111 replies to this topic

[C E Dwyer]

Khobai, on 03 December 2017 - 01:14 PM, said:

PSR should not reset

that defeats the whole purpose of PSR which is to keep new players and experienced player seperate

It's a given that a reset is a very bad thing.

P.G.I did a good thing to late in that they put a hard line on no T3 getting matched with a T1 if they'd done it 18 months earlier it would have been even better.

Now they need to adjust the scoring and get rid of the upward bias, wait three maybe four months and those that shouldn't be in their current tier would drop back, but there would be an increase in wait times for T1 people.

This way will not see much change quickly but I imagine that given six months people would be finding games generally closer, and it's really the only practical option left open without a complete change of system.

Of course only P.G.I know how well or how badly each Tier is populated, so we can only speculate

[C E Dwyer]

tker 669, on 03 December 2017 - 03:07 PM, said:

That would be true if matches were not frequently mixed with up to two tiers difference, sometimes more.

The tier gate failure was all tightened up.

Until about this summer your correct it was badly broken and the gates were letting T4's into matches with T1's

Now only T1's and T2's get placed together.

[The Lighthouse]

Based on the current population, separating players based on skill level seems no longer possible. Would be even more impossible when HBS game comes out and eventually MW5.

Here are my assumptions:

1) Despite popular belief, the number of players T1 is probably far smaller than any other tiers. Why? Because this game has such horrid player retention rate that we are keep losing players (just see how many "retired" players from Jarl's list). From personal experience I always have much easier time getting into games as smurf T3 account my main account. Only explanation is that there are far more players around T3 than T1.

2) So I still think the game's population distribution across tiers is pretty much similar to other games, MAYBE a bit swift to T1 side but not much.

3) So the overall tier system actually works... to a degree.


And here are my thoughts.

1) As someone who was involved in Heroes of Newerth for a while, the fact is ELO really does not work in team-based game. The most evident fact is in order for HoN's PSR system to work (which was pure ELO), S2 Games periodically had to "compress" the ratings... Just look at Dota 2 and tell me PSR works on team-based games. It simply doesn't. End of the discussion.

2) Just like real-life sports, game the game needs to use TSR rating, which is based on player's performance rather than team's W/L ratio, and MWO already has all the data needs : Match Scores.

3) The problem is that MWO does not use Match Scores to balance the game.... and I am not even remotely sure the match maker actually uses tiers either. From Taro's long ago post about the quality of the game, even tiers on team are not eventually distributed.

4) I think the first thing should be done is make sure the match maker evenly distributes tiers into teams first. If that does not work well, then use match score instead.

[UnofficialOperator]

Cathy, on 03 December 2017 - 05:00 PM, said:

The tier gate failure was all tightened up.

Until about this summer your correct it was badly broken and the gates were letting T4's into matches with T1's

Now only T1's and T2's get placed together.

Can't be right... I'm seeing really low wlr pilots in my matches...

[The Lighthouse]

Cathy, on 03 December 2017 - 05:00 PM, said:

The tier gate failure was all tightened up.

Until about this summer your correct it was badly broken and the gates were letting T4's into matches with T1's

Now only T1's and T2's get placed together.

I think they reversed that. There are reports Tier 4 Unit mates getting matched with/against Tier 1 folks. I think with such diminished population PGI was forced to open the floodgate again.

Then again, we should be getting hint that Russ suddenly wants for 8v8, that really should tell you guys the truth.

Edited by The Lighthouse, 03 December 2017 - 05:27 PM.

[Shadowomega1]

UnofficialOperator, on 03 December 2017 - 05:17 PM, said:

Can't be right... I'm seeing really low wlr pilots in my matches...

So of us have had their wlr pulled down heavily by players, that refuse to operate as a team. As well as pilot so poorly them being in the match is a hinderance.

[Brain Cancer]

UnofficialOperator, on 03 December 2017 - 05:17 PM, said:

Can't be right... I'm seeing really low wlr pilots in my matches...

Proof that no matter how poor the player, enough time played floats you to the top unless you fundamentally fail at MWO. Or just don't play much. Perhaps both.

[Unnatural Growth]

Cathy, on 03 December 2017 - 05:00 PM, said:

The tier gate failure was all tightened up.

Until about this summer your correct it was badly broken and the gates were letting T4's into matches with T1's

Now only T1's and T2's get placed together.

Where is your source for this? Please quote it. As far as I know, this was never fixed. Ever. I know for a fact that currently there are T3's and T4's dropping in with T1's.

[James Argent]

Cathy, on 03 December 2017 - 04:46 PM, said:

You can very much have negative win/loss and K/D and still progress providing the following happens.

Capper is talking about the mathematical impossibility of ANY game stat ratio being negative. You can't have negative wins. You can't have negative losses. Neither number can be negative, so the ratio is mathematically always positive (or zero, if you've never won a game). An 'undesirable' W/L ratio is less than 1.0, but any number between zero and one is still positive.

[Xmith]

UnofficialOperator, on 03 December 2017 - 03:42 PM, said:

Stats are publicly available. You can stay in T1 as punching bag.

A player's base stats are not publicly available. Using the leaderboard to judge a player's ability is a waste of time.

[Xavori]

The Lighthouse, on 03 December 2017 - 05:03 PM, said:

3) The problem is that MWO does not use Match Scores to balance the game.... and I am not even remotely sure the match maker actually uses tiers either. From Taro's long ago post about the quality of the game, even tiers on team are not eventually distributed.

Match score is insanely bad. It's so heavily slanted towards damage done that it's a pointless metric. And when you move to objective gameplay, it gets worse. A player who dedicates themselves to the non-combat objectives might very well win the match for his team, but in doing so, completely screw themselves over in terms of damage and kills which means poor score, poor c-bills, poor faction rep if playing faction...

[UnofficialOperator]

Xmith, on 03 December 2017 - 09:32 PM, said:

A player's base stats are not publicly available. Using the leaderboard to judge a player's ability is a waste of time.

Ok so what's your base stats? I'm curious to know where you are going with this?

Edited by UnofficialOperator, 03 December 2017 - 11:50 PM.

[UnofficialOperator]

Xavori, on 03 December 2017 - 09:53 PM, said:

Match score is insanely bad. It's so heavily slanted towards damage done that it's a pointless metric. And when you move to objective gameplay, it gets worse. A player who dedicates themselves to the non-combat objectives might very well win the match for his team, but in doing so, completely screw themselves over in terms of damage and kills which means poor score, poor c-bills, poor faction rep if playing faction...

Lol a pointless metric from a game mainly decided by damage and kills in most modes and even in conquest.

Hmm ok

/s

[Vellron2005]

Since, PSR tiers are not an approximation of actual skill, as much as effort and hours put in, who is allowed in "tier 1" and who's not makes no sense..

If we only looked at Win/Loss ratio, Kill/Death ratio and such epeen stats, only the top 20 tryhards would be allowed in.

People keep thinking that "Tier 1" means you're a god-like, super efficient one-man-army of a player.

WRONG.

Tier 1 means you're a veteran who's been around in this game. Means you played a lot. Means you know how the game works. Does not mean you are good with Gauss, or great with LRMs, or a great drop commander.

Only that you've been in the game a while, and have settled into your place in the game long enough to win some matches and not be a complete newbie.

Even T1's use LRMs, lose matches, get out-flanked, or caught in a bad position, or simply derp around and play with builds.

The quality of play does not go up as you go up in tiers.

The only thing there's more of in T1 is ego.

Edited by Vellron2005, 04 December 2017 - 01:50 AM.

[Vxheous]

Vellron2005, on 04 December 2017 - 01:48 AM, said:

Since, PSR tiers are not an approximation of actual skill, as much as effort and hours put in, who is allowed in "tier 1" and who's not makes no sense..

If we only looked at Win/Loss ratio, Kill/Death ratio and such epeen stats, only the top 20 tryhards would be allowed in.

People keep thinking that "Tier 1" means you're a god-like, super efficient one-man-army of a player.

WRONG.

Tier 1 means you're a veteran who's been around in this game. Means you played a lot. Means you know how the game works. Does not mean you are good with Gauss, or great with LRMs, or a great drop commander.

Only that you've been in the game a while, and have settled into your place in the game long enough to win some matches and not be a complete newbie.

Even T1's use LRMs, lose matches, get out-flanked, or caught in a bad position, or simply derp around and play with builds.

The quality of play does not go up as you go up in tiers.

The only thing there's more of in T1 is ego.

Well, there's XP bar Tier 1's, and then there's the truly Tier 1s that are the "super efficient one-man-army of a player". Unfortunately, the matchmaker can't tell the difference, nor are there enough "super efficient one-man-army of a player" Tier 1 players around at any given time to set off 12vs12 (or even 8vs8 matches) with just them. Just like when matchmaker was ELO based, and the top ELO players could sit around for 45 mins and not match into a game.

Edited by Vxheous, 04 December 2017 - 01:53 AM.

[Mightyrant]

and then there's the truly Tier 1s that are the "super efficient one-man-army of a player".

You're a "super efficient one-man-army of a player!"--@Vx


Take that.

Edited by Mightyrant Kell, 04 December 2017 - 09:21 AM.

[Xmith]

UnofficialOperator, on 03 December 2017 - 11:50 PM, said:

Ok so what's your base stats? I'm curious to know where you are going with this?

W/L 1.04
K/DR 1.53

This stat stays pretty much constant. The leaderboard stat does not.

Edited by Xmith, 04 December 2017 - 09:50 AM.

[Xavori]

UnofficialOperator, on 03 December 2017 - 11:58 PM, said:

Lol a pointless metric from a game mainly decided by damage and kills in most modes and even in conquest.

Hmm ok

/s

Two screenshots for ya then:
https://steamuserima...D79B01F77814A5/
https://steamuserima...57CC5E7E9232F7/

The things that should jump out at ya:
1. I did 1143 damage in an Archer Tempest in less than 5 minutes. This means I'm the mostest awesomest pilot to ever step in a mech. Go me! It's my birthday! It's my birthday!
2. We lost in less than 5 minutes. Maybe it's not my birthday after all.

Hell, I did twice the damage of anyone else on my team. I was totally rockin' it. And our VIP walked into the enemy and got blown to bits and that was that.

Also, my Tempest is 2xLRM20, 7 tons of LRM's, 2xSPL, ECM, TCMkIII, and a BAP. The entire reason it did so much damage while only getting 2 kills is that it's a damage farming mech. It gets locks fast, I vomit lurmies all over my target. Lots of pieces fall off the target until it finally keels over dead. Repeat until the enemy finally pays attention to me, which they never did in that match because they were too busy turning our VIP into slag.

There is nothing good about my Tempest build except that it farms damage and destroyed components (day 5 of Advent event here I come!). It's terribly inefficient. But hey, I got a bunch of damage which gave me a great match score, and that's all that matters, right?

[Brain Cancer]

Yeah. There's people who are veterans (heck, I'm one, I get those 3-4 kill games regularly enough), and then there really are the one-man killing machines who consider it just another day if they wipe out an enemy lance by their lonesome.

I consider a point of pride any time I make the top 200 on any positive leaderboard stat via solo play alone.



It's been a good season so far.

[Asym]

BC, 20 games? 20 games!
I have 76 games and spent a vast majority of that in a new/mastering Kit Fox and sit at .76 w/l 174 avg.... So BC, what does this infer about us? I played three times what you played in inferior equipment or mechs of different classes?

All of the data is meaningless because there is not a stable "expected" value or baseline metric that means anything....

I did have a few Kit Fox matches with 2+ kills and >500 damage! But, that depended on the assaults that I was protecting. They created the conditions for 2 + kills and 500+ damage because they have the combat power and mass to change the battle....not me.

Hope to see you out there again!