49 replies to this topic

[MrMadguy]

Mawai, on 26 April 2015 - 06:56 AM, said:

Sigh. Not this again.

Honestly, it likely has absolutely nothing to do with Elo.

In a PUG match a certain number of players run in random directions or just do their own thing. These folks usually either die first by running into the opponents or last by not helping the group that does run into the opponents.

Whichever side loses these folks first stands a much better chance of losing the match because they will be down 2 to 4 mechs almost immediately.

I see folks on BOTH sides doing this. If you are unlucky then the ones on your team die and if you are lucky it is the ones on the other team.

Once your side is down a couple of mechs it can be very easy for the game to steam roll from there ... especially if your team is not together and the opposition encounters several mechs. They roll over them and your fire is not focused and usually ineffective.

It has nothing to do with Elo .. it has to do with team work. A few words in chat or a few brief comments on coms can often get a PUG group working a bit more cohesively ... do that and your odds of winning probably go up 25%. Don't do it and you are depending on someone on the other side NOT doing it and the luck of the derp to fall your way as folks wander about randomly.

Some folks are in your Elo range because they aim well ... some are in the same range because they stick together and work well on a team. Get too many folks with good aim and bad team work ... they wander all over and die fast ... if you have the choice (which you usually don't) take team work over aim.

I really love those people, who always say, that "game magically lives on it's own" and devs are to weak to do something with it. One of those standard excuses. May be it's you, who don't understand, how ELO works? If I'm losing due to having bad players in my team, my ELO should drop and therefore MM should ensure, that I will start to be matched with better players. As simple, as that. If MM isn't working like this, which causes me to have extremely long lose streaks - then it fails.

[Novakaine]

Bishop Steiner, on 25 April 2015 - 01:13 PM, said:

So. I don't know if this is everyone's experience, but since Elo's have been introduced, I seem to be on a never ending, swinging cycle. 20(ish) Easy Wins, followed by 20(ish) crushing, curb stomping losses, being stuck on teams that I don't think even qualify for the Underhive.

Sure there are a handful of decent matches sprinkled in, but for the vast majority of the time, that seems to be the pattern. And it's pretty noticeable. I'll be going along OK; then suddenly, I find myself in one derp match, then another, and I know that for the rest of the day, things are going to go poorly. Switching Mech Classes, sometimes helps, as I thinK Elo is by Class, but even then, it's usually just forestalling the inevitable, ESPECIALLY if I play in the USA Evening time slots.

I should not that I am referring to the PUG queue here. And the stomps are not as bad, on average as before Elo, but dang it gets old, win or lose. It'd be one thing if it was like 5 wins, 5 stomps, rinse, wash and repeat. But going through so many losses in a row, usually makes me turn the game off. Part of it gets down to how fast losses snowball in a 12v12 focus fire environment. But probably 9 times out of 10 I can tell if we are going to lose before the first real shot is fired by the lack of coordination or other derp antics. And it really does seem to sing primarily between "good team" and "idiot team" with very little in between.

Why is it so hard to see more close, tight matches?

What's been your experience on the Mighty Elo Pendulum?

Come brother you know what the deal is.
Every time You, myself and many others logs in the server.
Our names are tagged.
Now I believe although I have no proof.
It's secretly based on in game experience not skill levels or elo whatever the hell that really is.
If you've been around a more than few years your'e considered an expert.
And are classed accordingly.
Now don't get me wrong I don't consider myself an expert by any means.
I'm a LRM pilot ya know.
But I do think the algorithm takes into account game experience.
When building a team.
Kinda makes you wonder.

[Turambar]

Gotta say this became very noticeable after Urbie patch for me.12-2 12-1 wins are super common now.

[KodiakGW]

LordKnightFandragon, on 26 April 2015 - 09:21 AM, said:

I dont understand how people can claim nothing is wrong when 98% of games end just like yours....4-12......and never better....that is a stomp......if it was more like 7-12, that was just a fairly bad game, but 4 or less to 12...that is an ass kicking.

I mean, look at your side match scores, you alone got your entire team's combined score...or just about.....yet for all your effort, you lose.

You are correct to notice the match score. I wouldn't focus on the kill difference between the two. I would focus on the Match Score and Damage difference. These tell you how many on the team knew how to do damage control, do team assisting functions, and hit enemy mechs.

Damage
Winners = 3984
Losers = 1856
Less than half (.47) the winners damage. Now if I had done the average damage of the rest of the team, 117, it would have been just over a third (.35).

Match Score
Winners = 777
Losers = 271
Just over a third of the winning team (.35). Now if I had done the average match score of the rest of the team, 19, it would have been well under a third (.29).

The white knights of the programming behind the match maker can attest to the fact that I've posted similar screen shots in other threads calling out poor match making. So this is not an isolated incident. This is more the norm than it should be.

But, they will continue posting that the math is right, stay the course, thousand points of light. Until somebody wakes up and makes any of the suggested changes to make it better, they'll be plenty of threads like this one that they can continue to make the same arguments.

[NRP]

I don't know. Sometimes I feel like I go on epic losing streaks (like 10-15 losses in a row), but my stats show I'm right at a 50/50 w/l ratio. Maybe I'm just a "glass is half empty" guy and only remember the losses?

[Bishop Steiner]

NRP, on 26 April 2015 - 11:52 AM, said:

I don't know. Sometimes I feel like I go on epic losing streaks (like 10-15 losses in a row), but my stats show I'm right at a 50/50 w/l ratio. Maybe I'm just a "glass is half empty" guy and only remember the losses?

actually you are saying what I am saying. I stay at 50/50.

The MM Elo is going to try to keep you there, and in general it works. I no e some argue the semantics, but the results is the same, a pretty standard 50/50 ish W/L.

Again, that I'm OK with, But the same should be accomplished by stacking the deck against me for 3-5 matches, balanced by 3-5 more favorable. Instead of like...20 of each. Would also be a lot less frustrating. (Sorry all the self fulfilling prophecy stuff some people are talking about like the MM/Elo issues are some metaphysical existential experience, I just don't get. Maybe scrubs do that, but few experienced players will).

Only thing I can think is the way the Elo score is recorded or adjusted is it takes that long for the MM to decide the discrepancy is enough to need to be balanced. Or, to make hos epicky about wording happier, before your Elo score range has stacked enough in either direction for it to have a notable impact on yor play capability, by running you up against those above or below your true skill level

[Wintersdark]

Bishop Steiner, on 26 April 2015 - 07:44 PM, said:

actually you are saying what I am saying. I stay at 50/50.

The MM Elo is going to try to keep you there, and in general it works. I no e some argue the semantics, but the results is the same, a pretty standard 50/50 ish W/L.

Again, that I'm OK with, But the same should be accomplished by stacking the deck against me for 3-5 matches, balanced by 3-5 more favorable. Instead of like...20 of each. Would also be a lot less frustrating. (Sorry all the self fulfilling prophecy stuff some people are talking about like the MM/Elo issues are some metaphysical existential experience, I just don't get. Maybe scrubs do that, but few experienced players will).

Only thing I can think is the way the Elo score is recorded or adjusted is it takes that long for the MM to decide the discrepancy is enough to need to be balanced. Or, to make hos epicky about wording happier, before your Elo score range has stacked enough in either direction for it to have a notable impact on yor play capability, by running you up against those above or below your true skill level

Its not just simantics, Bishop. The matchmakwr doesn't tey to get you to a 1.0 WLR, that's just what happens for the bulk of players as a result of what it does. Cause and effect are very different things.

You say:

Quote

Again, that I'm OK with, But the same should be accomplished by stacking the deck against me for 3-5 matches, balanced by 3-5 more favorable. Instead of like...20 of each. Would also be a lot less frustrating.

The matchmaker never "stacks the deck against you". That would imply that you are somehow important to the process, that the matchmaker is trying to accomplish something beyond just making as balanced a match as it can, and that isn't the case. The matchmaker doesn't make matches with the intent of one side losing. There is an prediction of which side will win, but that's independent of matchmaking and only comes into play for Elo modification afterwards.

I don't say this just to argue, but because its a very important difference. Even if you do understand, not everyone who is reading this will and it keeps the ridiculous notion going the the matchmaker is going to make you lose matches because you just won some.

Now, its possible to adjust the rate at which Elo changes, but its a double edged sword. The slower Elo changes, the more matches it takes to learn where you should be and the slower it can adapt to changing circumstances. So, it should move less in normal play, but if it does get away from where it should be it'll take even longer to get back. Changing faster means bigger jumps - quick adjustment to changing circumstances but more impact from individual matches so what is essentially bad luck can cause undesireably large Elo movement.

Adding group vs. Solo Elo scoring will help a hell of a lot, though. Group play can have a really weird effect on your own Elo score, depending on the difference between the group mates scores and is, IMHO, the largest contributor to peoples frustration with the system as a whole.

[Ghost_19Hz]

A lot of good discussion points being made here, I have similar experiences to the OP and many others, which is a worse experience since elo/12man introduced. And while i understand the elo system is not trying to make ppl lose its just trying to make an even match, the fact that the MM is controlled makes arguments for the side that it is "trying" to make people lose(it isn't, it probably just has bad data). The misconception arises from either 1 of 2 things: being that it either doesn't have enough ppl to create teams from, or it is just not functioning at the level expected of it (broken/bad data/algorithm or w/e you would like to call it).

Now, better players will always feel "sabotaged" in a "balanced" matchmaker, to prop up a worse team when paired with ppl way outside their skill level. Notice i said matchmaker, not elo. Except obviously elo is part of a matchmaker. See how ppl can feel oppressed by it? The other issue is that the DEVs see 50/50 win loss ratio as a job well done(hope that is not true). If that is the measurement for if its working, it will never make the game more fun to play. Save for maybe the ppl who lose 90% of their matches, they will now be carried in such an environment to reach a 50/50. Which is all just a symptom of low populations or bad metrics, not broken matchmaking or league systems. But since elo is what the MM draws conclusions from it essentially can "break" a good matchmaker system.

I think that if it were based off of average match score and separated by chassis would be much more accurate than win/loss predictions. Due to the fact that your team, even when organised drops are happening, is almost never a static variable for anyone ever.

Back in the days of no elo, we were nice to new players(most of us), helped them out. There was a bigger sense of community for me and there were 8-0 stomps sometimes but there was no one to blame(unless you're a jerk i guess) and boom you're in the next game before you can complain too much.

If we simply don't have enough players to create a good elo reference and match sets, throw the idea out. I don't mind trying to carry a team or helping new players. The Clan i belong to is also great with this sort of thing. We help new guys out where ever we can. Not all matches will be even but if your elo system isn't improving the experience, its doing more harm than good. W/o it though, its absence will be blamed for 12-0 stomps.

One thing to note about the idea of 12 man's creating a problem, i agree it changes the contributions of single players to much less than before and PGI has tried to alleviate it with different spawn locations(i think), and bigger maps. The maps are part of the problem but i think that they only exacerbate the issue that increasing the number of mechs in one location decreases the level of fun and level of strategy involved because the time to kill is lessened to near instant death. The different spawn locations has created slightly different matches, but only in the sense of "ok, where did the slow guys spawn, lets go meet them to ball up." Which is ok, i don't mind the different spawns but I'm not sure of the intent. Maybe someone who has been more active can fill me in on that.

Everyone is always so worried about matchmaker balance or if they are being rated fairly by it, balance is very important(game mechanics/rules anyway), its near impossible to get it exact, however, the beauty of a game is that it doesn't have to be precise in that manner. The weapons need tight balance, you can't balance the player skill and no game or sport does that to such extremes(argument for salary cap is there if someone wants). Games and sports have rules, and if you enter into a competition you may have an unfair advantage due to skill, that is exactly what the fair and balanced rules are there for, to figure out who has the "unfair" skill advantage. Nothing more really than seeing who's good at what and having fun while doing it.

The rules however, must be fair. The rules in a game are its mechanics and since humans are on their way mastering mathematics, game mechanic balance is possible. Let us all not forget that a game first and foremost should be fun.

One thing i'd like very much to point out is that, the grind is uncommonly brutal. If that is how it is going to be, expect more vitrol and hatred during matches, expect player retention to plummet, and more forum complaints on elo and matchmaking.
Make the game more fun, ppl buy more hats.

If your game is fun to play people won't mind being a 5-10% underdog due to their weapon/mech choice. No one will care, they will win and lose and have fun the whole time. If 12-12 is less fun in the current game maps/mechanics, lose it. So really, the micro managing players team placement, forcing CW maps into a choke, its easier to balance since your are micro managing the experience, but games aren't about being controlled, they are about freedom of choice and enjoyment.

TLDR; Get rid of elo for now, sponsor more tourneys, polish the weapons/chassis/maps, come back to elo later when you have the pop for it and use a more accurate metric this time around. Make the grind less brutal. Hats are to TF2 as Paints and Camo is to MWO.

Never lose primary focus of what a game should be at its core: Fun.

EDIT: run-on sentences and w/e.

Edited by Ghost_19Hz, 27 April 2015 - 01:50 PM.

[Telmasa]

FupDup, on 25 April 2015 - 07:05 PM, said:

I especially love the fact that there's UAVs all over the damn place in this picture...so true....

[FupDup]

Telmasa, on 27 April 2015 - 05:34 PM, said:

I especially love the fact that there's UAVs all over the damn place in this picture...so true....

Actually those are 3PV drones, at least I think. It would make sense, since the Atlas depicted is carrying bad players on his team (and 3PV is something that rarely gets used outside of newbies).