226 replies to this topic

[w4ldO]

PhoenixFire55, on 17 January 2020 - 08:04 AM, said:

That is your opinion and nothing more. A matchmaker that produces close, balanced matches is a proper matchmaker. Distribution of player skill across teams doesn't matter.

from the same paragraph

w4ldO, on 17 January 2020 - 06:53 AM, said:

if there are not enough players of equal skillset the matchmaker should balance out the teams.

could you please enlighten me on how the MWO matchmaker achieves such balance?

[PhoenixFire55]

w4ldO, on 17 January 2020 - 08:22 AM, said:

could you please enlighten me on how the MWO matchmaker achieves such balance?

First, your question has literally nothing to do with what we've discussed and just changes the subject for no reason other than you having nothing to say.

Second, it is impossible to say whether MWO matchmaker actually does or doesn't do anything worthwhile because the metric on which it is based (PSR) is broken by design.

Third, this topic is about Jarl's list. If you want do discuss matchmaker and how it works in detail then go and read up one of the 50+ threads about matchmaker on this forum. I don't see any point for anyone to repeat what we've been saying for years here yet again.

[Prototelis]

It only takes 24 players to make one balanced 12v12 match.


But anyways, the idea that someone who loses more than they win and has a low average match score to boot is worth listening to when talking about gameplay/strats is super lol.

Edited by Prototelis, 17 January 2020 - 10:18 AM.

[Marquis De Lafayette]

I think we can all probably agree that the matchmaker is so broken that it produces all kinds of matches, for everyone. Some matches have close to balanced sides (skill wise) and some matches are totally unbalanced, where you look at the other side and think “c’mon, why on earth are all those top guys on that same side” or vice versa where you think “man, I better get to work early to score even 400 damage...with all these good guys here on my side we have”.

Jarls accounts for all this, because the same stuff happens to everyone....the top guys are still the top guys with All the same stuff happening to them.

[w4ldO]

PhoenixFire55, on 17 January 2020 - 08:04 AM, said:

That is your opinion and nothing more.

it is also the opinion of numerous game developers where ELO systems are used. just look at the skill group graining CS:GO offers.
if you were to balance the teams by lets say average jarls rank you would make a few top players carry a bunch of bad players against a team full of average players.
obviously the best balanced match is one where everyones skill is so close to each other you would not be able to tell any difference.

PhoenixFire55, on 17 January 2020 - 09:50 AM, said:

First, your question has literally nothing to do with what we've discussed and just changes the subject for no reason other than you having nothing to say.

Second, it is impossible to say whether MWO matchmaker actually does or doesn't do anything worthwhile because the metric on which it is based (PSR) is broken by design.

Third, this topic is about Jarl's list. If you want do discuss matchmaker and how it works in detail then go and read up one of the 50+ threads about matchmaker on this forum. I don't see any point for anyone to repeat what we've been saying for years here yet again.

1. it has very much to do with what we discuss. with a working matchmaker the top players would face each other. therefore those players could not even rake up the stats like they do now, only the very top players could exceed the average numbers of the highest skill group. a working matchmaker most likely would render jarls in its current form near useless.
2. looks like you just made it possible. the matchmaker cannot do something worthwhile thanks to garbage in garbage out. tier system should be zero sum and reset to have any chance at fixing the matchmaker. but that would mean PGI did something.
3. i'd rather not, since i dont expect any answer to be found in those topics due to lack of communication and explanation from devs. but i sort of have a theory on how the matchmaker operates.

well ok then. lets get back on topic. not like it hasnt been discussed before, right?
so you want to take jarls as a measurement wether someones tactics has merit or not? fine do as you want.
but where do you draw a line for that? 98 percentile 95 percentile, 90 percentile, 85 percentile, 80 percentile? are you only ever take advice from someone being higher on the list than you are?

Prototelis, on 17 January 2020 - 10:16 AM, said:

But anyways, the idea that someone who loses more than they win and has a low average match score to boot is worth listening to when talking about gameplay/strats is super lol.

is this like ultra-partisan politics where a great idea of the ""wrong"" person is automatically deemed to be idiotic and false, while an absolutely bonkers idea of the ""correct"" person is treated like a gift from heaven? i hope not.
wether an idea could work or not you have to evaluate yourself. but if you blindly follow the mantra of "your stats suck therefore your ideas suck and i wont ever consider them" you are making a grave mistake.

imho jarls suffers from the broken matchmaker, too. you constantly get paired up with players you should not face.
it is only due to a large sample size for jarls to function as we know it and to hopefully even out the oddities you might see during your matches. and this game is very special when it comes to the amount of random B.-S. that can happen during a match. (compared to other shooters)

Marquis De Lafayette, on 17 January 2020 - 10:21 AM, said:

I think we can all probably agree that the matchmaker is so broken that it produces all kinds of matches, for everyone. Some matches have close to balanced sides (skill wise) and some matches are totally unbalanced, where you look at the other side and think “c’mon, why on earth are all those top guys on that same side” or vice versa where you think “man, I better get to work early to score even 400 damage...with all these good guys here on my side we have”.

Jarls accounts for all this, because the same stuff happens to everyone....the top guys are still the top guys with All the same stuff happening to them.

i would say it is not the "same stuff" that happens to everyone. rather the top notch players ARE the "stuff" that happens in your example.

Edited by w4ldO, 17 January 2020 - 01:10 PM.

[PhoenixFire55]

w4ldO, on 17 January 2020 - 12:45 PM, said:

1. it has very much to do with what we discuss. with a working matchmaker the top players would face each other.

And again, no. With a matchmaker working according to your opinion they would. With another properly working MM they wouldn't. Stop shoving your opinion in everyone's faces. There are many ways to make a properly working matchmaker and you will not be able to prove that yours is a better one.

More importantly, this have nothing to do with the question raised in the thread.

I will, in case someone doesn't understand, emphasize my main point once again. It doesn't matter what anyone thinks, doesn't matter what excuses, utter silliness and conspiracy MM theories one comes up with, a 90%+ Jarl's player is simply ways better than a <50% Jarl's player. Shared accounts, niche builds, yadda yadda, it all does not matter. He is simply better in a QP environment as clearly displayed by statistics you can't artificially farm, i.e. winning. And outside of the QP a 90%+ player will flat out wipe the floor with a <50% player in a 1v1 fight, even if he is drunk and playing with one eye closed. You can discuss a swing of ~5% due to factors such as running/not running meta and so on, but a Jarl's difference of more than that means one thing and one thing only, one guy is simply a better player than the other guy.

And going back to the title of the thread, i.e. Jarl's =/= knowledge, it goes exactly the same way. I would expect a 65% player and a 70% player to have a relatively similar degree of knowledge and expertise about this game. But there is absolutely no doubt whatsoever, that a 95% player has incomparably more knowledge, expertise and skill gameplay-wise than a 55% dude.

The most important thing to remember, is that when a guy has a career W/L of under 1.0, it means that whatever he does, it actively hurts his team chances to win, thus listening to any kind of advice given by such player is quite stupid.

[w4ldO]

PhoenixFire55, on 17 January 2020 - 01:11 PM, said:

And again, no. With a matchmaker working according to your opinion they would. With another properly working MM they wouldn't. Stop shoving your opinion in everyone's faces. There are many ways to make a properly working matchmaker and you will not be able to prove that yours is a better one.

you fail to realize that this games tier system would support such a matchmaker if PSR was a zero sum system and not upwards tendency.
this game uses PSR tiers to categorize players and it wants to use those tiers as skill groups for matchmaking.
someone put the code into the matchmaker to pair players of similar tiers in one match (we know that from news posts).
looks like PGI agrees with my idea of a matchmaker. they just got the PSR thing wrong as i cannot imagine PGI putting in an upward tendency towards tier 1 by intent.

just because somehow an MM is able to create a match that is somewhat balanced (a few top players carry a bunch of bad players against a team full of average players f.e.) doesn't make it a good MM. only a working one.
yet this game lacks even that.

would you rather play a match where the personal skill of all players is very close to eachother or play a match where the average skill between the two teams is equal? the latter comes with a plethora of consequences, the former does not.
most games with a matchmaker try to match up players with similar skill, hence the word matchmaker. it makes a match for people that should fit together.
also it is generally accepted that you can only progress in skill when you play against someone that is better than you rather then slaughter noobs all day.


proof enough? stop shoving your face into my opinion!

PhoenixFire55, on 17 January 2020 - 01:11 PM, said:

95 > 55

according to my research your math checks out.
but noone is arguing against that. this topic has been discussed ad nauseam and this thread was over from post 1. i'm raising the question wether we would have anything like jarls if the game was working like discribed (see the MM stuff above)

PhoenixFire55, on 17 January 2020 - 01:11 PM, said:

The most important thing to remember, is that when a guy has a career W/L of under 1.0, it means that whatever he does, it actively hurts his team chances to win, thus listening to any kind of advice given by such player is quite stupid.

career W/L is a very dumb measurement. that would mean you dont take progression into account. a player that sports a 2.0 W/L over the last 4 seasons but below 1.0 career W/L is still a liability by that logic.
also do you automatically discard every idea that comes from a player under your threshold (whatever it might be) without even thinking about evaluating it yourself?

Edited by w4ldO, 17 January 2020 - 03:56 PM.

[justcallme A S H]

Khobai, on 16 January 2020 - 09:37 AM, said:

and your point is?

jarls was not completely reset since people stopped playing in group queue. those stats from when people played in group queue are still very much there.

besides group queue isnt 100% dead and most group queue matches are stomps now. so thats continuing to taint the data. thats always been the problem with jarls list.

jarls list should only should solo queue stats. that would reflect an individuals skill way better.

Well given you don't play - you have no idea what you're talking about as usual.

That's the point.

Plus it is EASY to discern when GroupQ vs Solo stat players in Jarls. Even when people cross over between the two. You just need to know what your looking at. I mean we know you don't understand numbers, I've proven you wrong on so many occasions it's almost sad.

[justcallme A S H]

_Magno_, on 16 January 2020 - 12:23 PM, said:

As an example;
someone like RWTumbleweed does not have the stats as some here would insist is paramount to be listened to. But I am not listening to RWTumbleweed for mech builds or design; he consistently provides valuable information and comms in game that are NOT directly reflected in his stats.

He should NEVER be listened to.

Anyone that does means you're generally going to lose. As that is what he does, causes teams to lose.

Why? Because nothing, I repeat, NOTHING he says on comms is of any value of use to affect a win. It actively causes losses because people listen to a bad player telling people bad things over comms resulting in losses.

It's no more simple if difficult than that.

If you think he is good/useful - you're part of the problem if you cannot see how badly and negatively he impacts a match.

[RickySpanish]

Those stats are truly impressive. Yet I don't think I can ever recall a game where he has been on my team. We must play in different time zones given his sheer volume of games.

[FRAGTAST1C]

w4ldO, on 17 January 2020 - 03:55 AM, said:

Jarl does not display skill. rather it displays what numbers you can manage to archive using your personal skillset (stats are cheeseable afterall)
top players perform well in any match? what about when they face other top players? one team has to lose for the other be able to win, remember that
in a world where we had a proper matchmaker with even teams, top player would only face other top players.
the expected result of that would be top players having a 1.0 W/L

1. Stats can be cheese'd and you can clearly see who's doing it. It's evidently clear.

2. When top Players face other top players, their averages will still be better than most of the average players. Can you honestly tell me how many times an average player faced equal odds and did horribly? It's a lot more than when top players face each other.

3. There are many top players who are also streamers. They take pretty much anyone into their group and if they start winning constantly, they'll change sides in FW, still with their viewers in the group and still win. Not just win but also carry the team.

[justcallme A S H]

RickySpanish, on 17 January 2020 - 05:10 PM, said:

Those stats are truly impressive. Yet I don't think I can ever recall a game where he has been on my team. We must play in different time zones given his sheer volume of games.

Be thankful... I sometimes get lumped... It's painful, extremely painful as it is an uphill battle when the team is sent off to Zimbabwe cause of atrocious calls and "information".

[Marquis De Lafayette]

w4ldO, on 17 January 2020 - 12:45 PM, said:

i would say it is not the &quot;same stuff&quot; that happens to everyone. rather the top notch players ARE the &quot;stuff&quot; that happens in your example.

C’mon man. Anyone who is experienced in this game is going to pretty much have to get thrown in matches that are going to have tier 1 players of various quality. Top, good, average, bad and awful....plus all the minor gradations inbetween. If the top guys didn’t get thrown in with everyone else then the merely good (players) would just kick the stuffing out of everyone worse and then they would start thinking they are better than they are. At some point you have enough experience to play only “big boy/girls” (experienced players) and you have to just try to get better to have better results. There always has to be a top dog in games like this....maybe you are just disappointed it’s not you....idk...but you have to accept that some guys are just gifted at this and try to be the best you can be. I am not a top guy like some guys on this thread, but I am good enough to enjoy the challenge of trying to find a way to compete and win matches I am matched up against top level talent. Embrace the challenge and just learn what you can from facing some of these dudes, it will make you a better player.

[Dee Eight]

w4ldO, on 17 January 2020 - 02:57 PM, said:

career W/L is a very dumb measurement. that would mean you dont take progression into account. a player that sports a 2.0 W/L over the last 4 seasons but below 1.0 career W/L is still a liability by that logic.
also do you automatically discard every idea that comes from a player under your threshold (whatever it might be) without even thinking about evaluating it yourself?

Dumb players generally ignore progression because they're just that...dumb. Facts like progression get in their way of stat shaming others.

[MrMadguy]

_Magno_, on 16 January 2020 - 07:37 AM, said:

This is an unfortunate position to have. I've read plenty of responses that read as, "You have a 50% percentile on Jarl's list, so stop talking like you know things."
or
"Don't listen to this guy, he's ranked low on Jarl's list."
or
"pay attention to this guy, he's ranked high on Jarl's list.

I love Jarl's list, I love data, statistics, but I am not going to judge a player's value by Jarl's list.

Plenty of stats that will never be known from simple reference to Jarl's list.

1. The game does not capture information on whether you are tinkering around, letting friends and children play on your account, etc.. My kids play on my account, they don't suck, but they don't pad out in the positive direction either. I'm not going to limit sharing the enjoyment of fighting in giant robots to protect a standing that is used to distill humans into a simple value.

2. Focus on command work. I.e, if you're running around in an Artic Cheetah with ECM, TAG and 1 large laser, you will rarely get above 300 damage even if you ONLY focus on damage and rarely get kills so your kill count and score will not go up, but you could be doing a fabulous job spotting and communicating information to make your company mates more effective.

3. Playing all the mechs. I've played dozens of games with the Viper, Mist Lynx, etc., and NOT with meta builds. Plenty of better mechs. If I continue to challenge myself, enjoy something different, understand what truly makes for a good vs bad mech through experience/gameplay, my stats are going to be suppressed. But my knowledge is not.

4. Gaming the statistics. Plenty of ways to do this.
- Only play in meta-build mechs
- hang back and wait for team mates to soften opponents, then wake up to clean up on open torsos late game. You folks are junk humans and I hope all your ice cream tastes of swampy pork.
- Always select scenario and map locations suited for your build. (I love selecting hot maps and conquest with my slow *** laser vomit Thunderbolts).

Jarl's list is fun to peruse. I'm glad that PGI collects this information. But we should be careful how we drive this information to mean false things.

I've got tired of telling, that may be AvgMS correlate with skill, but doesn't actually represent it. Because 500MS in Tier 5 and 500MS in Tier 1 - are two completely different things. Because when you're skilled and you play against noobs - you can earn higher AvgMS, than when you play against players with equal or higher than yours skill. Because your enemies and teammates allow you to do it. Enemies let you kill them without any resistance and teammates don't steal your kills. Only true skill rating would do it. And current we don't have one.

Edited by MrMadguy, 18 January 2020 - 01:00 AM.

[Ssamout]

MrMadguy, on 18 January 2020 - 01:00 AM, said:

I've got tired of telling, that may be AvgMS correlate with skill, but doesn't actually represent it. Because 500MS in Tier 5 and 500MS in Tier 1 - are two completely different things. Because when you're skilled and you play against noobs - you can earn higher AvgMS, than when you play against players with equal or higher than yours skill. Because your enemies and teammates allow you to do it. Enemies let you kill them without any resistance and teammates don't steal your kills. Only true skill rating would do it. And current we don't have one.

Yawn.

Tier 5 accounts are new accounts and that can be seen in the stats. If you average 500ms in tier 5 you are not in tier 5 for long, then you rise to general population and if your only way to get 500ms average is by harvesting spring potatoes your stats will show this.

And tier 1 is full of noobs.

[FRAGTAST1C]

MrMadguy, on 18 January 2020 - 01:00 AM, said:

I've got tired of telling, that may be AvgMS correlate with skill, but doesn't actually represent it. Because 500MS in Tier 5 and 500MS in Tier 1 - are two completely different things..................

Find out how quickly you will reach T3 and start playing against T1s if you get 500MS as a T5. You'll be surprised at how bad your argument is up there.

[w4ldO]

FRAGTAST1C, on 17 January 2020 - 06:37 PM, said:

2. When top Players face other top players, their averages will still be better than most of the average players. Can you honestly tell me how many times an average player faced equal odds and did horribly? It's a lot more than when top players face each other.
3. There are many top players who are also streamers. They take pretty much anyone into their group and if they start winning constantly, they'll change sides in FW, still with their viewers in the group and still win. Not just win but also carry the team.

2. yet when top players face other top players you should see their magical performance numbers on jarls go down compared to when they face worse players. if a match is filled with top players, someone has to eat the curb making them perform (in term of numbers) worse than they used to against mediocre players.
3. well yes. the only thing that changed is the faction side they play on. there is no reason for them to perform different.

Marquis De Lafayette, on 17 January 2020 - 07:29 PM, said:

C’mon man. Anyone who is experienced in this game is going to pretty much have to get thrown in matches that are going to have tier 1 players of various quality. Top, good, average, bad and awful....plus all the minor gradations inbetween. If the top guys didn’t get thrown in with everyone else then the merely good (players) would just kick the stuffing out of everyone worse and then they would start thinking they are better than they are.

At some point you have enough experience to play only “big boy/girls” (experienced players) and you have to just try to get better to have better results. There always has to be a top dog in games like this....maybe you are just disappointed it’s not you....idk...but you have to accept that some guys are just gifted at this and try to be the best you can be. I am not a top guy like some guys on this thread, but I am good enough to enjoy the challenge of trying to find a way to compete and win matches I am matched up against top level talent. Embrace the challenge and just learn what you can from facing some of these dudes, it will make you a better player.

1. i mentioned the upward tendency of the PSR system that is one of the underlying problems this game has. those top players we are talking about belong into tier 1. the "bad and awful" ones clearly dont.
2. i guess it is great that i also mentioned this: "also it is generally accepted that you can only progress in skill when you play against someone that is better than you rather then slaughter noobs all day."

Ssamout, on 18 January 2020 - 01:41 AM, said:

Tier 5 accounts are new accounts and that can be seen in the stats. If you average 500ms in tier 5 you are not in tier 5 for long, then you rise to general population and if your only way to get 500ms average is by harvesting spring potatoes your stats will show this.

wasn't there some sort of forum post a là "how many games till tier 1"?

[FRAGTAST1C]

w4ldO, on 18 January 2020 - 02:34 AM, said:

2. yet when top players face other top players you should see their magical performance numbers on jarls go down compared to when they face worse players. if a match is filled with top players, someone has to eat the curb making them perform (in term of numbers) worse than they used to against mediocre players.
3. well yes. the only thing that changed is the faction side they play on. there is no reason for them to perform different.

Are you worried about making sure THEIR stats tank instead of improving yours? You can do that by the following methods....

1. Stop dying within the first minute of enemy contact

2. Get 500 damage minimum

3. Get a kill or two

[w4ldO]

MrMadguy, on 18 January 2020 - 01:00 AM, said:

I've got tired of telling, that may be AvgMS correlate with skill, but doesn't actually represent it. Because 500MS in Tier 5 and 500MS in Tier 1 - are two completely different things.

Ssamout, on 18 January 2020 - 01:41 AM, said:

Tier 5 accounts are new accounts and that can be seen in the stats. If you average 500ms in tier 5 you are not in tier 5 for long, then you rise to general population and if your only way to get 500ms average is by harvesting spring potatoes your stats will show this.

FRAGTAST1C, on 18 January 2020 - 02:34 AM, said:

Find out how quickly you will reach T3 and start playing against T1s if you get 500MS as a T5. You'll be surprised at how bad your argument is up there.

am i lacking reading comprehension or are you guys talking into the same direction?