40 replies to this topic

[Elizander]

Scout Derek, on 17 March 2017 - 06:38 PM, said:

I know the myth that tiers don't matter....

but please, just... reset them, this patch or the next. Some matches are way too lopsided and quite frankly.... it gets pretty horrible.

I mean, I like racking up the kills, but honestly I feel bad for the guy that I backstab because he didn't turn around after the second to third SRM 22 Pack into his back :/

And then being on the recieving end of it... Appogee knows what I'm talking about. He and I were on opposite sides. He kept getting a bad team while I kept getting the overall better one.

I can't help but feel this way whenever my team goes potato and I carry too much:

The only way to maintain 'quality' of higher tiers is to have a limited amount of the active population stay in Tier 1/2. This means better performing players will push potatoes down to tier 2/3.

[Weeny Machine]

Kin3ticX, on 17 March 2017 - 10:46 PM, said:

You need to fix the title of this thread doe


"Hi community I am just your friendly tier 1 neighbor and I think we should reset the tiers so I can have 2 weeks of slaughtering grade AAA potatoes"

Yup, that so totally doesn't happen currently, right?

Elizander, on 18 March 2017 - 04:52 AM, said:

The only way to maintain 'quality' of higher tiers is to have a limited amount of the active population stay in Tier 1/2. This means better performing players will push potatoes down to tier 2/3.

I'd love that.

Edited by Bush Hopper, 18 March 2017 - 05:06 AM.

[MacClearly]

I don't think there is any hope unless they revamped the whole tier system or increased the metrics. As a player with accounts in tier 4 through 1, I assure you that there is a difference and a reset would be pretty painful. Now while I would like a better system or stricter,I am not convinced it would make much of a difference. Quick play is the lowest form of the game and most of the population lives there full time. Even at tier 1 most players are frustratingly unable to even do team basics. Nascar, standing behind someone trading and blocking their retreat, plugging up firing lines and so on are common and normal behaviour.

[Ghogiel]

I find it funny how people QQd about the Elo MM we had when it was zero sum, post match it adjusted skill rating amounts based on who was playing who and on W/L only, as opposed to moving the rating based on flat values according to match score + W/L, took into account mech weight classes (we had 4 Elo ratings, 1 for each mech class).

I'm not sure which was complained about more, PSR or Elo lol.

We need at least 2 ratings as well, group and solo, which was also a big failing of the Elo MM.

Edited by Ghogiel, 18 March 2017 - 08:44 AM.

[Mister Blastman]

Ghogiel, on 18 March 2017 - 08:43 AM, said:

I find it funny how people QQd about the Elo MM we had when it was zero sum, post match it adjusted skill rating amounts based on who was playing who and on W/L only, as opposed to moving the rating based on flat values according to match score + W/L, took into account mech weight classes (we had 4 Elo ratings, 1 for each mech class).

I'm not sure which was complained about more, PSR or Elo lol.

We need at least 2 ratings as well, group and solo, which was also a big failing of the Elo MM.

The problem with both systems is they don't grade on the curve.

ELO was win/loss only, which is a poor measure in a team game where your teammates are a strong influencing factor. It needed to be win/loss + your contribution factor (match score) and then compared with the running average to factor your gain/loss in ELO points based on a weighted scale.

ELO was too simplistic.

The other problem was the matchmaker averaged the ELO scores to build a team, not considering that three goods and nine bads don't balance well against twelve competents, especially in a game that rewards focus fire and cover.

PSR is an experience bar. It is ********. The matchmaker needs to take into account individual ability and that individual ability needs to be constantly held accountable in the statistical database based on the ever evolving player population... thus rating on a curve.

[Ghogiel]

Mister Blastman, on 18 March 2017 - 08:52 AM, said:

ELO was too simplistic.

Nah it was too complex, which is why we got PSR. XD

[Mister Blastman]

Ghogiel, on 18 March 2017 - 08:54 AM, said:

Nah it was too complex, which is why we got PSR. XD

I don't think it was. Knowing what we know about those who do, I'm more inclined to believe it was a clunky mess that only half-worked half the time, and the other half it exploded in disaster.

[Ghogiel]

Mister Blastman, on 18 March 2017 - 08:52 AM, said:

The other problem was the matchmaker averaged the ELO scores to build a team, not considering that three goods and nine bads don't balance well against twelve competents, especially in a game that rewards focus fire and cover.

I believe so too, but it was at least self correcting. You can't be good according to the MM if you keep losing so you'll just end up on the team of competents if you can't carry enough bads. PSR is can't even tell the difference between a bad, competent or a good player.

Edited by Ghogiel, 18 March 2017 - 08:58 AM.

[Bud Crue]

If I've made it to T2, and am still rising; I'm pretty sure a reset won't do any good.
The manner in which the game scores you, advances you and penalizes you all need to be redone. The impact of group play vs solo play. The impact of other's performance. The impact of the mech, its weight and its "tier" also need to be taken into account. Etc.

I think it safe to say that none of that will ever happen; and if that's the case then a reset wouldn't really help anything.

[Trashhead]

I think it would be better over all to not reset the tiers,
but rather distribute the players over the 5 given tears based on their skill in regards to the skill of all otherplayers... like in a top 10.

I.E. if you constantly score low you stay in T5 or T4,
if you score very high you end up in T2 or T1.

As it stands today, every single player would end up in Tier 1 if they just played long enough.
(At least, this is the impression I got, since I am tier 3 now and still raising, despite being mediocre at best.)

[Abisha]

Tier rating say's nothing about how good you are only K/D rate gives a clue.

[Deathlike]

Ghogiel, on 18 March 2017 - 08:43 AM, said:

I find it funny how people QQd about the Elo MM we had when it was zero sum, post match it adjusted skill rating amounts based on who was playing who and on W/L only, as opposed to moving the rating based on flat values according to match score + W/L, took into account mech weight classes (we had 4 Elo ratings, 1 for each mech class).

I'm not sure which was complained about more, PSR or Elo lol.

We need at least 2 ratings as well, group and solo, which was also a big failing of the Elo MM.

Some people simply cannot accept that they are not good. That's the reality for you.

I know there were flaws in ELO, but if we knew what our ELO values were, I bet they were more accurate than what we have now.

It would be nice if people accepted that winning is solely dependent on whether or not you work together, in conjunction with being able to run a decent enough build. However, we can't have that in MWO because... participation trophies and all.

Edited by Deathlike, 18 March 2017 - 01:41 PM.

[Dryderian]

El Bandito, on 17 March 2017 - 09:17 PM, said:

PSR is never gonna be accurate unless its upwards bias is removed. Failing that, a reset is needed. T1 is way too diluted ATM.



But that has nothing to do with the PSR system, which is the main topic of this thread. And end game really should not be accessible to new players easily. I personally would not have allowed CW to be shown on player menu until they played certain amount of matches, or hit certain tier.

It is so frustrating to have a pugger that dies with all 4 mechs even before mid-game, or that pugger who drops with a 90 ton Cyclops as his last mech.

My point is, unless MWO can a keep a big enough player base to let the algorithm do it's work in quick plays or FW, the problem can't be solved and you keep players with the endgame, which is faction warfare.

[stealthraccoon]

I've been in T5 since tiers were introduced - it's never moved far as I can tell.
I think playing lights has put me there; I get 3 kills, 400 damage, and my team loses, I'm lucky to get an '=' out of it.

[DAYLEET]

Scout Derek, on 17 March 2017 - 06:38 PM, said:

I know the myth that tiers don't matter....

but please, just... reset them, this patch or the next. Some matches are way too lopsided and quite frankly.... it gets pretty horrible.

I mean, I like racking up the kills, but honestly I feel bad for the guy that I backstab because he didn't turn around after the second to third SRM 22 Pack into his back :/

And then being on the recieving end of it... Appogee knows what I'm talking about. He and I were on opposite sides. He kept getting a bad team while I kept getting the overall better one.

I can't help but feel this way whenever my team goes potato and I carry too much:

Making everyone T5 wont prevent you from playing with T4-t3. There is not enough people to enforce tiering or 3/3/3/3.

[Mister Blastman]

Abisha, on 18 March 2017 - 09:43 AM, said:

Tier rating say's nothing about how good you are only K/D rate gives a clue.

K/D alone isn't sufficient. It must be used in conjunction with average match score and win/loss ratio, and then compared with your peers.

Edited by Mister Blastman, 18 March 2017 - 02:15 PM.

[Moonlight Grimoire]

Resetting tiers would be a half decent idea, I know when PSR first came out and I got to T4 the game got way better than the hell of T5, now that I am almost to T2 and it has been a long while since the PXH came out, the quality of matches has hit T5 levels constantly. Even if I see T1's on my team or the enemy team (names I know from forums that show the T1 badge) quality still seems horrid.

PSR worked until people got sorted into eternal T5 or slide up into T1. Like once out of T5, it is hard to go down in PSR, like really hard.

Return to ELO, redo match score to weigh things that actually win matches which are components destroyed, KMDD, UAV kills, UAV detection, that stuff instead of kills. Have the stats for pilots under the stats tree count KDR and kills based on KMDD's and not actual kills. Push people to actually play and not splash mechs to get that last hit for the kill. That will mess with people's precious KDR. Have PSR and Tier's still exist but more for showing standard deviations from "average". T5 being -2 deviations, T1 being +2 deviations, roughly. Dump new players in the bottom bracket, we don't expect much of them so they earn their ranking. With an ELO more focused on performance than kills, damage, and wins (reward smarter playing) and rewards consistent players with higher ELO ranking it helps. Cull people who haven't played say 60 matches in the past 30 days (comes out to around 10 hours over 30 days, generous in my opinion) from the ELO ranks to remove inactive players from recalculating at the beginning of the month where each deviation begins and ends.

ELO did give better matches than what we have now. PSR gave better matches for the first 4-6 months it was out. Now PSR has settled and it is bad. Removal of PSR and switching it to a standard deviation rating would allow for removal of the "valves" that PGI has which artificially makes more "buckets" of players. This will help with faster match making, or just go full ham and don't use ELO for match making and just do first come first serve, do it's best to get 3, 3, 3, 3 for both sides within a certain tonnage range and not care about skill, would certainly mix up the playerbase in who they face. PGI also needs to find a way to have 8v8 and 12v12 running at the same time, have MM randomly pick 8v8 or 12v12 with leaning more towards 12v12 servers when you have a high flow rate of players through MM, push more players to 8v8 servers when flow rate is slow. Helps speed up match making as you are dealing with a third of the players required to run a match.

Still, this doesn't address the economy of the game, which is too low for the average player, takes too long for them to get a new mech after the cadet bonus, premium time with the low economy is near worthless at 50%, if it were 100% at the current price we could have a conversation.

[InspectorG]

Bush Hopper, on 18 March 2017 - 12:21 AM, said:

Newsflash: currently really good pilots (no, not me) are already slaughtering the potatoes anyway

And on a good day, the potatoes slaughter each other!

[Weeny Machine]

Abisha, on 18 March 2017 - 09:43 AM, said:

Tier rating say's nothing about how good you are only K/D rate gives a clue.

I'd prefer match score. Some people are like vultures and save their alpha to get the kills despite another one set it up.

InspectorG, on 18 March 2017 - 02:29 PM, said:

And on a good day, the potatoes slaughter each other!

What is that called? "Mashed potatoes" errr I mean "matched potatoes" ?

Edited by Bush Hopper, 18 March 2017 - 04:08 PM.

[Mystere]

It is pointless to change the matchmaker -- or even to have one, for that matter -- unless the population numbers improve.

It will just be a waste of time, effort, money, and other limited resources.