46 replies to this topic

[Gwei Loong]

kyfire, on 13 June 2017 - 06:52 AM, said:

An interesting concept, but.........There's one major flaw (as I see it). A player's W/L ratio is NOT an indication of the player's ability. Think about this......A player drops and dies within the first minute of the match without firing a single shot. His team goes on and wins the match, the player that died without firing a single shot is credited with a win. Did he win? Not in my opinion. He didn't contribute to the win.
Does the MM need to be fixed? Yes it does but some better metric needs to be found that is fair to all.

It is possible to write an algorithm that does not measure skill at all, except perhaps by stacking the odds against only the most skilled players in the game when using the solo que.
This is about preventing losing streaks for both new players and active players making the grind equal for everyone.
Players that want group experiences where their skill is pitted against another in a competitive fashion always have MRBC, or hopefully the new 8v8 mode.
No one ever says "I want a really good group where we all work together, all our builds complement each other, and everyone is on VoIP... Think I better chose solo que."

Kaeb Odellas, on 13 June 2017 - 07:02 AM, said:

Everyone has their string of bad luck and bad matches, If you're losing far more than you're winning over the long term, it's pretty safe to say that you're doing something wrong.

Ok fair enough but with three accounts going and not a single similarity between any of them I fail to make the correlation. It is like I already stated "Each account has its own stats and to tell you the truth they aren’t even close. When I compare their stats I have one account that's .88 w/l, one that is 1.8 and another that's 1.34. So this idea about the stats of an account accurately measuring ones skill cannot hold much value otherwise all of my accounts would have the same stats."
The only thing I can say is when I play this account I play it only in the solo que. My K/D ratios are a lot closer but it's not a good way to compare what we are talking about either.
I welcome PGI to look at the overall performance of the groups I have been having lately and then think about how they wish to balance their game. I have been playing long enough that you can see there has been changes in solo performance related to population and this relaxed matchmaker. My overall w/l stats are decreasing over time and not because I play worse. In the mean time I'll keep having fun and taking the piss but I am telling you unless you have twisted sense of humor your not going to find this game interesting.

Edited by Gwei Loong, 13 June 2017 - 08:09 AM.

[Gwahlur]

Kaeb Odellas, on 13 June 2017 - 07:02 AM, said:

No, kill/death is incredibly easy to game. You can inflate your K:D by last-hitting enemies or shutting down and hiding during certain defeats. You can't inflate win/loss. You can only reduce it by throwing matches.

Everyone has their string of bad luck and bad matches, If you're losing far more than you're winning over the long term, it's pretty safe to say that you're doing something wrong.

Well, if someone inflates their k/d, they'll be rated as a very good player, and thus get teamed with worse players, which over time will deflate their k/d again.

I still see k/d as the best metric to use

Edited by Gwahlur, 13 June 2017 - 07:45 AM.

[Gwei Loong]

Gwahlur, on 13 June 2017 - 07:45 AM, said:

Well, if someone inflates their k/d, they'll be rated as a very good player, and thus get teamed with worse players, which over time will deflate their k/d again.

I still see k/d as the best metric to use

Well I see the point. K/D does sort of accurately judge whether or not a person knows how to target and or focus with a team. It also unfortunately punishes light mechs as statistically they have the highest to K/D ratios.



So just going off of what we know about K/D ratios won't make the matchmaker egalitarian. Thank Kiiyor for this information by the way.

Edited by Gwei Loong, 13 June 2017 - 08:02 AM.

[Birthright]

Sorry but there is not matchmaking in place.

Im in Tier 4, still I get matched with or against Tier 1 players like that MoltenMetal youtube guy.

He's like tier 1 since forever, why would i ever be matched with him?

Hate to say it but there is absolutely no matchmaking in place.

[Angel of Annihilation]

I generally win more than I lose but I have my fair share of stomps, getting stomped and relatively well balanced matches so I don't see the problem. I think the problem is that most people don't realize that it only takes one badly out of position mech on your team getting killed to create the domino effect that leads to a loss not an unbalanced match maker. You need to understand that a single mech loss is around an 8% firepower loss, two mechs is 16%. Three 24%. If your team hasn't made substantial inroads to killing off an equal amount of enemies by this time, your going to lose.

[Davegt27]

when I first started I had disco after disco

on one map I had a 100% disco rate

when I did find out about stats I had 50/50 win loss rate I was pretty shocked
now PGI has an even more robust system
I am always impressed by PGIs computing power

people wanted balance well that's just what your going to get so you might as
well start sucking it up right now

[Beaching Betty]

Burke IV, on 13 June 2017 - 06:36 AM, said:

Losing is good for you. It teaches you to play better

True. It will make you think on what you did wrong and from there, you can improve yourself to be much better or less careless.

[Gwei Loong]

Viktor Drake, on 13 June 2017 - 08:41 AM, said:

I generally win more than I lose but I have my fair share of stomps, getting stomped and relatively well balanced matches so I don't see the problem. I think the problem is that most people don't realize that it only takes one badly out of position mech on your team getting killed to create the domino effect that leads to a loss not an unbalanced match maker. You need to understand that a single mech loss is around an 8% firepower loss, two mechs is 16%. Three 24%. If your team hasn't made substantial inroads to killing off an equal amount of enemies by this time, your going to lose.

Ok we know whats going on. I think most people understand this but it doesn't change whats happening in the case of disco's, people not playing the objective in domination, guys that go yolo, or groups that don't work together. Welcome to the solo que.

Beaching Betty, on 13 June 2017 - 11:03 AM, said:

True. It will make you think on what you did wrong and from there, you can improve yourself to be much better or less careless.

Maybe your right and in the end I'll keep on advertising the way I do and playing my games but in the end were not really the one's losing. Fact remains PGI is losing it's player base.
I am just suggesting we take a look at what Blizzard is doing and why millions of people play their games. You cannot say that one very successful company who wrote the book on this kind of marketing is not doing anything right and why argue your points with me when you can just look at the ratings on steam.

Edited by Gwei Loong, 13 June 2017 - 11:39 AM.

[Beaching Betty]

Gwei Loong, on 13 June 2017 - 11:36 AM, said:

Ok we know whats going on. I think most people understand this but it doesn't change whats happening in the case of disco's, people not playing the objective in domination, guys that go yolo, or groups that don't work together. Welcome to the solo que.



Maybe your right and in the end I'll keep on advertising the way I do and playing my games but in the end were not really the one's losing. Fact remains PGI is losing it's player base.
I am just suggesting we take a look at what Blizzard is doing and why millions of people play their games. You cannot say that one very successful company who wrote the book on this kind of marketing is not doing anything right and why argue your points with me when you can just look at the ratings on steam.

Thanks for the feedback, I agree on what you said here tho, I just did not fully read your post, I was just scrolling through the comment and saw "Losing makes you better" and then I quoted and thats it..

Edited by Beaching Betty, 13 June 2017 - 12:00 PM.

[Humpday]

I'm fairly certain the MM will try to trend you toward a 1:1 win loss best it can. I've noticed if i start getting above 1.3 ish, suddenly there is a lonnnnnnnnnng string of straight of rofl stomps. Not only that you van visually see a play style sway, the team will scatter all over the map for some unknown reason, where as a more experienced team of random players will generally stay close.

And you know how those go, 4-5 people get rolled in the first 2 minutes and if you're in anything other than a light or a fast medium that can quickly reposition, you're screwed, no matter how good you are. One can only carry so much every game, when you put up 3-4-5 kills with over 1k damage a the rest put up 200...its like ...well crap.

Meh...i think i'm used to it by now and no longer get surprised. What I do when i start noticing that, is a switch out of QA and go to group, and vice versa.

Oh btw, I didn't read the OP post...it was too long. So sorry if I appear to be rambling and completely off base.

Edited by Humpday, 13 June 2017 - 12:03 PM.

[Kaeb Odellas]

Gwahlur, on 13 June 2017 - 07:45 AM, said:

Well, if someone inflates their k/d, they'll be rated as a very good player, and thus get teamed with worse players, which over time will deflate their k/d again.

I still see k/d as the best metric to use

Getting the last hit in on an enemy doesn't make you particularly skilled, nor is it more likely to have you win. In some cases, pushing for the kill is the wrong thing to do.

For instance, if you strip an enemy of its last weapons in a brawl, spending any more time trying to kill it when there are other, more dangerous targets to engage isn't the best use of your time. A smart player switches targets when his current target is no longer a threat. A light mech that sneaks up and manages to cripple an enemy assault would be wiser to run away than to stick around and try finish his target and risk getting killed himself. Rushing a position in a fresh mech to secure a kill on a wounded enemy isn't a smart play when it his friends instagib you in response.

And the ability to kill isn't the be-all-and-end-all of player skill. A pilot with relatively mediocre piloting or gunnery skill but a fine grasp on tactics and communication can lead his team to victory without killing a single mech. An Assault pilot spearheading an assault and taking the brunt of the enemy's fire may not get a whole lot of kills, but his ability to absorb and spread damage can prolong his time as the primary target and allow his teammates to push in for the win.

[Gwahlur]

Kaeb Odellas, on 13 June 2017 - 01:01 PM, said:

Getting the last hit in on an enemy doesn't make you particularly skilled, nor is it more likely to have you win. In some cases, pushing for the kill is the wrong thing to do.

For instance, if you strip an enemy of its last weapons in a brawl, spending any more time trying to kill it when there are other, more dangerous targets to engage isn't the best use of your time. A smart player switches targets when his current target is no longer a threat. A light mech that sneaks up and manages to cripple an enemy assault would be wiser to run away than to stick around and try finish his target and risk getting killed himself. Rushing a position in a fresh mech to secure a kill on a wounded enemy isn't a smart play when it his friends instagib you in response.

And the ability to kill isn't the be-all-and-end-all of player skill. A pilot with relatively mediocre piloting or gunnery skill but a fine grasp on tactics and communication can lead his team to victory without killing a single mech. An Assault pilot spearheading an assault and taking the brunt of the enemy's fire may not get a whole lot of kills, but his ability to absorb and spread damage can prolong his time as the primary target and allow his teammates to push in for the win.

And that would balance itself out, as those people would get paired with less good players so they had less kills to "secure", meaning lower k/d, meaning they would end up at the skill level they should.

[Kaeb Odellas]

Gwahlur, on 13 June 2017 - 01:04 PM, said:

And that would balance itself out, as those people would get paired with less good players so they had less kills to "secure", meaning lower k/d, meaning they would end up at the skill level they should.

Why are you so tied to K:D as the defining matchmaking metric? K:D rewards builds and playstyles that are good at securing kills, not at securing wins. The objective of the game is to win, not just kill. K:D would unfairly punish any build that isn't "big pinpoint alpha" because those are the kinds of builds that secure kills.

[Gwahlur]

Kaeb Odellas, on 13 June 2017 - 01:30 PM, said:

Why are you so tied to K:D as the defining matchmaking metric? K:D rewards builds and playstyles that are good at securing kills, not at securing wins. The objective of the game is to win, not just kill. K:D would unfairly punish any build that isn't "big pinpoint alpha" because those are the kinds of builds that secure kills.

You have a point there about some builds getting more kills than others. Still, kills are more in an individual players hands than wins, so k/d is better than win/loss.
I can't really think of any build-independent metrics to use, can you?

[Kaeb Odellas]

Gwahlur, on 13 June 2017 - 01:39 PM, said:

You have a point there about some builds getting more kills than others. Still, kills are more in an individual players hands than wins, so k/d is better than win/loss.
I can't really think of any build-independent metrics to use, can you?

Winning a single match may not be entirely within your control, but you definitely have an influence on it, and over the course of dozens or hundreds of matches, you can easily see an individual player's ability to win.

[Mister Blastman]

Gwei, when you figure out how to use line breaks and paragraph breaks, I'll bother reading your posts. Until then, my eyes are ignoring them.

[BumbleBee]

I'd rather it be tied to AVG Match Score over last 25 games, per weight class.
It's not a perfect metric, but it may be better than what we have now.

The "XP Bar" would be better off trying to slowly push everyone into Tier 3, sorted by a bell curve. That way only the truly Tier 1's and 5's would stay where they belong.

Edited by BumbleBee, 13 June 2017 - 04:30 PM.

[LordNothing]

i dont really like that mm actually just fixes games to keep a 1:1 w/l.

in ye olde days of yore, if you found yourself loosing too much, you could always change servers and find players of your own skill level. when we had control over our servers you could have one for every skill bracket. but when you have forced 12v12 and a forced match maker you have no choice at all in anything. even the maps are pushed on you. maybe qp isnt where we need to focus efforts, id love to see more robust custom games with a lobby browser. that way when i want to play "24/7 terra therma skirmish no layzors t2 only", i can.

Edited by LordNothing, 13 June 2017 - 05:17 PM.

[Ghogiel]

PSR and current MMing seems to work. I top the leader boards in alts in various stats 100% solo in QKDs with a T3 account.

But then it kinda stops working once you max out the XP bar.

Elo was technically better when we had that but it doesn't feed good players wins and kills like PSR system, Elo was harder because you had to carry potatoes but weren't allowed to farm potatoes, unless the other side also had a player with max or near max elo, in which case it probably would have to fill both sides with some potato ratings to save the max Elo players from waiting 15mins for a match. .

Edited by Ghogiel, 13 June 2017 - 11:03 PM.

[Revis Volek]

Like other said, its kinda the way it goes with single spawn games. Snowballing may occur.



But the population, or lack there of, is another issue taxing any MM you use in this game.