23 replies to this topic

[Pardo Kerensky]

I seriously cannot understand why we cannot have a player skill rating that actually takes into consideration how high your actual skill in the game is instead of how much you have played and how many times you got carried in your trash built trashmechs
why cant we just have an ELO based PSR based on average matchscore
[Redacted], would it be so hard to implement?
explain to me kindly what Deep Rooted Cryengine Errors™ are stopping you from implementing a better system
I have recently come back to the game after quitting shortly after the first clan release and the games have become atrociously unbearable
90% of the time I score the highest matchscore and highest dmg in my team and my team still gets stomped because 8 out of 12 players cannot even deal more than 100 dmg while being in 70 to 100 ton mech weight range
more and more long time good players / pro players are fleeing from your game in favour of better vehicular based games
a lot of good players / pro players I was friend with or I used to see regularly are now gone
are you seriously waiting for every good player to leave?
is that your objective?
or most likely you dont care about the good skilled players, about the game being well crafted and fun for everyone, the only thing you ever cared about since the very beginning are the atrociously bad whale players that you can milk on regular basis at every MechPack™ release, so you do everything possible to cater to them, first of all by implementing a so called 'Player Skill Rating' that doesnt take into account the actual player skill in the slightest
because god forbid if we actually separate the trashes from the skilled players, because that would hurt the feelings of the average whale boomers to end up being forever segregated to play with terrible players of their own skill level
so they just get carried in their awful builds while not being able to aim, pilot nor position at all and slowly end up in tier 2 and 1 simply by playing like trash 12 hours a day everyday and getting carried
and you call this system a player SKILL rating?
I dont even know what to say anymore
coming back was a mistake
[Redacted]

Edited by draiocht, 07 June 2018 - 08:37 AM.
staff abuse

[Davegt27]

" how many times you got carried in your trash built trashmechs"

lol

[Villainy]

I understand the frustration mate, but whew lad. You feel better?

[Savage Wolf]

Because you cannot in any way reliably track individual performance no matter how your algoritm. It is virtually impossible and no one has even gotten close to it. Every attempt ever made turns into what we also have today: people statpadding instead og playing to win and that ruins teamwork.

The only stat you can really use reliably is win / loss, but in a team game that does not track the individual performance, only the team. But those are the terms and limitations a game is subject to.

But don't take my word for it, just ask Extra Credits

[Pardo Kerensky]

Savage Wolf, on 07 June 2018 - 12:25 AM, said:

Because you cannot in any way reliably track individual performance no matter how your algoritm.

yes you can
by linking the ELO/PSR whatever you wanna call it directly to your matchscore
>play well and get higher than team average matchscore and you climb up based on how much higher than the average team score your score actually is
>play like trash and do awful and not contribute anything to your team and you drop down no matter if its loss or if you get carried and win anyway

this is how it should be

contribute to the team with matchscore and you will climb either if its a loss or a win

I seriously do not know what is so goddamn hard about implementing a system like this
[Redacted]

Edited by draiocht, 07 June 2018 - 07:39 AM.
staff abuse

[Savage Wolf]

Pardo Kerensky, on 07 June 2018 - 12:45 AM, said:

yes you can
by linking the ELO/PSR whatever you wanna call it directly to your matchscore
>play well and get higher than team average matchscore and you climb up based on how much higher than the average team score your score actually is
>play like trash and do awful and not contribute anything to your team and you drop down no matter if its loss or if you get carried and win anyway

this is how it should be

contribute to the team with matchscore and you will climb either if its a loss or a win

I seriously do not know what is so goddamn hard about implementing a system like this
[Redacted]

Then explain to me how to accomplish this and go down in history as the person who achieved what no other game developer so far has been able to pull of.

What you want to do can only be done if we solve this fundamental problem first or put in MWO terms: Actually invent an algorithm that can produce an accurate matchscore and somehow prevent people from statpadding. Otherwise, you solve nothing.

[Redacted]

Edited by draiocht, 07 June 2018 - 07:43 AM.
staff abuse, Quote Clean-Up, replies removed

[Pardo Kerensky]

Villainy, on 07 June 2018 - 12:15 AM, said:

I understand the frustration mate, but whew lad. You feel better?

no I dont feel better at all
yesterday I played all afternoon and all evening until late night EU timezone and 90% of my games were like this
do highest damage and highest matchscore not just in my team but in the entire goddamn match and still lose
and the day before has been the same and the day before exactly the same all the way since I started playing again
this was the scoreboard of the last game I had last night after which I just stopped out of sheer frustration and depression

[Redacted]

Edited by draiocht, 07 June 2018 - 08:39 AM.
name & shame policy

[Anjian]

Maybe matchmaker can't compare [Redacted] and matchmaker them properly.

I do come from a game where skill based or at least win potential based matchmaking is in effect, and guess what, people are "tanking" their stats deliberately so they face easier competition and seal club.

Edited by draiocht, 07 June 2018 - 07:49 AM.
inappropriate reference

[Savage Wolf]

Anjian, on 07 June 2018 - 01:19 AM, said:

Maybe matchmaker can't compare [Redacted] and matchmaker them properly.

I do come from a game where skill based or at least win potential based matchmaking is in effect, and guess what, people are "tanking" their stats deliberately so they face easier competition and seal club.

Ah, a new low for online play. Not only can a personal score lead to statpadding, it can also lead to stattanking. Somehow I'm not surprised. But both are examples of a personal score system making people play worse and not better. So yeah, PSR has to go.

Edited by draiocht, 07 June 2018 - 07:49 AM.
Quote Clean-Up

[Peter2k]

Savage Wolf, on 07 June 2018 - 12:58 AM, said:

Then explain to me how to accomplish this and go down in history as the person who achieved what no other game developer so far has been able to pull of.

What you want to do can only be done if we solve this fundamental problem first or put in MWO terms: Actually invent an algorithm that can produce an accurate matchscore and somehow prevent people from statpadding. Otherwise, you solve nothing.

[Redacted]

Why worry about perfection when all you have to achieve is better than now?
MM is basically just trying to keep vets away from new players (according to Paul), and from daily anectodal stories on the forum it doesn't even do that.

I wish MM would at least try to balance out a few key things across the team, like both teams get an ecm mech and some LRM boats; similar tonnage and so on.
Instead of putting all LRM mechs and all ecm mechs with tag and NARC on the same team on a map that favors LRM.
Personally I think a few tweaks would go a long way.
I'm not sure my view on the matter is skewed, but when we had ELO for weight classes it seemed better.

Also what does stat padding give you in MWO terms?
Playing every game against P R O T O N and the like, really good players?
Think that's fun for someone who "cheats" it's way to the top?
I played a handful of games where the MM decided putting P R O T O N, Wispy and some other really good players on the enemy team was a great idea instead of shuffling them between both sides.
Game was over faster than the loading time.
3 times in a row.
I'll never forget.

If anything you need to concern yourself with ELO tanking so elite players can't just club seals.

Edited by draiocht, 07 June 2018 - 07:45 AM.
Quote Clean-Up

[Savage Wolf]

Peter2k, on 07 June 2018 - 01:29 AM, said:

Also what does stat padding give you in MWO terms?
Playing every game against P R O T O N and the like, really good players?
Think that's fun for someone who "cheats" it's way to the top?
I played a handful of games where the MM decided putting P R O T O N, Wispy and some other really good players on the enemy team was a great idea instead of shuffling them between both sides.
Game was over faster than the loading time.
3 times in a row.
I'll never forget.

If anything you need to concern yourself with ELO tanking so elite players can't just club seals.

Yeah, ELO was better. It wasn't fantastic, but at least it was much more accurate than PSR.

Since PSR is also tied to rewards, cheating yourself to higher tiers also grants you more c-bills. And if you learn how to let yourself get carried by the good players, you can continue in Tier 1, especially since the game encourages people to do it, Tier 1 is actually full of statpadders, so you don't even end up against that many good players.

And stat tanking is basically you learning how to win with the least gain in PSR meaning you win, but is able to drop regardless. But the upwards bias of PSR might prevent this. Only good thing about it though.

[No One Lives Forever]

PSR did work better than its working now. They "loosened" the program cause there's just not enough players and "searching" times were too great.
PGI needs to do more for the game, better marketing, we need more players! They need to drop that Minimally Viable Product BS and take more pride in their work, in their game!

Edited by No One Lives Forever, 07 June 2018 - 01:51 AM.

[El Bandito]

Pardo Kerensky, on 06 June 2018 - 11:32 PM, said:

I seriously cannot understand why we cannot have a player skill rating that actually takes into consideration how high your actual skill in the game is instead of how much you have played and how many times you got carried in your trash built trashmechs
why cant we just have an ELO based PSR based on average matchscore

Cause that's just as bad as PSR, since it shafts Light and Medium pilots.

[Ssamout]

El Bandito, on 07 June 2018 - 02:44 AM, said:

Cause that's just as bad as PSR, since it shafts Light and Medium pilots.

It doesnt shaft light and medium pilots so hard than several assault pilots doing 50 dmg on a regular basis shafts your whole team and will to play this game.

Why not just put in a small multiplier rather than dump the idea?

Edited by Ssamout, 07 June 2018 - 03:49 AM.

[LordNothing]

* thinking potato's shooter

Edited by LordNothing, 07 June 2018 - 03:50 AM.

[Scyther]

OP, your frustrated rant is pretty much in line with the same rants you will see on World of Tanks, Overwatch, LoL, CoD etc etc.

Some of those games have skill based matchmaking. Some of those games' devs swear up and down that testing has shown that 'true' skill based MM is the last thing players want. 'Standout' players no longer stand out and just feel like an average shmoe in the match. Or worse, like they are on the low end of the skill pool for that match.

TBH, for a game that balances weight/class as well as PSR, MWO doesn't do "horribly". Far too many stomps, but every game has those... MWO just has more. The stomps have been getting more frequent over time (IMO - I don't have actual data to show this). I believe this is due to making matches from an ever-shrinking player queue that is being split more ways. Solaris may be pulling some of the 'match carriers' off to fights where they only need to carry themselves.

MWO has additional MM problems aside from QP, in that Solaris divisions, FP total lack of MM, QP group balancing, weight/tech/skill balancing... all present different challenges.

People talk about 'balance the tech' but Clan tech is designed by the IP as inherently superior. So that's another balancing issue. MWO actually needs a sort of 'asynchronous' balancing/MM technique that takes weight/tech/Mechlab build/pilot skill and availability into account to make a more 'close to even, or at least interesting' match composition.

That sounds a bit complicated, but I've posted the bare-bones design of similar algorithms a couple times in MM threads. Interestingly, those posts get zero likes, zero comments, no criticisms... heck, nobody even makes a reference to them or to actual MM algorithm design. Perhaps PGI aren't the only ones who have trouble with the details of MM coding concepts.

Edited by MadBadger, 07 June 2018 - 04:02 AM.

[RickySpanish]

With respect to the OP, this thinking man's shooter is more like chess with a pidgeon.

Edited by RickySpanish, 07 June 2018 - 04:05 AM.

[TWIAFU]

Pardo Kerensky, on 07 June 2018 - 01:02 AM, said:

no I dont feel better at all
yesterday I played all afternoon and all evening until late night EU timezone and 90% of my games were like this
do highest damage and highest matchscore not just in my team but in the entire goddamn match and still lose
and the day before has been the same and the day before exactly the same all the way since I started playing again
this was the scoreboard of the last game I had last night after which I just stopped out of sheer frustration and depression

Real Mechwarriors do that in Urbies and don't need to use 11 others as armor shields.

RickySpanish, on 07 June 2018 - 04:04 AM, said:

With respect to the OP, this thinking man's shooter is more like chess with a pidgeon.

To quote Tom Hanks;

There is no thinking in QP.

[LowSubmarino]

I think it would be possible to find a skill rating which is much more accurate. Wouldnt be perfect but much more reliable.

If you look at a given mechs alpha and dps potential then you can see quite clearly what he or she accomplished within a match. More emphasis should be put on dmg/componente/kills and assists while some of your team mates are still alive.

If you have an energy Marauder IIC with lets say an alpha of 60 - 70 or even more and okayish dps and that player does 100 - 150 dmg on average after 20 matches then that player simply cannot trade yet. Cannot trade at all. Barely gets out 2 - 3 alphas of that. Melts without ever accomplishing much. How much dmg is concentrated on single components is another good indicator. Somebody diahing out 600 dmg (without strikes) and 50 - 75 % of that dmg lands on e.g. side torsos or even legs and is dished out not over hours after team has already died....then Id say thats good.

Somebody with 20 - 30 dmg/brawling alpha (e.g. ac 20 centurion) that is in close proximity to team, absorbs lots of dmg while landing like 30+ alphas is obviously trading pretty well.

Dont know how exactly the stats are determined but somebody landing max 1 - 3 alphas over a number of matches....is cannonfodder.

I dont think any of that is relevant though as we simply dont have enough players to make use of a better rating system.

[Asym]

If you were seriously going to take on "effectiveness", you'd have to determine what to measure... A light pilot and an assault pilot both my have skill, but they aren't the same.

Rewards and points need to be identified and qualified "by role" so that all classes are proportionally rewarded at the most basic levels and then, significantly rewarded is the pilot "exceeds" role expectations.... Each mech you own and have spec'd out should have a combat effectiveness metric. You all complain about crappy builds ! So, reward effective builds and penalize crappy one. Crappy ones produce far less $$$ and exp......

Your effectiveness and the mech effectiveness combine in a combat potential number. Those numbers, not tiers are how an ELO based MM chooses teams AFTER the MM orders the match by type.....first. 3 assualts, 3 heavies, 3 mediums and 3 lights OR, any combination of weights that "balance" the combat potentials and mechs.... That way, combat potential itself determines the balance....

Just a thought.