37 replies to this topic

[El Bandito]

Then why is there not a single T1 player with negative WLR?

All of them have more than 50% WLR, and guess what? If you have more than 50% WLR in the highest tier, then you are indeed a skilled player.

[Thorqemada]

Bcs PSR is a Playtime + Success metric and these Players were the first to meet the requirements.

[El Bandito]

Thorqemada, on 27 September 2015 - 08:35 AM, said:

Bcs PSR is a Playtime + Success metric and these Players were the first to meet the requirements.

Yep. Seems some people are forgetting the "Success" part.

[Khobai]

The skill level at tier 1 still varys wildly

5 tiers wasnt enough to truly seperate the playerbase out based on skill level alone

You would probably need at least 7-8 tiers for that

[GRiPSViGiL]

You are a mission to make it mean you're better than everyone else. So sad. As said there is still a wide range concerning skills at that tier.

Edited by GRiPSViGiL, 27 September 2015 - 09:15 AM.

[El Bandito]

GRiPSViGiL, on 27 September 2015 - 09:14 AM, said:

You are a mission to make it mean you're better than everyone else. So sad. As said there is still a wide range concerning skills at that tier.

If I was in a mission to make myself appear better than everyone else, I wouldn't be flaunting my T2 badge.

Edited by El Bandito, 27 September 2015 - 09:18 AM.

[Chuck Jager]

Khobai, on 27 September 2015 - 09:12 AM, said:

The skill level at tier 1 still varys wildly

5 tiers wasnt enough to truly seperate the playerbase out based on skill level alone

You would probably need at least 7-8 tiers for that

Who say matchmaker goes directly off of the tier. There could be a 5000pt spread amongst all tiers and MM chooses a spread of 10000 to 2250 with certain rules and algorithms.

[El Bandito]

Chuck YeaGurr, on 27 September 2015 - 09:18 AM, said:

Who say matchmaker goes directly off of the tier. There could be a 5000pt spread amongst all tiers and MM chooses a spread of 10000 to 2250 with certain rules and algorithms.

T1-T3 drop together so I would not be surprised if there is a large skill difference between team members. And even within T1, certain players are simply God among men.

Edited by El Bandito, 27 September 2015 - 09:22 AM.

[GRiPSViGiL]

El Bandito, on 27 September 2015 - 09:17 AM, said:

If I was in a mission to make myself appear better than everyone else, I wouldn't be flaunting my T2 badge.

Just conditioning everyone for when you get that tier 1 badge.

[EgoSlayer]

El Bandito, on 27 September 2015 - 09:04 AM, said:

Yep. Seems some people are forgetting the "Success" part.

The problem is "success" only means a 51% win rate.

Edited by EgoSlayer, 27 September 2015 - 09:27 AM.

[El Bandito]

GRiPSViGiL, on 27 September 2015 - 09:25 AM, said:

Just conditioning everyone for when you get that tier 1 badge.

And I will be getting it through my own skill, not lack of it.

EgoSlayer, on 27 September 2015 - 09:27 AM, said:

The problem is "success" only means a 51% win rate.

Teams in popular games advance to, and win World Championships with 51% win rate.

Edited by El Bandito, 27 September 2015 - 09:31 AM.

[EgoSlayer]

El Bandito, on 27 September 2015 - 09:28 AM, said:

Teams in popular games advance to, and win World Championships with 51% win rate.

Sure, but you know what you call a medial student with a C - Average?
Doctor.

Doesn't mean they are the best.


Yet a 51% win rate in MWO gets you to tier 1, the best.

[Fleeb the Mad]

According to the graph that was posted in the PSR thread, it's impossible to get to T1 by playing lots with no other metrics. If you win but perform poorly your PSR does not change. Players can't be carried up to the upper tiers in solo queue.



While I'd personally argue that all PSR metrics are skewed for players who frequent the group queue, particularly those who play on large teams as opposed to puggling it, one needs to have regular 'high match score' performance in order for PSR to rise. I say that with the expectation that a player will typically see higher match scores on wins than on losses if their performance is consistent, because winning teams offer more scoring opportunities. But since a 'medium' score loss results in the same drop as a 'high' score win, that tends to imply for a 50/50 win/loss rate that their PSR would remain at even keel if they didn't have occasional awesome matches and few or no flubs.

So...I actually have some difficulty coming to the conclusion that a player, without continuous improvement, would rise out of any tier they're in. I suppose to know for sure we'd have to actually see where the match score bars are set and how much a 'large' rise/drop is in relation to a 'small' one.

Edit: I'd also mention that it's a bit fallacious to use Win/Loss rate as a success metric. If all the top tier players have greater than 50% win rates...how exactly does that work out when they play each other and are isolated from the bottom tiers? The only way that sort of thing works is if their stats are skewed by CW or Group queue matches.

Edited by Fleeb the Mad, 27 September 2015 - 09:57 AM.

[Mechwarrior Buddah]

El Bandito, on 27 September 2015 - 08:12 AM, said:

Then why is there not a single T1 player with negative WLR?

All of them have more than 50% WLR, and guess what? If you have more than 50% WLR in the highest tier, then you are indeed a skilled player.

According to Russ, everyone will eventually be t1 but also according to Russ, everyone's skill will eventually plateau. So... T1 is where the bads go to get stomped

Brilliant plan

Fleeb the Mad, on 27 September 2015 - 09:46 AM, said:

According to the graph that was posted in the PSR thread, it's impossible to get to T1 by playing lots with no other metrics. If you win but perform poorly your PSR does not change. Players can't be carried up to the upper tiers in solo queue.



While I'd personally argue that all PSR metrics are skewed for players who frequent the group queue, particularly those who play on large teams as opposed to puggling it, one needs to have regular 'high match score' performance in order for PSR to rise. I say that with the expectation that a player will typically see higher match scores on wins than on losses if their performance is consistent, because winning teams offer more scoring opportunities. But since a 'medium' score loss results in the same drop as a 'high' score win, that tends to imply for a 50/50 win/loss rate that their PSR would remain at even keel if they didn't have occasional awesome matches and few or no flubs.

So...I actually have some difficulty coming to the conclusion that a player, without continuous improvement, would rise out of any tier they're in. I suppose to know for sure we'd have to actually see where the match score bars are set and how much a 'large' rise/drop is in relation to a 'small' one.

Soi we yet KNOW what "low" "medium" or "high" are in terms of numbers?

[Ultimax]

EgoSlayer, on 27 September 2015 - 09:27 AM, said:

The problem is "success" only means a 51% win rate.

Tarogato's thread where he is collecting stats says otherwise.


http://mwomercs.com/...ts-for-science/

Tarogato said:

73 players collected total

Tier 1:
12 players
6500 matches average
2.34 KDR average
1.44 WLR average

Tier 2:
21 players
3412
1.77 KDR average
1.33 WLR average

Tier 3:
25 players
2642 matches average
1.39 KDR average
1.17 WLR average

Tier 4:
9 players
2064 matches average
1.20 KDR average
1.05 WLR average

Tier 5:
6 players
3980 matches average
0.59 KDR average
0.99 WLR average

So...

T1: 72% wins average
T2: 67% wins average

Edited by Ultimatum X, 27 September 2015 - 09:56 AM.

[EgoSlayer]

Ultimatum X, on 27 September 2015 - 09:56 AM, said:

Tarogato's thread where he is collecting stats says otherwise.


http://mwomercs.com/...ts-for-science/






So...

T1: 72% wins average
T2: 67% wins average

Actually that doesn't disprove or counter it. First it's an average, second it takes somewhere around 5-7 thousand matches (assuming it's a linear progression from tier 2 to tier 1) to get to Tier 1 with a 51% average. Matches that have to have been played since January.
The average number of matches played, KDR, and W/L ratio in the above stats are total history since the stat reset in April 2014, so the number of matches in that list doesn't indicate only PSR matches and likely half or more of the matches are from elo time.

[Koniks]

The reason Elo was better is because it weighted wins against strength of opponent.

PSR treats winning and match score the same regardless of who you play. With good matchmaking, that's not a problem because in theory you should be matched against people of equal skill or likelihood to win.

This was and is problematic with how our group queue matchmaker averages your group's Elo and PSR for matchmaking. But it's still a problem with solo queue with how the release valves worked and still work.

[Thorqemada]

The graph shows 6 cases where your PSR climbs (4) or stays equal (2) and 2 cases where your PSR decreases - the Chance to climb is bigger than to fall.

My Tier 3 PSR did rise despite doing pretty moderate (50/50 W/L + many underwhelming lost matches) in my Awesome - it shrinks and grows but in the end it grows slowly bigger...

Edited by Thorqemada, 27 September 2015 - 11:26 AM.

[Fleeb the Mad]

Mechwarrior Buddah, on 27 September 2015 - 09:52 AM, said:

Soi we yet KNOW what "low" "medium" or "high" are in terms of numbers?

I'm not sure what this is asking, but as far as I can tell, no. We don't actually know what the numbers or the score thresholds are.

Ultimatum X, on 27 September 2015 - 09:56 AM, said:

Tarogato's thread where he is collecting stats says otherwise.


http://mwomercs.com/...ts-for-science/


So...

T1: 72% wins average
T2: 67% wins average

I'm inclined to dismiss that outright based on the tiny sample size, and also the fact that the statistics are going to have a massive bias because they include both group queue and CW numbers as well as historical numbers from previous matchmaking systems. I don't believe anyone can actually maintain a 70% win rate in the solo queue, particularly when the matchmaker will put these people against each other.

It cannot be a trend in public queue gameplay to have a win rate greatly deviating from 50% over hundreds of games, in a system that will continually make your matches harder as you win more. The only way that sort of thing happens over a large number of games is if they skew the results by playing on stacked teams in the CW or group queue.

That in itself says nothing for the skill of the individual player in a meaningful context.

[Kira Onime]

Mechwarrior Buddah, on 27 September 2015 - 09:52 AM, said:

According to Russ, everyone will eventually be t1 but also according to Russ, everyone's skill will eventually plateau. So... T1 is where the bads go to get stomped

Pretty sure that was Paul saying that.