126 replies to this topic

[Bishop Steiner]

El Bandito, on 23 September 2015 - 04:00 AM, said:

Post your own average WLR, KDR and XP earned per match first. They are all on you stats page. Meanwhile I shall call you Ti-fiveium.




That's cause good players win more consistently.

not a given in solo drops. Team queue, yes. Seen and been in many solo drops where good player might go down swinging but realistically had no chance of winning
(using today's system, which of course, supposedly would not allow this unbalance, it would expressed as a Tier 1 dropping with 11 Tier 4 and 5s, facing a team composed of Tier 3 and 2s. During the Elo era, this was a pretty common occurrence in solo queue)

[Mawai]

El Bandito, on 23 September 2015 - 04:00 AM, said:

Post your own average WLR, KDR and XP earned per match first. They are all on you stats page. Meanwhile I shall call you Ti-fiveium.


That's cause good players win more consistently.

"That's cause good players win more consistently."

No, If the matchmaker is working properly then good players will be matched against good players and they will win exactly 1/2 the time over the long term (unless they keep upping their game and keep getting better ... but there is usually a limit to that).

If you find that you are constantly winning more than losing then you are constantly getting better (which is awesome) or the matchmaker or ranking system is broken and constantly matches you against folks who aren't actually as good as you are.

[Mawai]

Tarl Cabot, on 23 September 2015 - 05:24 AM, said:

Understand that your PSR, which was seeded from your average Elo (which was per weight class), is now one number regardless of weight class.

The Elo seeding initially put you in Tier 5.

PSR primary component is determined on if you won or lost the match. MWO is a team-focus game and movement up depends on how that co-op team performs together. If for both sides there is no real communication (more than just Hi, lock targets, stay together) and the team as a whole does not have any situational awareness then it will be a crap-shoot on who wins. Primary ingredient to winning is one side being slightly more focused/organized/communicative via chat/VOIP and some luck. It is unlucky to turn a corner and run into 4 opponents and dying to focus fire, it is lucky if running into and through same opponents and not be rip to shreds. Add communication and/or situational awareness of your team who may have been close behind you, upon seeing the red doritos, they rush the four and takes them out.

After that it is what you did in the game, as noted PSR takes in several factors, and PSR puts a different weight on each action that PGI has not released. Damage is simply only one component.

The group game. It is similar to the above. If there is no real communication between the groups, it can fall to pieces quickly for one side. Focus fire will tend to be more prevalent higher in tiers you go, thus mechs will drop more quickly. I would say that a Tier 1/2 solo queue could look more like a match in the group queue.

Hi. Nope. Elo was not used for tier seeding. (Unless you have a quote where a dev said otherwise?)

"All current players will be seeded into whichever Skill Tier best aligns with the historical data already present in their account. The historical data pool used for generating your Tier placement goes back until January 2015."

Basically they took all the game results for every player since January 2015 and calculated what tier you should be in based on that data. Whether they seeded everyone as newbs and then applied the PSR change based on every match played since January 2015 or some other method isn't clear but it was based on performance data from this year rather than historical Elo rating.

If you haven't played that much this year or were concentrating on leveling up mechs then that may help explain a lower level tier since it will be more strongly influenced by the reference point that PGI used for evaluating the starting point in their intiial tier calculation when using the historical data (i.e. if everyone was started off in tier 4 as if they were newbs and historical data then applied to determine tier then the number of games played since January 2015 might be a significant factor in the current tier/PSR assignment.)

Edited by Mawai, 23 September 2015 - 07:03 AM.

[El Bandito]

Mawai, on 23 September 2015 - 06:49 AM, said:

"That's cause good players win more consistently."

No, If the matchmaker is working properly then good players will be matched against good players and they will win exactly 1/2 the time over the long term (unless they keep upping their game and keep getting better ... but there is usually a limit to that).

If you find that you are constantly winning more than losing then you are constantly getting better (which is awesome) or the matchmaker or ranking system is broken and constantly matches you against folks who aren't actually as good as you are.

You need to check out dictionary bro, cause the words "consistently" and "constantly" do not mean the same thing.

Bishop Steiner, on 23 September 2015 - 06:48 AM, said:

not a given in solo drops. Team queue, yes. Seen and been in many solo drops where good player might go down swinging but realistically had no chance of winning
(using today's system, which of course, supposedly would not allow this unbalance, it would expressed as a Tier 1 dropping with 11 Tier 4 and 5s, facing a team composed of Tier 3 and 2s. During the Elo era, this was a pretty common occurrence in solo queue)

Good players will still do better than average in Solo-Q overtime. IMO, Group-Queue is less accurate in measuring a person's individual skill, cause he could be riding on a good group's coat tail all the time.

I respect people with good Solo-Q WLR more than those with good Group-Q score, cause that seems to indicate his personal skill more.

Edited by El Bandito, 23 September 2015 - 07:07 AM.

[MechPorn]

As with everything in MWO (this is not a jab at PGI or their system) it is nothing but a grind. You want to improve your overall tier, get in a group of 11 like minded players, obtain TeamSpeak (easiest place to get a room is NGNG) and grind with them.

The biggest difference between tier 5 ad tier 1 is that the tier 1 players are mostly in coordinated groups that work/practice (calling targets, giving roles and communicating) together and tier 5 is not.

[Sjorpha]

This is might be a good example that winning is still the main thing, and all the doom and gloomers who thought damage padding would work to boost your tier are wrong, which is exactly as it should be.

How effective are those 4-800 damage? How good are you at tanking and playing aggressive enough.

Assault LRM boats would be a prime example of a mech that can put out those numbers while still putting you in tier 5, because it's going to cause your team to get rolled up if you hide all that armour in the back instead of tanking your share in the front like you should.

Edit: Not meant as a jab at Titannium, I have no idea how he actually plays. But the description of assaults putting out huge damage "every other game" while still being ranked low is consistent with how LRM assaults should fare in an accurate PSR system. If it is the case this is good news.

Edited by Sjorpha, 23 September 2015 - 07:13 AM.

[Lugh]

I wouldn't sweat it OP. I restarted playing in May and am only just barely in to tier 3.

[Mawai]

El Bandito, on 23 September 2015 - 07:02 AM, said:

You need to check out dictionary bro, cause the words "consistently" and "constantly" do not mean the same thing.


Good players will still do better than average in Solo-Q overtime. IMO, Group-Queue is less accurate in measuring a person's individual skill, cause he could be riding on a good group's coat tail all the time.

I respect people with good Solo-Q WLR more than those with good Group-Q score.

Sorry I know what "consistently" means.

In the context you used it ... I interpreted it to mean that good players will win more than 1/2 the time. Was that incorrect? Are you saying that winning "consistently" means that you win as much as you lose? Winning constantly would imply winning every match.

The problem here is this. You are a good player. If the match maker puts someone just as good as you are on the other team ... which it is SUPPOSED to do? No? In that case, then both players are "good" ... they both win "consistently" which MEANS that over time they will win 1/2 of their matches.

It is simple math. It has nothing to do with how good the players are but if you have good players on both teams ... whether it is solo or group, then they will WIN pretty much exactly 1/2 their matches. IF you or your team wins more than 1/2 of the matches ... i.e. wins "consistently" ... at least that is how I interpret the word ... then you are being matched against players that are NOT as good as you since if they were as good then you would win 1/2 the time.

I hope that is clear. If good players win more than 1/2 of their matches then something is broken ...

1) The matchmaker does not have a good number for the skill level of the good player and is constantly underestimating them
2) The good player is constantly getting better leading to (1) above
3) The ranking system is broken so that the number the matchmaker is using does NOT actually represent the player contribution to WINNING the match so they will win more than expected.
4) The matchmaker is broken and fails to put two teams of "good" players into the match on a regular basis.

Any of the above will result in a player getting more wins than losses ... BUT ... if the ranking system is accurate and the matchmaker forms even matches then you will will 1/2 and lose 1/2.

For example, you can take the 24 best players in MWO at the highest level of tier 1 ... put them into a match ... and you know what? 1/2 will win and 1/2 will lose ... consistently .. in every match. Mix them up, throw them into match after match, the absolute best players ... and 1/2 will win and 1/2 will lose. If they are all ranked about the same then over time their W/L for each and every one of them will be 1/1.

[Lily from animove]

Titannium, on 23 September 2015 - 06:24 AM, said:

yep. because it count mine first ever 100-150 trial/beginner matches with 50-100 damage score.
That part , you "old" players, already skipped by starting 2015 stats with mechs I even dont have now.

Seems fair to me.
PGI, way to go. Discriminating new players for being new.

there should be apossibility for a palyer to "reset" his stats, the current tracked stats then would go into archives stats where they will pile up over the entire career, and a Player can then start freshly on that statspage. Would help trackign stats of new builds and compare them.

Imperius, on 23 September 2015 - 06:45 AM, said:

It's ok I'm lower tier 3 and I have no idea why or do I care. Speculation says it's total games makes the biggest difference.

yes thats your problem, not enough game,s come bakc in 1000 more games and you are probably mid or high T2.

Mawai, on 23 September 2015 - 07:21 AM, said:

For example, you can take the 24 best players in MWO at the highest level of tier 1 ... put them into a match ... and you know what? 1/2 will win and 1/2 will lose ... consistently .. in every match. Mix them up, throw them into match after match, the absolute best players ... and 1/2 will win and 1/2 will lose. If they are all ranked about the same then over time their W/L for each and every one of them will be 1/1.

not necessarily, this only happens when those top 24 would have the same skill. lets say 3 of them have exceptonal better skills and 3 of them somehwat lower, but still all 24 are the best 24. then depending ont he randomness of MM the 3 exceptionals will have a higher W/L and the lower 3 a lower W/L than 1/1

because the system has a top and a bottom and when you are on the top there is nowhere to go "further" to achieve the 1/1 by having correct matchhing opponents.

Same goes for the lower end.

Edited by Lily from animove, 23 September 2015 - 07:29 AM.

[fat4eyes]

Mawai, on 23 September 2015 - 06:49 AM, said:

"That's cause good players win more consistently."

No, If the matchmaker is working properly then good players will be matched against good players and they will win exactly 1/2 the time over the long term (unless they keep upping their game and keep getting better ... but there is usually a limit to that).

If you find that you are constantly winning more than losing then you are constantly getting better (which is awesome) or the matchmaker or ranking system is broken and constantly matches you against folks who aren't actually as good as you are.

Another possibility is that you bring something to the battle that match score can't quantify (or doesn't quantify well) and so doesn't feed into the matchmaker. I'd think that good commanders/tacticians fall here. Maybe scouts as well, but to a lesser extent because scouting bonuses do feed into match score. Basically any skill that allows your team to work together better than the other team, in a "whole is greater than the sum of its parts" way.

Mawai, on 23 September 2015 - 07:21 AM, said:

For example, you can take the 24 best players in MWO at the highest level of tier 1 ... put them into a match ... and you know what? 1/2 will win and 1/2 will lose ... consistently .. in every match. Mix them up, throw them into match after match, the absolute best players ... and 1/2 will win and 1/2 will lose. If they are all ranked about the same then over time their W/L for each and every one of them will be 1/1.

You're assuming that they're all equally skilled. If one of them is so far above the others then the team that guy is on will win more often. Also true if one of them can lead better than the others, because whichever team he's on will win more.

Edit: reply to another post

Edited by fat4eyes, 23 September 2015 - 07:31 AM.

[Bishop Steiner]

Sjorpha, on 23 September 2015 - 07:09 AM, said:

This is might be a good example that winning is still the main thing, and all the doom and gloomers who thought damage padding would work to boost your tier are wrong, which is exactly as it should be.

How effective are those 4-800 damage? How good are you at tanking and playing aggressive enough.

Assault LRM boats would be a prime example of a mech that can put out those numbers while still putting you in tier 5, because it's going to cause your team to get rolled up if you hide all that armour in the back instead of tanking your share in the front like you should.

Edit: Not meant as a jab at Titannium, I have no idea how he actually plays. But the description of assaults putting out huge damage "every other game" while still being ranked low is consistent with how LRM assaults should fare in an accurate PSR system. If it is the case this is good news.

just don't think the W/L should really be weighted heavily on individual performance.

Can't count the number of times I have had massive individual matches (and I brawl, as you well know) with 5 plus kills, 700 dmg..... and the team still lost the match. a great example is this one

Brawling Gargoyle, took the lead, killed all 6 mechs in the first 3 minutes of the match......team was up 11-6 when I went down...and they still lost the match.

Why does that impact my ranking? I did what an Assault is supposed to....played wedgebreaker and killed HALF THEIR TEAM.

[Bilbo]

Imperius, on 23 September 2015 - 06:45 AM, said:

It's ok I'm lower tier 3 and I have no idea why or do I care. Speculation says it's total games makes the biggest difference.

It really doesn't. I'm just shy of Tier 3.
Wins / Losses 3,814 / 3,486

[Kyynele]

Titannium, on 23 September 2015 - 03:42 AM, said:

Id like to know, if T1 pilots could write here their average stats, ie KDR + damage and cbills earning (non +30%), to compare.

Here you go.



I pug a bit too much, have a need to level up all the mechs, and like to play underdog mechs like Locusts and Cicadas in tournaments. And sometimes just because they're fun.

[DaFrog]

KBurn85, on 23 September 2015 - 05:35 AM, said:

mate, in this game, the bottom line is you either derive joy from stats, or you derive joy from piloting mechs.

I'm a Tier 3 player who goes for stats. My mechs are timbergod, whale, cheater. They are meta to the hilt. I've even made an excel spreadsheet to count the number of MPLs that goes into the timbergod. I got sick of chasing KDRs and quit sometime ago. It was too much pressure to keep up 3.0 on certain mechs. I'm only back to check my tier.

Look at my CTF. It sucks. I'm probably T5 on that mech.

If you don't chase the meta, but pilot for the joy of piloting many different mechs, then you will never be happy if you care an iota about tiers.

Ultimately, this is a very very competitive game, because there is no objective other than chasing kills previously, and now, chasing tiers, so the higher tiers are filled with nothing but meta.

For me, it feels like work, rather than play, which is why I quit. Ask yourself why do you play.

No. This is not a competitive game. This is PUG. If you wanna play competitive, go sign up in the tournaments.
And b4 you actually think of being a smartass about my comment, know this: my tier is better than yours, and I am not playing competitive one iota.

[Lily from animove]

Bishop Steiner, on 23 September 2015 - 07:25 AM, said:

just don't think the W/L should really be weighted heavily on individual performance.

Can't count the number of times I have had massive individual matches (and I brawl, as you well know) with 5 plus kills, 700 dmg..... and the team still lost the match. a great example is this one

Brawling Gargoyle, took the lead, killed all 6 mechs in the first 3 minutes of the match......team was up 11-6 when I went down...and they still lost the match.

Why does that impact my ranking? I did what an Assault is supposed to....played wedgebreaker and killed HALF THEIR TEAM.

nice example.

But the current PSR ratign will probably NOT make you drop, so you will stay.

While those 98 damage guy wil also "stay" So by PSR it emans you both equal in skill what you delivered in this matchup. Seems not correct or what do you think?

furtherthat 197 damage gyu in the winning team will go up, because he realyl did much better than you?

this is a prime example of a match and how PSR rating we currently have is NOT good.

tbh takign those matches scores into account, I would rate 4 people of the winning team upwards, because they did good. 2 of the losers as well, because they did equally good.

many winners did not even good, they should simply drop in rating, and some of the losers as well (below 200 damage guys) . the rest. The rest did a normal game which neither deserves dropping or raising just because they netted the win on their team. They were playing ok. and so should stay where they were.

Kyynele, on 23 September 2015 - 07:31 AM, said:

Here you go.



I pug a bit too much, have a need to level up all the mechs, and like to play underdog mechs like Locusts and Cicadas in tournaments. And sometimes just because they're fun.

whats your estimated number of games you did in 2015?

Edited by Lily from animove, 23 September 2015 - 07:39 AM.

[Daiben]

Titannium, on 23 September 2015 - 06:24 AM, said:

yep. because it count mine first ever 100-150 trial/beginner matches with 50-100 damage score.
That part , you "old" players, already skipped by starting 2015 stats with mechs I even dont have now.

Seems fair to me.
PGI, way to go. Discriminating new players for being new.

Tiers are not prizes, they are there to try to have you matched against players with a similar skill level. You have nearly 3000 games now and those 100-150 games as a new player only only a minor influence on your overall PSR rating. For those 3000 games you have a 0.92 kdr, a 0.9 WLR and roughly an average damage per game of 200-250. Therefore tier 5 does appear to be the appropriate place for you to be placed in for you to have fair games.

Its not about discriminating against new players, its about preventing them from being completely crushed every game against high tier players. You claim to be doing 600-800 damage every other game but that doesn't appear to be backed up by your stats. Your direwolves average 300 damage while your atlases are around 250. That means for every 800 damage game you probably have 9 games where you are doing 150-300 damage. For a tier 1 or 2 player those numbers would be low even for most mechs let alone a direwolf. Keep in mind your stats have been earned against other tier 4-5 level players; how do you think you would do if you were suddenly put up against tier 1-2 players? Do you think the game would be more enjoyable?

If you don't want to be at tier 5 then you just need to put out good performances more consistently. It could be a skill issue but it could also just as easily be a problem with your builds, your in-game positioning or your general strategies. Find out which area is the issue and work on it and you will probably find yourself in a higher tier soon enough.

[Bulletsponge0]

Buster Highman, on 23 September 2015 - 04:10 AM, said:

Where in the settings can I change my tier to 5?

Sorry buddy, you're Tier1forLyfe

[Tatula]

Played about 10 matches last night. I got either the green up arrow or the red down arrow. Didn't see a yellow equal sign. Not once. Here's an example:

http://i.imgur.com/SCCkGJS.jpg

2 kills, 365 damage, 3 assists, a loss. I would think that should have earned me a yellow instead of a red.

Edited by Aloha, 23 September 2015 - 07:56 AM.

[Kyynele]

Lily from animove, on 23 September 2015 - 07:36 AM, said:

whats your estimated number of games you did in 2015?

Probably something like 500-600 regular queue matches this year.

Also, if you do decently in a lost game, your rating still goes up.

[Revis Volek]

Titannium, on 23 September 2015 - 03:56 AM, said:

Sirmad, i know about this, but that not answered my question.
doing 600-800 damage in assaults, and system said , you are T5. why ?

quul, i saw you twice in my rounds yesterday.
I got 7 green, 3 reds from 10 mathces, my T5 progress bar moves +- 0.5mm. so how many games must i win to get to T4 ,and how many to T1 ? 2000/30000 ?

If you do 600-800 dmg every game and get a kill or two you would not be in tier 5, Im sorry but i simply do not believe you.
Give us some screen shots or some proof, maybe you found an issue but but i doubt it.

I avg 550 dmg a game in Mediums (my most played) and lights and i am a 85% filled bar tier 3. No way you are that consistent and that far down into the underhive.