the time though the 12 remaining enemy mechs after winning often have quite a few mechs 1 shot away from death, this is why you see 10 or 11-0's often as well, as say a single few medium lasers TK one guy due to his situation or killed him by contact damage or one of your guys doing a sparta charge to get 1 of the damaged guys before dying.

True. Question remains tho, why are easy kills only on one team and not on both? If some people are just easy kills, why are they balanced with not-so-easy kills on the other team?

If teams are really balanced, the 12-0 just can not happen. Not in the current solo/group queue without comms. It is impossible.

[PhoenixFire55]

Wesxander, on 29 January 2015 - 08:15 AM, said:

Match maker is screwed up from several perspectives:

1) Ranks players by ELO only, which is worthless, when you consider that ELO requires a consistent team or player to work correctly. Example NFL 12 man part of the team does not change to much via offense or defense. Therefore ranking football teams by ELO works as the team stays mostly the same, even one injury only slightly modifies that. Solo play in say tennis ELO would work. In games where you get a new 12 man team every match just about ELO is worthless. It has some value for novice player but I would wager by 1,500 or so games the player is veteran, and will have learned the game well enough that ELO does not truly represent their skill at that point.

Yes and no. MM does not use Elo scores as in "team-Elo" scores. The team Elo is simply the summ of players Elo scores, unlike the NFL example you brought up, where it is still a team, i.e. people constantly playing and training together. In terms of Elo 2+2 does not necesserily equals 4, there are different mentalities, players who prefer to push hard and charge in often and players who prefer to sit back and let the enemy come at them. Obviousely different people might or might not work well together when on same team. Same goes for mech builds.
Obviousely a tag-narc-equipped scout and an LRM boat will work well together, but same scout will be wasted with a peek-a-boo sniper etc.

Wesxander, on 29 January 2015 - 08:15 AM, said:

2) ELO does not factor in weight class of mech when deciding. Hey that 85 ton Stalker to ELO is the same as 100 ton Direwolf on the other side. Anyone will tell you the stalker is out matched but ELO doesn't think so. In fact this is what is sometimes causing the 12 and 0 stats games we see a lot. Hey cicada your up btw I (MM) forgot to mention I am putting you against a clan SCR with bap should be good game you and him have the same ELO score btw......

True this. And its actually very easy for the devs to make Elo score variant-dependant (or at least chassis-dependant), not class-dependant. IMO it will improve things greatly, but I guess they can't be bothered with MM that is already working as intended^TM, 12-0 ggclose.

Cicada and Streak-Crow is a bad example because its rather complicated. Streak-Crow is only effective against targets up to 40 tons. If the Cicada happens to be the only such mech on his team, it only means that SCR will be utterly ineffective against all other targets, which will eventually compensate the possible easy early kill of that Cicada (if it comes to that at all). On the other hand, SCR does not have an ECM slot and can't inpoint-backstab enemy assaults all that well.

DireWolf vs Stalker is a better example, but I'd rather use smth as Dire vs Awesome. In case of equal player skill there is no denying the fact that Dire will contribute more towards the win. Stalker tanks and boats too well to write it off so simply.

Wesxander, on 29 January 2015 - 08:15 AM, said:

3) ELO does not care about special equipment: ECM. BAP, TAG. I have seen many matches were one side gets literally 3 ecm mechs and 2 BAP mechs while the other side gets none, or if they are lucky one ECM mech to counter the enemies 4 or 5. Hardly balanced.

Again, yes and no. ECM does not stop bullets. BAP and TAG do not increase the accuracy of bullets, so each of these items is meaningless by itself. In solo queue chances of all the ECM-BAP-TAG guys getting into a team with all the LRM-SSRM guys are rather negligible. Having 24 random people in a match more or less evens things out. In group queue, you can have such disbalance happening tho, because for example one 4-man can bring a pair of scouts and a pair of LRM boats, while an opposing 4-man will bring (for whatever reason) flamer-MG Hunchbacks. Elo score can't do anything about such team 'chemisrty' and 'derp-a-mech' builds.

Wesxander, on 29 January 2015 - 08:15 AM, said:

MM does not appear to balance tonnages either.

Tonnage alone means nothing, its is how you use it. Had a match on Viridian where my team had lots of fast lights and mediums, we were out-toned by about 120-150 tons. Those guys knew what to do and just kept running around map harassing slower opponents who had trouble navigating the 'you-can-stuck-here-even-if-you-think-you-can't' swamp. If anything, sometimes extra tonnage is a liability. Does not happen too often tho as the only tactics teams use in PUGs is 'big-blop-of-death-goes-counter-clockwise-merry-go-round'. Used to be different in conquest and assault modes before dear devs started messing with basecap speed and dumba$$ turrets.

Wesxander, on 29 January 2015 - 08:15 AM, said:

Clan mech regardless of ELO should be considered at minimum 10 to 25 tons more than they are rated.

Not with IS quirks. No.

Wesxander, on 29 January 2015 - 08:15 AM, said:

Expert pilots with huge win records should have 10 to 15 tons added to their weight range when facing a guy who is only played 100 games to their 3000.

The fact that a pilot is an 'expert' is (or should be) already reflected in his Elo score. Just like his overall experience should be reflected by the Elo score or whatever score used to determine his skill. Tonnage has nothing to do with this.

Edited by PhoenixFire55, 30 January 2015 - 04:51 AM.

[PhoenixFire55]

Meta 2013, on 29 January 2015 - 05:15 PM, said:

First, As long as they base match making on the ELO system, and Allow New players to drop with Senior Players, Match,maker will NEVER be able to actually give Good matches. The chance is always there, but it will be pure dumb luck.

Problem is if they don't drop with veteran players they'll have nobody to drop with. There aren't that many new players to have a new-player queue, partially (if not mainly) because when they were enough new players the PGI didn't bother to fix this issue, thus driving those new players and potential future new players away.

We are crying for 3 years that something needs to be done about new player experience. In-game or at least official video tutorials, hell a fkn game MANUAL would be a start. But devs don't care, hoping that some day in the far future they'll make a good game out of MWO and people start playing it, only thing it doesn't work that way. The game already has a very bad rep and its only getting worse, due to multiple ever-present issues like hit-reg or rather lack thereof.

Wesxander, on 29 January 2015 - 08:15 AM, said:

New players start at a middle ELO,so you get a total noob in a new mech with out even basics unlocked, or worse one of the idiot trail mech builds that has enough ammo for 9 shots ... MM can easily give you a whole team of those, while putting all veterans on the other side, based solely on they're Elo rankings.

Not in a middle, no. But too close to the middle anyway. I think it was 1200 out of possible 4000.

Wesxander, on 29 January 2015 - 08:15 AM, said:

I've asked before...and will again, we need some tags for folks who have reached a certain number of drops ... ie like 1000, 2500, 5000, 10,000, 25,000 ..etc so we can see the experience levels, and know why we probably got roll stomped.

There is no reason to hide Elo scores. Much less reason to not allow us to see other people stats in game. Total number of matches included. Does not affect MM, but at least new players will be able to see that a 'guy-1' played 10000 matches, so it is probably wise to ask an advice from him rather then from a 'guy-2' who played 10 matches.

IraqiWalker, on 29 January 2015 - 07:23 PM, said:

There was always the choice to make "called shots". Not aimed shots (Those require TCs). Big penalty, but you can definitely slap multiple shots on the same location.

With like 1 out of 10 chances. Here you just point and click. I truly miss the time of 6 PPC Stalkers ...

[PhoenixFire55]

Think I'm done doing MM posts for today. Always makes me angry when I recall CB matches and realize how since then this game turned into a complete piece of sh!t.

[Meta 2013]

To : #46 PhoenixFire55

There was one detailed explanation on exactly how elo worked when they re-did the match maker a while back. the range is 0 to 2800, new players get an elo of 1400 for the first 100 matches, of each weight class. Then the real elo is calculated. Yes you can come up with a very close approximation of where your sitting for your ELO. pretty much 90% or more of what's going into your ELO is your win/loss in that weight class. I'd hunt that down for you, but the forum search tool is a joke.

You said hardly any new players, beg to differ, all I bloody see is new players, and you can tell the trial mechs right off. The folks who pick assaults run out of ammo after the first mech they engage, its not unusual (not the norm though) to see 2 lances of folks with 20 to 80 damage total each.
----

All in all several folks have come up with many reasons leading to he roll stomps, all are valid and add in to the roll stomp formula.

The real point is ... PGI will not address it.

The CW race was a Joke, sure you have it now, but in the end, it all still comes down to ... Players want a Game that ...<READY PGI> .... WORKS.

Over 20 years ago, Valve software got hit reg right. sure, always problems when you mix high and lo pings together, you folks cannot get it right even when pings are close. Time to tuck your tail between your legs ...again, and get some real help from people who actually know what they are doing and get this fixed. Call Valve, and who knows they may take pity on you. There is no saving pride now, we all know you cannot do it yourselves.

The rush to get cw done has put so many bugs in the UI, it has become very unstable, almost every match you see folks crash at the start and rejoin, myself included. Just random crashes to desktop, for myself it is not a great number, but still at least once daily, sometimes more. Yes the software is complex and as they have stated they are learning, yep we get that, but Fix the product you have before you jump into the fire an add more stuff you cannot make work, that still inherently suffers from the Game Basics you have ignored all along. You may not be ignoring them, but your not fixing them either.

Time to #$%^ or get off the pot. So many good players have just lost all interest in the game, guys that would have been here for the next 20 years spending cash, just gone, because you cannot do one simple thing, make the game work. That means balance matches, and get hit reg working, and lastly give a purpose to all the grinding. You won't even give us little Tags that signify achieving certain milestones you can display next to your pilot name in game ... a simple thing like hitting 2500 drops, or 10000, or 25000. Something that simple, gives folks something to shoot for.

Lastly : FOR THE LOVE OF GOD, do something to make it easier for new players. -make it easier for new players to get into decent mechs and get them upgraded , that will ultimately help with the Topic at hand - roll stomps.

Fixing the trail mechs, and putting realistic, builds, not the bogus crap you do now for most of them, will go along way to making matches better. Put double heats sinks on them, put AMS on EVERY trial mech, so folks can learn what an AMS net can really do. This isn't pro builds, this is pug dropping, get over yourselves and make good builds for the trail mechs. Put at least 2 tons of ammo per 10 launcher, not 1 ton per 20. Put 3 tons for ballistics, these are simple things that can really help sway the roll stomp formula a little. Not a fix, but helps eliminate some of the contributing factors. This is doable, Plenty of folks here will flood you with good trail mech builds, so what's your excuses for not doing such a simple thing to help make the game a little better?

We get you need t make money, but if you keep driving folks away, and keep the new player experience a grind of frustration, not going be anyone left to entice with your premium stuff. Also get real, your sale prices are what you should really be charging for the content. 30 dollar mechs that are worse than the c-bill versions... come on ... you want example? Pretty Baby - I rest my case. Fair is Fair, you did an excellent job on the Ilya, the fire brand, and a few others have note worthy reasons, like not putting all weapons in one arm, etc. But, some suffer from your being scared of pay to win, and were nerfed too much to be useful. Doesn't need to be pay to win, doesn't have to be a power house, but for 30 bucks...darn it, make it fun to drive.

You charge German Engineering quality prices for a sub k-mart quality product. Your not there, you need to be in the volume market right now. 99 cents a color, not 4 bucks. Camo patterns.. per chassis, not for that price. If I buy it, I should own it for any mech I own.

Just sayin.

Edited by Meta 2013, 30 January 2015 - 02:52 PM.

[IraqiWalker]

PhoenixFire55, on 30 January 2015 - 12:53 AM, said:

LMAO ... take an Atlas and go 1v1 against an Ultra Crab or an Ultra Whale. I'd love to see you survive for even 15 seconds let alone 30. Thats with double armor and 1.4 heat sinks. But of course keep telling yourself its longer.

Take those same builds into TT and tell me how long your opponent's mech lasts (hint: with called shots, an ultra whale can kill another assault pretty much within the turn. Aimed shots with a TC are even better).

You also seem to be forgetting that there was specialized ammunition in TT that could make the kill even faster.

You have played TT, right? 6 UAC5s is 12 heat for 60 damage (An atlas has all of 30 ish internal HP on it's CT), if they don't jam. With aimed shots, and a decent pilot, you can place all 60 points of damage into one section. Killing your opposing mech in 1 turn.

Should I bring Gausszilla onto the table? An Annihilator with 5 Gauss Rifles Sure, you can't slap TCs worth a damn into that one, but the 75 point alpha almost guaranteed a kill a turn.


While it's true we have twice the armor, and HP, but triple the fire rate, at least there are some restrictions here.

Edited by IraqiWalker, 30 January 2015 - 08:58 PM.

[PhoenixFire55]

IraqiWalker, on 30 January 2015 - 08:58 PM, said:

Take those same builds into TT and tell me how long your opponent's mech lasts (hint: with called shots, an ultra whale can kill another assault pretty much within the turn. Aimed shots with a TC are even better).

Did you even play TT ever before? An Ultra AC fires ONCE per 10 seconds turn. And there is a high chance that it'll jam for the rest of the game if you double fire it. Called shots or no called shots, 5 UAC Dire will do 25 damage all over the Atlas, 50 damage if you are lucky and maybe 20-30 next turn depending on how many UACs got jammed. Not nearly enough to drop a mech with 40+ armor on legs, 50+ on CT and sides.

IraqiWalker, on 30 January 2015 - 08:58 PM, said:

You also seem to be forgetting that there was specialized ammunition in TT that could make the kill even faster.

Irrelevant as we do not have it in MWO.

IraqiWalker, on 30 January 2015 - 08:58 PM, said:

You have played TT, right? 6 UAC5s is 12 heat for 60 damage (An atlas has all of 30 ish internal HP on it's CT), if they don't jam. With aimed shots, and a decent pilot, you can place all 60 points of damage into one section. Killing your opposing mech in 1 turn.

I did, which can't be said about yourself looking at these comments. Atlas has 62 internal CT structure and up to 60 armor on front CT. That is 120 damage do kill it via CT only shots, which never happen in TT to begin with, you are lucky if half you shots are going to hit CT. Not sure what kind of TT you mean, but you could never place all your shots into same location in the original game, it was designed that way.

IraqiWalker, on 30 January 2015 - 08:58 PM, said:

Should I bring Gausszilla onto the table? An Annihilator with 5 Gauss Rifles Sure, you can't slap TCs worth a damn into that one, but the 75 point alpha almost guaranteed a kill a turn.

An Annihilator with 5 Light Gauss Rifles for starters, thats 50 points of damage per turn, which will again be spread all over the mech. If 3 out of 5 hit CT you'd be damn lucky.

IraqiWalker, on 30 January 2015 - 08:58 PM, said:

While it's true we have twice the armor, and HP, but triple the fire rate, at least there are some restrictions here.

lol Restrictions? Like what? PUGs habit of sitting in one spot doing nothing?

[Av4tar]

How about AI with same mechs like attackers?

[saKhan Steiner Lawl Kerensky]

Chaosity, on 23 January 2015 - 11:09 PM, said:

Title says it all.

And before somebody says be a better pilot, one bad player does NOT make for continuous 12-0 games, but a bad Match Maker does.

Partly PGI's fault, is that they're very far from any previous vision of the Mechwarriors and light years away from Tabletop. So if people haven't been along for the ride for all of the doublearmor/almostdoubleheatsinks/nerfweapons/buffweapons/ecm/hitregjoke/ghostheat/clanop/isquirks, people come in with nostalgic memories of what was fun or worked for them in the past and get a kick in the teeth if they don't sit down and read a few years worth of patch notes... because it doesn't really SAY up front in MW:O how MUCH they've band-aided the mechanics of this game.

Second, you can't fix stupid.

PGI tried to compensate for stupid and people who are in general bad pilots. ELO can't compensate for ignorant builds, unaware people in a team environment, and poor choices that players make, especially for Innersphere mechs.

In my honest opinion, I think that's why Clan does so well, despite all the nerfs. They are forced to play with less choices that would FUBAR their mechs, and I would argue that it makes their mechs and pilots more efficient resulting in more wins. When new players buy an IS chassis or a new engine without researching how well it'd do, they're screwed into that choice and play a horrible build until they get enough cbills MANY games later to fix it or try something else. If a person hates a load out on a Clan mech, they just buy a separate arm or torso and fix the weapon loadout instead of another IS engine and chassis/upgrades/weapons to make a meta build.

Long story short:
-New players are face with too many unknowns about game mechanics which impact greatly their team contribution.
-New players who purchase and customize IS mechs will most likely have a more difficult time than Clan mechs due to nostalgia, unknown adjustments of game mechanics, and too much choice for IS mechs which will lead to less than optimum builds that will impact team play.
-Matchmaker can't compensate or balance teams for horrible mech builds (new player or not).
-You can't fix stupid.

[Willard Phule]

Wow....this is hilarious.

So many people that KNOW how the Elo/matchmaker works, yet, there hasn't been a post about it for over 6 months.

The MM was rewritten in June of 2014. Since then, we've asked Karl to give us a comprehensive post about how it works...only to be ignored.

Keep thinking you know what it's doing. It's hilarious to sit and watch you all bicker about stuff you know nothing about.

[Black Ivan]

Indeed. The magcical match maker . PGI would slowly need to do a serious overhaul on that thing after release of the Resistance package. IMO there are many things they should add to take into account for match making.

[DeathlyEyes]

I want to see what happens to solo queue with Elo disabled. Elo was put in to balance groups. Elo now just decides some matches before they are played by stacking one team to account for one players really high elo in solo queue.

[PhoenixFire55]

Willard Phule, on 01 February 2015 - 08:05 AM, said:

Wow....this is hilarious.

So many people that KNOW how the Elo/matchmaker works, yet, there hasn't been a post about it for over 6 months.

The MM was rewritten in June of 2014. Since then, we've asked Karl to give us a comprehensive post about how it works...only to be ignored.

Keep thinking you know what it's doing. It's hilarious to sit and watch you all bicker about stuff you know nothing about.

Its just as hilarious when you claim that you know it was rewritten in June 2014 only based on smth smb from PGI told you. They claim they fix hit reg in pretty much every second patch, but we all know its not entirely true, or more like entirely not true. According to this logic it was just as hilarious to discuss the MM for the past 2+ years, simply because all you knew about it was based on what PGI told you about it, which can be anything really.

People are at least trying to draw attention to the issue the existance of which can't be denied (unless you are on an island^TM I guess). We work with what info we have, as for your post, it helps us how exactly? It seems to me the only one here who is bickering about stuff he knows nothing about is you.

[PhoenixFire55]

DeathlyEyes, on 02 February 2015 - 02:35 AM, said:

I want to see what happens to solo queue with Elo disabled. Elo was put in to balance groups. Elo now just decides some matches before they are played by stacking one team to account for one players really high elo in solo queue.

Nothing much. If you simply draw 24 people out of the waiting queue and randomly split them into two teams you'll have a high chance of ending up with a balanced game. Its all according to LLN. Judging from my own experience the current % of 12-0 / 12-1 roflstomps is about same as predicted statistically for random MM.

[Willard Phule]

PhoenixFire55, on 02 February 2015 - 03:43 AM, said:

Its just as hilarious when you claim that you know it was rewritten in June 2014 only based on smth smb from PGI told you. They claim they fix hit reg in pretty much every second patch, but we all know its not entirely true, or more like entirely not true. According to this logic it was just as hilarious to discuss the MM for the past 2+ years, simply because all you knew about it was based on what PGI told you about it, which can be anything really.

People are at least trying to draw attention to the issue the existance of which can't be denied (unless you are on an island^TM I guess). We work with what info we have, as for your post, it helps us how exactly? It seems to me the only one here who is bickering about stuff he knows nothing about is you.

Look.....we know crap all about how the Matchmaker and the Elo thing works now. Everything out there on the web is from prior to June of 2014.

June of 2014 is when both the clan invasion was released and they re-wrote the matchmaker to take into consideration "team average" Elo scores.....as opposed to trying to make even teams with individual Elo scores. That part is described very plainly in the patch notes...go see for yourself.

Since then, we have heard NOTHING about how it works, even when directly asked by a forum moderator (Egomane).

So, everyone's opinion of how it works or doesn't is valid. Until someone can prove anybody's theory wrong with hard evidence, everyone is right.

[Platinum Spider]

Willard Phule, on 02 February 2015 - 04:39 AM, said:

So, everyone's opinion of how it works or doesn't is valid. Until someone can prove anybody's theory wrong with hard evidence, everyone is right.

That is not how i works. Stating something with no evidence doesn't make it right, just unsubstantiated. They can't all be right, but they can all be wrong.

[Electron Junkie]

DeathlyEyes, on 02 February 2015 - 02:35 AM, said:

I want to see what happens to solo queue with Elo disabled. Elo was put in to balance groups. Elo now just decides some matches before they are played by stacking one team to account for one players really high elo in solo queue.

So much this... PUG drops only FFS.