52 replies to this topic

[OnLashoc]

Stoicblitzer, on 17 June 2013 - 11:56 AM, said:

here is my tongue-in-cheek reply: l2p and working as intended™.

i like how people argue that allowing 8 mans back into the pug queue will cause roflstomps....like they don't happen anymore.

<3

We tried to warn them that removing 8 mans was not the answer (those of us knew that baddies were just bad, and needed something to blame for their awfulness).

I think my stats have actually went up because I no longer have 8 mans to counter my own group. Now its 2-4 trained teammates with 4+ pugs which I assume is against 8 pugs with a couple maybe on the same team.

[JokerVictor]

Zerberus, on 17 June 2013 - 02:13 PM, said:

No, it isn`t.

CS:Go uses Elo, as does Guild Wars, WoW used it for the Arena until 2010, and numerous other online games which would require more than precursorily reading the sources I freely provided in a previous post. If you don`t read documents handed to you, you are handicapping yourself. L2Research in a manner that someone with a "science background" should have been taught. This is stated in this form solely in counterpoint to your equally condescending L2Argue.

In light of your proven capacity for in depth research as well as your obviously holier than thou self-perception, I don`t think I need to concern myself with the rest of your statement.

Lol, arrogance without evidence wins arguments, yeah!! Speaking of holier than thou....

Also, failing to address the points you know I'm right on doesn't mean you can ignore them.

And I'm sorry, I forgot some details in the first go round... so let me add them. From your freely provided 'research materials' (lol, wikipedia).

Quote

In Guild Wars, Elo ratings are used to record guild rating gained and lost through Guild versus Guild battles, which are two-team fights. The initial K-value was 30, but was changed to 5 in January 2007, then changed to 15 in July 2009.

Not rated on an individual basis.

Quote

Counter-Strike: Global Offensive also uses the Elo rating.

A game that has NO CLASSES OR PREGAME CUSTOMIZATION OF ANY KIND and a very healthy population. Works fine there.

So, LoL is the only class based game that still uses Elo for individual random matchmaking. Happy now? Now rebut the rest of it you arrogant b*stard.

Edited by JokerVictor, 17 June 2013 - 02:39 PM.

[Dracol]

Just to stir the ELo pot. In some regards MW:O elo can be compared to a sport utilizing the same type of system.

The argument presented against the comparion was that MW:O teams are random groupings, while sports Elo is based around a set team that does not change.

But, in sports, teams have second strings. From game to game, the team members that get the majority of playing time can change due to injuries, giving rookies oppurtunities against weaker opponents, etc.

In Mw:O, some matches will be comprised of the top players of their elo bracket and at other times, the second string who's elo has been carried by more experienced players will be fielded.

Edited by Dracol, 17 June 2013 - 02:45 PM.

[JokerVictor]

Dracol, on 17 June 2013 - 02:44 PM, said:

Just to stir the ELo pot. In some regards MW:O elo can be compared to a sport utilizing the same type of system.

The argument presented against the comparion was that MW:O teams are random groupings, while sports Elo is based around a set team that does not change.

But, in sports, teams have second strings. From game to game, the team members that get the majority of playing time can change due to injuries, giving rookies oppurtunities against weaker opponents, etc.

In Mw:O, some matches will be comprised of the top players of their elo bracket and at other times, the second string who's elo has been carried by more experienced players will be fielded.

A valid point, but the game calculates the teams' Elo value on a per match basis. Maybe when CW goes in and there's an actual elo score assigned to your unit will we see some improvement here. Doubtful, though.

If you've ever 4-manned regularly you've probably seen this happen... where you as a team go on a hot streak and win 8-10 matches in a row, after which you get consistently paired with just terrible players, generally just barely out of their initial 25 matches or sometimes still in them.

It really doesn't matter what the other team consists of at that point, you were paired with steering wheel drivers to bring your average team elo down so you could be dropped in the match. You almost always lose these matches, because 4 really good players are still going to lose to 8 average players after their 4 training wheel teammates get themselves killed without doing anything. Unless, of course, you get extremely lucky (which does happen from time to time).

[Loonix]

JokerVictor, on 17 June 2013 - 02:00 PM, said:

I do have to provide a better alternative to be right? No, I don't. It's not my job to design these systems, but I'm perfectly qualified to point out their flaws as an end user.

Quite right.

Anyone with a decent, high school grasp of math can understand the devs explanation (credit to the devs post; you'd have to be fairly bad at math not to ). And anyone with a reasonable amount of in game experience can imagine the implications.

If I can say one thing positive about the system, it should certainly do a good job of keeping everyone's win/loss close to a flat 1for1 - something I assume all games like MWO aim for. But will it make for consistently compelling game play? Barring anything I've missed, or the devs haven't told us, no it will not. There are simply too many important variables not in the mix. It shouldn't even need explaining.

In any case, I'm not sure matchmaking can truly be addressed properly until general weapon/load out balance has been addressed.

[Dude42]

Yes I love how when I do well and win lots of games, instead of being paired with and against players of higher skill, I just get more nubs on my team.

Working as intended.

[Crockdaddy]

I'd be happy with some transparency into my ELO score, that would help me personally. I know my winning percentage is around 61% in assault and 55% in conquest matches (not surprising considering I play Assault / Heavy mech's often). My K:D hovers between 2.2 to 2.3 ... it goes up when I run the fat mechs and drops when I play in the spider. Ironically, when I group up, I die more often, in part because my role is often meat shield for my more talented teammates. I'd say I pug 1/3 of all my games and am group 2 to 4 on TS on the other 2/3. If my grouped team runs optimal builds, we win 80% of the time if not a bit more. If a few of us are leveling new mechs ... we usually lost 60 to 70% of the time.

Much of the experience simply depends on what build I / we take and what we are trying to accomplish. In 8 mans, we tend to lose more often, but that also depends on whom we face as well. (Like when we run into the semi pro ppc sniper teams we die often as they are quite good).

So I have many different experiences within the game, relatively speaking it is more negative than positive since the implementation of ELO, but likely I was used to winning 90% of my matches prior to ELO. What would make me happy is to have a transparent ELO score.

Oh and at the very least, consistent class balancing. If we have 3 Heavies, it would be nice to know we aren't going against 3 stalkers.

[Kazly]

I should start this by stating I have an inherent problem with any system designed to make me lose every other match, regardless of my skill level.

Who does ELO serve? To make your random drops during beta more fun? It’s not working for me.
Let community warfare sort ELO out. Random drops should be just that – random. In CW, we’ll see what TEAMS are the highest ranking, which are the lowest. But there’s a key word here – team. Team team team. People playing together all the time. It’s these folks that ELO is needed for, because the team becomes an individual entity and can properly be adjusted with a rating system.

Random drops? Scoring me based on 13 other ******* variables? Screw that. Don’t judge me because the two lights on my team decided to mix it up and die instead of watching the caps, even after 6 of us rolled through your teams but can’t make it across alpine peaks to a cap point at 60-70kph before the timer ticks off at 750 points – the same amount required by the way in River City, a map 1/5^th the size of Alpine.

Bah.

ELO?

ELO is not needed in random drops.

Edited by Kazly, 18 June 2013 - 10:33 AM.

[Appogee]

Can we hear from some of the most skilled players what their experiences are like?

Tonight I dropped with several people who couldn't even steer their Mechs, let alone land a shot on target. Fair enough, everyone has to learn the game. I don't blame them for that. But why am I dropping with them on my team...?

After 1000 games of MWO I want to lose because I either made a mistake or because my opponents were better than me. Not because half my team were beginners.

Edited by Appogee, 18 June 2013 - 12:25 PM.

[Dude42]

Appogee, on 18 June 2013 - 12:10 PM, said:

Can we hear from some of the most skilled players what their experiences are like?

Tonight I dropped with several people who couldn't even steer their Mechs, let alone land a shot on target. Fair enough, everyone has to learn the game. I don't blame them for that. But why am I dropping with them on my team...?

After 1000 games of MWO I want to lose because I either made a mistake or because my opponents were better than me. Not because half my team were beginners.

If only we had some way to determine who is most skilled.

I feel like I'm quite skilled, but for all you or I know, since we can't see ELO ratings, or any kind of skill bracket, I'm a total scrub. All I know for sure is that I pretty much always at or near the top the damage/kills/score at the end of the round. I've played about 1700 matches since I joined a couple months ago. My first ~700 matches were somewhat bad... For one thing I kept getting dropped from matches due to malformed packets, and after ~700 games had a KDR of 0.46, and a W/L ratio of 0.65.
Since they fixed they malformed packet thing last month my KDR has climbed to 1.25, and W/L is now 0.96. So you must imagine that after the first 700 or so, my stats have been skyrocketing. I also do not own a FOTM mech(no highlanders, cataphracts, jagers, or stalkers). So I imagine if I gave in to the cheese I'd be doing quite a bit better. I'd be happy to post all my stats here for all mechs if requested.

(If you're curious I have COM, RVN, BJ, HBK, DRG, AWS, and AS7, 3 of each, for a total of 21 mechs).

Anyway, that being said.... I still get paired with total newbies, lots of them. People who have trial mechs, with locked arms, have all weapons in the same group, alpha strike every shot, miss, and then overheat. I still regularly see things that make me want to cringe. If you look around in some other threads you can find some screens I posted that illustrate the point. It's not just my team either. There are plenty of wins we get because I exploit the nubs on the enemy team. Its easy to cockpit shot someone once you realize they cannot hit you, even if you stand stock still in front of them. Every match comes down to 2 or 3 people on each team, the rest are filler.

Since they introduced the new screenshot feature I have been screenshotting the match results nearly of all my rounds(good or bad, win or lose). I browse through them to see names repeated, and see how they do. If you make notes of who the good players are, and what they're driving that day, eliminating them first almost always ensures an easy win. I feel like there's times they do that to me too, lol. Work the system

I see the same names all the time, which vary depending on when I'm playing. I know who to watch out for, and I know who I can walk up to can cockpit with impunity. It helps a lot.

Edited by Dude42, 19 June 2013 - 02:05 PM.

[Gaan Cathal]

Neverfar, on 17 June 2013 - 01:22 PM, said:

If I had MC for every time someone in the group (often the guy that's semi-idle or died first) says everyone else is a noob/tryhard/terribad/whatever, I'd have every Mech in the game.

Oh lord, the amount of times I amble up to the enemy base three minutes into an Assault match and find a perfectly stationary mech on their spawnpoint that 'suddenly wakes up' when I shoot it in the arse. I swear it's these guys. Probably think they can swoop in and get lots of cheap kills halfway through to make the KDR look good or something.

[Gaan Cathal]

Appogee, on 18 June 2013 - 12:10 PM, said:

Can we hear from some of the most skilled players what their experiences are like?

Tonight I dropped with several people who couldn't even steer their Mechs, let alone land a shot on target. Fair enough, everyone has to learn the game. I don't blame them for that. But why am I dropping with them on my team...?

After 1000 games of MWO I want to lose because I either made a mistake or because my opponents were better than me. Not because half my team were beginners.

Because of the way Elo works actual new players - which in this case is actually anyone new to a weight class - will enter into the system in the middle of the ratings. If you persistently find you're getting paired with and against obviously new people, then you're likely in the 'average' band of Elo. If it's only with, not against then you're probably slightly higher than average and get yanked in to compensate for them being outclassed.

Also, if he's still around, Wispsy. Indisputably very, very ******* good light pilot. And high-Elo gameplay isn't exactly friendly to lights.

[Dude42]

Gaan Cathal, on 19 June 2013 - 02:15 PM, said:

Because of the way Elo works actual new players - which in this case is actually anyone new to a weight class - will enter into the system in the middle of the ratings. If you persistently find you're getting paired with and against obviously new people, then you're likely in the 'average' band of Elo. If it's only with, not against then you're probably slightly higher than average and get yanked in to compensate for them being outclassed.

Also, if he's still around, Wispsy. Indisputably very, very ******* good light pilot. And high-Elo gameplay isn't exactly friendly to lights.

I understand what you're trying to say, but I think what you mean is not that the enemy team won't have newbies on it, it's that the enemy team will have less newbies than your team, if you are higher-ELO.

The problem is that at certain times of day the player pool is too small for matchmaker to consistently be able to put together a "high elo"(or "low elo") match. As you are aware in the elo distribution curve, the populations thin way out at the extreme levels of skill. So we end up with newbies and pros thrown together along with regular guys in "medium averaged" elo games. That happens to everyone. You really notice problems after win streaks(for me it occurs at about a 2.4-2.6 W/L ratio with a mech) when matchmaker starts dividing up the nubs less evenly. The enemy team may only have 1-2 new players, but your team now has 4-6. Especially at certain times of day. In the evenings, at least for me, its a lot better, and for a few hours you get some real good games. But then comes late night, and all the messed up matchmaking you can handle.

[Reported for Inappropriate Name]

what are you driving? if you're solo I really recommend driving something either with ecm, or a catapult with two gauss rifles. The objective here is to use something that is not teammate dependent to get within firing range of the enemy and get damage out.

also, it's impossible to win every game in a random team environment(especially one that forces you to play with players outside your skill range, good or bad), but the best way to mitigate losses definitely lies with ecm, or extreme range and damage output.

think of it this way, if you're in front of your teammates, you're relying on them to be damage effective. If you're behind your teammates, you're relying on them to make targets of themselves so the opponents don't shoot at you. Of course if you're in an atlas, everyone will come to huddle in your shadow, so it's more important to pick an appropriate attack vector with them in consideration. Driving something big and imposing will signify status, less competent players in mediums and heavies will go "woah, look at that atlas painted so outrageously, i bet he knows what he's doing", and that ladies and gentlemen, is why the D-DC is the best lone wolf assault money can buy.

but seriously, if you want people to pay attention to what your mech is doing, friend or enemy, paint it like construction equipment.



also when are we getting the ability to change the name of our mechs? I don't like having the same name as a mech in the game, who knew we'd be getting a highlander named heavy metal.

Edited by Battlecruiser, 19 June 2013 - 03:13 PM.

[Dude42]

Battlecruiser, on 19 June 2013 - 03:03 PM, said:

what are you driving? if you're solo I really recommend driving something either with ecm, or a catapult with two gauss rifles. The objective here is to use something that is not teammate dependent to get within firing range of the enemy and get damage out.

also, it's impossible to win every game in a random team environment(especially one that forces you to play with players outside your skill range, good or bad), but the best way to mitigate losses definitely lies with ecm, or extreme range and damage output.

think of it this way, if you're in front of your teammates, you're relying on them to be damage effective. If you're behind your teammates, you're relying on them to make targets of themselves so the opponents don't shoot at you. Of course if you're in an atlas, everyone will come to huddle in your shadow, so it's more important to pick an appropriate attack vector with them in consideration. Driving something big and imposing will signify status, less competent players in mediums and heavies will go "woah, look at that atlas painted so outrageously, i bet he knows what he's doing", and that ladies and gentlemen, is why the D-DC is the best lone wolf assault money can buy.

Actually...

My stats indicate that seismic sensor is more valuable than ECM as far as W/L(and KDR) goes, by a lot But if you can have both, take both.

Edited by Dude42, 19 June 2013 - 03:10 PM.

[Reported for Inappropriate Name]

Dude42, on 19 June 2013 - 03:09 PM, said:

Actually...

My stats indicate that seismic sensor is more valuable than ECM as far as W/L(and KDR) goes, by a lot But if you can have both, take both.

the whole point of ecm is to get close enough to use your big weapons. The whole point of seismic sensor is to let you know when someones close enough to kill you. I definitely recommend both. Every battle is about two things, knowing when the enemy is going to hit you, and hitting the enemy before they know you're it is coming..

Although personally, I might as well be an esper when it comes to tracking opponents (although this doesn't mean i'll be able to do anything about whats coming towards me). I don't need seismic but that doesn't stop me from using it, for the rare occasions where my predictions are wrong.

Edited by Battlecruiser, 19 June 2013 - 03:22 PM.

[Stoicblitzer]

although i may not be uberskilled, i play with people who i believe are in the top 5-10% of Elo. our experience seems to be exactly the way others in this thread have described theirs. 4 high elo+4 low elo or trial mechs. sometimes, it will be 8 high elo vs 4 high elo+4 low elo. I guess in that circumstance, the MM is trying to give the 8 high elos a win. this arbitrary system forces people to take the best mechs in the game to compensate for the low elo players. this is why you see the tonnage + ppc arms race. of course, i have no way of determining for sure who is high/low elo but we typically see players from top competition teams so i am assuming they are high elo.

Edited by Stoicblitzer, 19 June 2013 - 03:21 PM.

[Dude42]

Battlecruiser, on 19 June 2013 - 03:17 PM, said:

the whole point of ecm is to get close enough to use your big weapons. The whole point of seismic sensor is to let you know when someones close enough to kill you. I definitely recommend both. Every battle is about two things, knowing when the enemy is going to hit you, and hitting the enemy before they know you're it is coming..

Although personally, I might as well be an esper when it comes to tracking opponents (although this doesn't mean i'll be able to do anything about whats coming towards me). I don't need seismic but that doesn't stop me from using it, for the rare occasions where my predictions are wrong.

I'm not disagreeing, just want to point out that the increased W/L from seismic(and ECM) is due to the ability to watch the enemy team from safety and relay that information to your teammates.

The increased KDR on the other hand is from exactly what you describe, seeing before being seen, and not being taken by surprise.

As is plainly evident from my mech stats, high KDR =/= high W/L.

Edited by Dude42, 19 June 2013 - 03:27 PM.

[Reported for Inappropriate Name]

Dude42, on 19 June 2013 - 03:26 PM, said:

As is plainly evident from my mech stats, high KDR =/= high W/L.

my kdr is bad because i mess around with sub optimal builds. however i'm done with that. the next thing to enter my garage after d-dc will either be be the gauss catapult once again, or another atlas so i can continue to work towards that extra module slot. it could go either way because properly fitting a gauss catapult costs around 13 mil total


also, as far as my opinion on elo goes, the last time i played this game for a large amount of time in a 4 person group, we would win two, and then like clockwork get a game missing two teammates
see, we even made a thread about it back in april.
http://mwomercs.com/...wheres-my-team/

Edited by Battlecruiser, 19 June 2013 - 03:36 PM.

[Stoicblitzer]

Battlecruiser, on 19 June 2013 - 03:31 PM, said:

i mess around with sub optimal builds. however i'm done with that. the next thing to enter my garage will either be the gauss catapult or another atlas

uhhh....the atlas and catapult are currently suboptimal mechs. stalkers are the best.