1911 replies to this topic

[CyclonerM]

Heffay, on 10 July 2014 - 12:45 PM, said:

No, that is not your job as a Dire Wolf. Your job is to do what the drop commander tells you to do. If he says go sit on Sigma and defend it from lights, THAT is your job. If you win the match with zero damage doing what the DC tells you to do, you did your "job" far better than the Dire Wolf who did 1400 damage and lost the match.

And the person who does his job better more often over time has a higher Elo score to reflect that even though he is doing an average of 100 points of damage in a Dire Wolf, his 70% win percentage shows that he's doing it right.

That is a better player than a high scoring player with a 30% win rate.

I would say that the Locust's job is scouting, not what the DC tells you. If we talk about 'Mech role. Should you get a high Elo because your DC told you to use your Locust to lead a push or something like that?

[Cimarb]

Heffay, on 10 July 2014 - 12:45 PM, said:

No, that is not your job as a Dire Wolf. Your job is to do what the drop commander tells you to do. If he says go sit on Sigma and defend it from lights, THAT is your job. If you win the match with zero damage doing what the DC tells you to do, you did your "job" far better than the Dire Wolf who did 1400 damage and lost the match.

And the person who does his job better more often over time has a higher Elo score to reflect that even though he is doing an average of 100 points of damage in a Dire Wolf, his 70% win percentage shows that he's doing it right.

That is a better player than a high scoring player with a 30% win rate.

Sigh, no it is not. If your drop commander is dumb enough to tell 1 of his 3 assault mechs to guard Sigma from a light pack, which will more than likely obliterate a lone assault, then you have already lost the match.

If I am grouped - note my unit affiliation, so I DO drop as a group regularly - then I am in my Dire Wolf, and my job is to deal as much damage, and kill as many enemy team members, as possible. If I was assigned to a different lance, like the one with a Cicada, then my job would be different. Same with a Locust or whatever else you wanted to use as an example. I am often solo, though, and in those cases the "leadership stat" is very elusive. Success as solo usually has more to do with coordinated individualism than any actual commanding.

W/L is a measure of leadership and cooperation. I totally agree with that. As a soldier, my success should be judged on how I perform MY role, not how everyone else performed theirs.

If I am in a war zone and our leader makes a mistake and we get ambushed, our team is more than likely going to "lose" the engagement. My ability to succeed, by surviving, saving as many of my allies as possible, and killing enough enemies to make them hopefully retreat (or fending them off long enough for our own retreat), is what is going to earn me awards, praise, and honor, regardless of whether I survive myself or not.

If I die in combat, I hope that I will have my Elo increased through a Medal of Honor like I just described, rather than have it lowered just because some noob LT called artillery on our own position.

300 is the story of how the Spartans lost so amazingly that it became legend and changed the tide of war. Yet, in MWO it would have just been a big fat "L" and drop in Elo.

[Heffay]

CyclonerM, on 10 July 2014 - 01:10 PM, said:

I would say that the Locust's job is scouting, not what the DC tells you. If we talk about 'Mech role. Should you get a high Elo because your DC told you to use your Locust to lead a push or something like that?

If it leads to winning... absolutely.

Would you consider an Adder to be a scout mech just because it's a light? Should you judge it's ability to fulfill some arbitrary "scout" metrics just because it's the same weight as a Jenner?

Cimarb, on 10 July 2014 - 01:12 PM, said:

If I am grouped - note my unit affiliation, so I DO drop as a group regularly - then I am in my Dire Wolf, and my job is to deal as much damage, and kill as many enemy team members, as possible.

Yeah? And if your DC tells you to assault a fortified position in a suicide charge, knowing that you'll be lucky to get off one alpha before being destroyed in order to give time to flankers to get into position, did you help your team by doing very little damage, or hurt your team?

Your job isn't to do damage. Your job is to win.

[SirDubDub]

Still. If your team gets itself into a situation that requires a Dire Wolf to defend itself against lights, things have gone belly up.

[Heffay]

SirDubDub, on 10 July 2014 - 01:32 PM, said:

Still. If your team gets itself into a situation that requires a Dire Wolf to defend itself against lights, things have gone belly up.

Fine. How about using the Dire Wolf to just hold a cap point? Take away the "from lights" part.

The point is a Dire Wolf can contribute significantly to a victory just by doing the simple task of area denial, even if he does no damage.

[Wintersdark]

Cimarb, on 10 July 2014 - 12:33 PM, said:

I get that making a "better" system would be difficult. Making anything better is likely more difficult than taking the easy route.

You say this like PGI is just taking the path of least resistance.

Making a better system isn't just difficult, it is in some ways the holy grail of competitive ranking systems. Because they are [I]hard[/]. Not "Its a lot of work" hard, "mathematicians who are devoted to this sort of work haven't found a significantly better system" hard.

Quote

I am not saying W/L does not matter - it does - but it is not the ONLY factor that matters, nor is it a good representation of an individuals' skill, which is why we currently have brand new players being lumped in with players like me with over 4,500 matches (and I'm sure I am probably not up to the number of matches any of you three are).

Do you really understand what Elo is and how it is calculated and modified?

Its not based on your WLR. That's a simplified way to explain it, but that isn't what it is or how it works.

You can't just add things to it, because it is not a score derived from wins and losses.

So, no, WLR isn't everything. Its not even significant. What is significant is is whether teams with you on them beat teams ranked higher than yours more often than not.

A single game is never indicative of personal performance, and Elo has nothing whatsoever to do with single match performance. It is not a rating of your skill, either, though that too is a simplified way to look at it.

All Elo is, is a measure of how much your actions on average contribute to team victory. Its not a measure of how good a scout you are, because that is unimportant.

Maybe you dropped in a Locust. Ideally, you'd be scouting. A poor player will do that... Even if its unnecessary in the context of their match. Sure, he can spot for LRM's, but you have none. A good player will adapt, and switch roles on the fly, contributing to his teams victory instead of tilting at windmills. As Heffay suggested, a Direwolf sitting on an important cap point and doing zero damage [I]but discouraging a critical cap[/b] can have a significant impact on victory, whereas if he'd hared off and killed a couple mechs his stars would look better... But, oh, cap loss.

See, Elo isn't a system to tell you How Good You Are. We do need a system for that, mind you! But not for matchmaking.

All the matchmaker should care about is how much you contribute to victory, because it is ultimately the matchmaker's goal to strive for matches with very close odds of success either way.

Elo works for that. There is a substantial amount of research that had gone into it, including how it works when applied to team games, and the data has been overwhelmingly clear - it works. Its imperfect, yes. Its unreliable game to game, but it DOES lead to well ranked teams with correct odds for predicting the outcome of the match.

[Wintersdark]

CyclonerM, on 10 July 2014 - 01:10 PM, said:

I would say that the Locust's job is scouting, not what the DC tells you. If we talk about 'Mech role. Should you get a high Elo because your DC told you to use your Locust to lead a push or something like that?

Remember, Elo isn't a ranking of how awesome you are. There is a reason its private.

So "getting a high Elo" is no more a reward than getting a low Elo is a punishment.

Mech role is irrelevant. Only results matter. If you lead a push in your locust, and you where successful then absolutely yes, that's exactly what Elo is for - your taking actions that contribute to team victory. In that case, what has happened is you have made a correct judgement call - you'll probably die and get poor stats as a result, but that was what your team needed to win. Do that once successfully and 10 times unsuccessfully? Elo decreases (because you're stupid). Do that once successfully, scout successfully, beat your opponents, and your Elo score goes up...

... And in return, you get to fight other players who also make better judgement calls as a "reward". Elo in action.

This is why you can have a high Elo and be a poor pilot - you're not a good MechWarrior, but you help teams win.

Or you can have a low Elo score and be an outstanding pilot - maybe you just Rambo too much, and while you're death incarnate, you're never where you're team needs you to be.

Both of those cases represent Elo functioning 100% correctly.

[Void Angel]

Like what was just pointed out above - Elo is NOT determined by your win/loss ratio. It's determined by the probability of your match results compared to what actually happened. This allows your rating to change over time to more closely match your skill, thus placing you with others of the same general skill level.

[Cimarb]

Void Angel, on 10 July 2014 - 03:23 PM, said:

Like what was just pointed out above - Elo is NOT determined by your win/loss ratio. It's determined by the probability of your match results compared to what actually happened. This allows your rating to change over time to more closely match your skill, thus placing you with others of the same general skill level.

Totally a matter of phrasing. Elo goes up when you WIN despite the prediction, or LOSE despite the same thing. It is still determined by your W/L stat, but dependent on the predicted outcome. Very minor difference.

[Wintersdark]

Cimarb, on 10 July 2014 - 04:20 PM, said:

Totally a matter of phrasing. Elo goes up when you WIN despite the prediction, or LOSE despite the same thing. It is still determined by your W/L stat, but dependent on the predicted outcome. Very minor difference.

May, not does. If you beat a much lower ranked team, your Elo doesn't increase. This isn't an issue at the middle of the road, but it becomes significant at high and low Elo.

Regardless, Elo is an internal matchmaking system, not a rating of how "good" you are at anything in particular except winning. That's what its for, to match players of relatively equal "winningness" against each other, thus to bring predicted match results to 50:50 - basically, so that when you drop in a match, the outcome isn't basically predetermined.

[Void Angel]

See above; the reason people are on insisting on correcting this point is that it's incorrect - not poorly phrased.

[Cimarb]

Wintersdark, on 10 July 2014 - 04:57 PM, said:

May, not does. If you beat a much lower ranked team, your Elo doesn't increase. This isn't an issue at the middle of the road, but it becomes significant at high and low Elo.

Regardless, Elo is an internal matchmaking system, not a rating of how "good" you are at anything in particular except winning. That's what its for, to match players of relatively equal "winningness" against each other, thus to bring predicted match results to 50:50 - basically, so that when you drop in a match, the outcome isn't basically predetermined.

Void Angel, on 10 July 2014 - 05:29 PM, said:

See above; the reason people are on insisting on correcting this point is that it's incorrect - not poorly phrased.

Let me reset this a bit...

The matchmaker is currently borked big time. We have brand new players playing with those that have played for years. We have tourney teams playing against casual players. The current system, however you want to phrase it, is not working.

I want to play with against players that are of equal skill and/or experience as myself. I do not want to have to explain how to chainfire weapons, or what the button for air strike is. I want brand new players to play with people that are also learning, so they are spared the comments that come from us experienced players when they cannot even hit an Atlas walking by them.

I believe that basing our "Elo" on multiple statistics will make for a much better system of matching players. That, combined with using BV versus weight class, will make for a much better experience for everyone.

[Alaskan Nobody]

Cimarb, on 10 July 2014 - 06:01 PM, said:

We have brand new players playing with those that have played for years.

This bit is problematic (at best!)

Any evidence (or better yet) proof that those new players were not in a group with others?

Nothing any matchmaker can really do when that happens after all.
Either the new players get dragged all the way up, the old players get dragged "down", or they both get put in the middle
NONE of which are very good solutions.
One of which has to happen.
(Private matches excluding, of course )

No - I am not trying to say you are wrong
But that is an angle that should be considered before we decide that, as you put it:

Cimarb, on 10 July 2014 - 06:01 PM, said:

The matchmaker is currently borked big time.

Edit:
To clarify - the bulk of my matches before my current break were helping train newer players.
One of which announced he would no longer be dropping with me, because our combined Elo was taking him "too far up" (Actually it was just a bad set of game - yes those happen, even to me - but he somewhat overwhelmed at that point, and was not really listening to that)

Similarly - what Koniving does.

So long as new players CAN group with more experienced ones, what you are describing there IS going to happen.

Edited Edit: spelling, sorry for the spam

Edited by Shar Wolf, 10 July 2014 - 06:09 PM.

[Heffay]

Cimarb, on 10 July 2014 - 04:20 PM, said:

Totally a matter of phrasing. Elo goes up when you WIN despite the prediction, or LOSE despite the same thing. It is still determined by your W/L stat, but dependent on the predicted outcome. Very minor difference.

This isn't true. If the Elo disparity between the teams is high enough, it's possible to LOSE Elo points if you win against a much inferior team.

[Cimarb]

Shar Wolf, on 10 July 2014 - 06:05 PM, said:

This bit is problematic (at best!)

Any evidence (or better yet) proof that those new players were not in a group with others?

I am talking about the solo queue primarily, and Russ admitted (today?) that it is solo-only right now, and likely for the immediate future as well - no groups are currently dropping in the solo queue, at all.

[Gasoline]

I think we're on the edge of derailing the thread.

@Karl:

A question that hasn't been answered yet and I think is pretty important is this (Took me some time to dig it up again...). Sorry for full quote, but it's back on page 47 .

Koniving, on 08 June 2014 - 11:14 PM, said:

Hey Karl. Koniving here.
Reading the past couple of pages I see everyone has you pretty busy. I wasn't going to say anything, but apparently you would be better suited to talk to about this than mostly anyone. Since you're already looking into the crosshair anyway...

I've had a super simple solution to concentrated damage being too high and too perfectly focused. "Head bob" is what it's simply called. I'm sure you know what I'm talking about. In MWO the player's head never bobs. Some would complain of nausea or some other excuse to preserve their perfect point and click adventure.

Anyway to wrap this up quickly, in first person the aim is drawn from the pilot's head (not the camera but the head) or so the assumption goes. That or the body; I'm assuming the head. In third person there is a different point of reference, the cockpit. This is not in the same place as I am well aware. But in third person, the aim moves with the mech. If the mech is prone to hobble left and right, the aim will hobble left and right. If the mech bounces, the aim will bounce. If the mech jerks from behind being hit, the aim also appears to jerk from being hit.

This is already in the game in the third person perspective, but it'd be amazing if also applied in first person.

This would go a long way toward an alternative to delayed convergence that is 100% pinpoint, but not 100% "perfect." In a way it'd act similar to the "to hit" modifiers of tabletop for stationary shooting, walking while shooting, running while shooting and jump-jetting while shooting.

Two vids. Long one demonstrates it with several mechs and many ranges. During the Banshee I score multiple headshots through "timing" my shots with a full charge bounce. Short one is very simple, demonstrates it with a Jager at basically one range.

Long version.


Short version (for the time-constrained).


I believe almost every result of this concept would be beneficial to gameplay. Higher skill quotient for perfect shots while moving. Perfect hits rely heavily on 'timing' in a way that cannot be realistically achieved by a macro. Players slowing down for more 'guaranteed to hit' shots for those not so good with their 'gunnery' skill (lol). Mech crosshair movement patterns would be based on their animations, making them feel more unique.

(Edit: "Behind" in place of "being"....amazing what typos come up at 3:14 AM.)

[Alaskan Nobody]

Cimarb, on 10 July 2014 - 08:00 PM, said:

I am talking about the solo queue primarily, and Russ admitted (today?) that it is solo-only right now, and likely for the immediate future as well - no groups are currently dropping in the solo queue, at all.

Wait time breaks opening?

IE the filters try to prevent that - but as you wait longer the net gets wider?


I am thinking about what you said about adding more filters in - but I am not seeing how that would really work.
I do not see ANY of the other stats contributing to the team's victory, and what are we fighting for if not to win?

Kills?
High kill score does not mean you helped a lot.
It helps more than the other stats - but then Kill Stealing actually becomes a problem

Damage?
Low total damage is frequently better (more efficient kills)
It is also a sign of not having helped much

Survival?
Then you have the problem of running and hiding being the defacto method at the end of matches

A combination of them all?
Well since NONE of them actually mean anything on their own it might work....

But then which end of damage should they be aiming for?
Damage/Kills?
Great - I run out in my locust, wait till everyone is almost dead (hiding at the far end of the map to ensure I live that long) then vulture.
THAT will surely help my team!

Kills/Assists?
Suicide farming!



Should they screw their team over for survival, or suicide rush?

Only stat that actually MEANS you helped - is whether your team won or not.

All the damage means nothing if you lost
All the KILLS means nothing if you lost.







Which brings up an interesting thought. (probably not directly brings up though)


You can place all the filters you want - all the BLOCKS you want.
But if there are not enough people in the que, those filters HAVE to be lifted, or people will stop playing.

The problem you present - is probably not going away regardless of what kind of matchmaker we have, or what filters it has.

IE: It takes to long to find a match!

[Goose]

So we had our first sale on 'Mechs in , right? How'd that turn out?

[Jody Von Jedi]

Goose, on 11 July 2014 - 12:35 PM, said:

So we had our first sale on 'Mechs in , right? How'd that turn out?

I bought 3 Atlas that weekend, but I may have been the exception.

[Bilbo]

Jody Von Jedi, on 11 July 2014 - 12:39 PM, said:

I bought 3 Atlas that weekend, but I may have been the exception.

Nope. I bought several as well.

Edited by Bilbo, 11 July 2014 - 01:00 PM.