66 replies to this topic

[KHETTI]

I rarely play MWO these days , so when i do i'm rusty as f**k, my awareness has diminished as has my gunnery skills.
Its pretty sad when i drop in an Adder Prime and outscore the heavies on my team combined, usually results in me logging out again for another couple of weeks.

[Davers]

KHETTI, on 05 September 2016 - 10:24 AM, said:

I rarely play MWO these days , so when i do i'm rusty as f**k, my awareness has diminished as has my gunnery skills.
Its pretty sad when i drop in an Adder Prime and outscore the heavies on my team combined, usually results in me logging out again for another couple of weeks.

Every weekend is some crummy event that seems to reward horrible play. I think you would get better matches in the middle of the week.

[MW222]

Bob Jenkins, on 05 September 2016 - 08:51 AM, said:

It's not really any better, for me personally, in fact group que is even worse than solo que. Matchmaker is the single most frustrating thing in this game. I would play CW 100% of the time but that game mode is dead, its population is barely existent.

Wish I could say this is a rare thing, but here is an example of the other team having EIGHT assaults to our SINGLE assault. Five of those assaults were Kodiaks. Can't make this stuff up.

https://www.dropbox.....06.13.jpg?dl=0

One in ten ratio of "have a chance of winning" to "no chance" sounds pretty accurate.

Well first say "Thanks PGI!" and if you are wondering it's called load balancing and the sorting algorithm took to long so engage the subroutine call "Dump to Que", rinse and repeat.

[Cabusha]

Match maker has always been crap in this game, made even worse since 12v12 because it just snowballs that much faster. I don't expect to have wins every match, but I do expect to see stuff that's reasonably close. But MM being MM, it will eventually start farming you to the other team. Generally after 2 or 3 straight matches of seeing LORD, 228, MS, etc on the other team, I pull the plug and stop playing. The match maker bias is strong.

[DAYLEET]

What do you want: wait 10-15 minutes to drop or drop fast. Chose wisely.

[C E Dwyer]

oneda, on 19 August 2016 - 04:06 AM, said:

Pls.

I just want to understand how this is possible.

My mind cannot comprehend such deeds I saw in the last 10 games.

Its beyond weird or just bad game play. Its not even human.

I honestly believe their are nonhumand entities in mwo and prolly all over the world by now.

I think that is the explanation.

PSR means people with limited skills will always sooner or later end up in T1.

you've played the game long enough to know this game has some major flaws in it, due to whom the Lead designer is, and nothing will change, by posting what most people have realised by now.

Other than to give us entertainment and know, another has joined the group of those that play with no hope of a game that works, but continue regardless .

DAYLEET, on 05 September 2016 - 12:00 PM, said:

What do you want: wait 10-15 minutes to drop or drop fast. Chose wisely.

This is the issue and the real reason for PSR's upward trend

T1 complained about wait times, so the 'buckets' yes those things were opened faster and faster, knowing that oneday everyone will be T1 and buckets won't matter ever again..

[MW222]

DAYLEET, on 05 September 2016 - 12:00 PM, said:

What do you want: wait 10-15 minutes to drop or drop fast. Chose wisely.

Never had to wait that long before PGI IMPROVED IT.

[Idealsuspect]

FupDup, on 19 August 2016 - 07:46 AM, said:

I think that the tier system shouldn't be static, it should be relative. What I mean is that the top tier should be dynamically calculated to encompass the best players while the bottom tier should be calcuated based off of the worst.

When the best get better, the top tier rises up with them. Likewise, the bottom can go downwards if potatoes manage to potato harder.

This does technically mean that your tier could actually decrease while you're not playing, if everybody around you got better during your absence.

Too complicated for this firm ...

[Tibbnak]

I think that showing someone's tier in the tab score board would go a long way to helping clear up any misunderstandings of matchmaking. You know what tier someone is, and you can visually see how the matchmaker functions.

[Navid A1]

Idealsuspect, on 05 September 2016 - 01:23 PM, said:

Too complicated for this firm ...

I'm sure they can do that.
But they do not F*ing care.

All their idea of balance and MM is doing group queue streams where Tina is protected by a meta lance.

If they really want to get an idea about this game... they must forget about work for one week and play solo non-stop... both in quick play and FW.

[MW222]

Tibbnak, on 05 September 2016 - 01:34 PM, said:

I think that showing someone's tier in the tab score board would go a long way to helping clear up any misunderstandings of matchmaking. You know what tier someone is, and you can visually see how the matchmaker functions.

DING! DING! DING! We have a winner!

[Goose]

Just remove the damn release valves …

[Navid A1]

Tibbnak, on 05 September 2016 - 01:34 PM, said:

I think that showing someone's tier in the tab score board would go a long way to helping clear up any misunderstandings of matchmaking. You know what tier someone is, and you can visually see how the matchmaker functions.

That will not solve the problem.

the problem is this the PSR system itself:
The tier system is off at the moment... the number of players are so low that the MM can not go with ±2 Tier.
I have had people saying that they are T4s and 5s over VOIP.

A better way to go is to base PSR simply on match score average of a given team
If you score higher than average, you go up based on how higher than average you were
If you score lower than average, you go down depending how low you were from the average

Edited by Navid A1, 05 September 2016 - 10:00 PM.

[smokytehbear]

crashlogic, on 20 August 2016 - 02:07 PM, said:

There are three kinds of lie: lies, damned lies, and statistics. Saying win loss averages out is a statistical platitude. The reality that in every match some player go down who don't deserve to and some players go up who don't deserve to. The deterministic variable is winning and losing. It is much easier to play well and lose with a bad team because oh yeah, its a team game.

This quote again...

Edit: I don't even know what you're trying to say. I honestly can't tell whether you agree, disagree, are calling me a charlatan, are condescending by reminding me it's a team game... I just don't know man, but I had written a response to go with what I assumed you meant off of a best guess, so I'll just leave it here anyway. /edit

By "average out" I was talking about stochastic eventualities. If you're a good player, you can still lose. Eventually though, if you're a good player, you will win more than you lost. Measuring something like a player's skill with deterministic means (which I would argue is most definitely not a W/L record but instead whatever this abomination of match score we have now is) would be next to impossible for an algorithm. There's just way too much stuff in a game to be able to attach numbers to it, and trying to with a simplistic effort is bound to wind up in failure. It's not even really statistics, it's simpler than that.

As an example, temperature is a measure of average random kinetic energy in a substance. Tons of molecules are flying around at all kinds of different speeds, crashing into each other and slowing down or speeding up. The match score system is akin to trying to find out how fast they're going by plucking each one out of the air and using a radar speed gun to count it up, and doing this 10^23 times. The intelligent way to do it on the other hand, is to use a thermometer. Using a stochastic means to determine a general bias or trend in a highly random environment is almost always the way to go.

Yes, it's a team game. You're part of that team, and so are 11 other people. If you play well consistently, you will help your team and more often than not, you will win. You will still lose ocassionally though, and even the worst player ever will still win occasionally. Neither of these two instances invalidate the entire process from still being generally accurate, and comparatively, way more accurate than you'd get using a deterministic method like match score and counting how many enemy mechs you spotted first.

Which reminds me, the problem with the temperature analogy is that the molecules aren't being rewarded for being deemed "fast." Instead in the real case, people know how to score well and they can and do exploit that. How many times have you seen the radar lock box jump back and forth rapidly between a few mechs? That's somebody "scouting" for you to get those Cbills. He's contributing a whole lot right? Seems just as much as the scout that actually does his job and spots the KDK-3 flanking your heavy lance, let's give them the same reward.

If you find a way to exploit winning, let me know. Actually, let everyone know, since I mean that's the whole point right? I think I have an idea how though myself, it's called actually being good.

Edited by smokytehbear, 06 September 2016 - 12:48 AM.

[Willard Phule]

PGI has NEVER been able to make any form of a balanced match. Hell, they didn't even try with FW...

The original matchmaker (you remember that whole Elo nonsense) was just a temporary fix that they tried to turn into a thing. Same with PSR.

Keep in mind that these are the people that brought you "ghost heat"....as opposed to simply changing the allowable rate of fire. Then they realize that their hit registration is bad, so they fix that....then go around "nerfing" everything in the name of "balance."......right up to when their registration breaks, yet again.

It's almost as if they don't care, as long as people keep paying real money for fake mechs.

[adamts01]

Dread Render, on 19 August 2016 - 06:06 AM, said:

This board has been full of posts about it for a very long time.
Your just noticing now?

When PSR was first introduced it wasn't that bad, but slowly people have ground up their XP bars and now tier doesn't mean a thing, except for tier 5 LRM assaults, everyone should judge them. So it absolutely has been getting worse, and everyone has a different threshold for ********. The white knights say otherwise, but I'm convinced the population is quickly dwindling.

Mystere, on 19 August 2016 - 07:34 AM, said:

Come to think of it, how should such a reset be done? Are you thinking of bringing everyone back to T5? Or are you thinking of using some other method?

Resetting everyone to 5 with a PSR that doesn't trend upward would be great. Also, only 3 tiers and only 33% of players can make it to tier 1, so someone has to knock down a tier 1 to move up from tier 2. As soon as the tier 1s and 2s fill up anyway. It wouldn't be so bad, I took a 6 month break from electricity, lived in my hammock on the beach, came back and made tier 1 in a week-long MWO binge. Things would settle down nicely in a month or so.

[Navid A1]

Just resetting PSR will accomplish nothing. It will be a temporary band aid.

PSR variation should be based on how much more you scored above team average in a specific weight class regardless of win or loss...Score higher than average and go up. Score lower and go down.

Only then it will mean something other than an XP bar.

And that is only the simplest way. If you really want to do a proper skill rating. You MUST have a rating for different mech builds.. a dynamic value (affected by weight class, weapons and heat management) that act as a weight multiplier to your match score.

[justcallme A S H]

smokytehbear, on 19 August 2016 - 03:07 PM, said:

No. Get rid of PSR and implement something that actually considers skill, not raw damage.

I'm almost max T1 now after QP'ing again, in less than 4 weeks... I did a total of around 150dmg in a light earlier tonight in a match. I popped a perfect UAU and lots of "UAV Locked Damage/ECM counter". My PSR went "^" - in a win.

I was expecting it to be neutral at best, likely to go down as it was a rubbish game overall.

Next match, I did 500dmg/1 kill and a few components in an assault and got the "=" - in a loss.


I also thought it was based heavily on damage, if it is damage related then it must be somehow tied into weight class to some extent or winning carries far more weight.

I dunno but either way it doesn't make sense and it is not as heavily tied to damage as people think.


I can also confirm T1 down to T5s in solo queue are all being lumped in together.

Edited by justcallme A S H, 06 September 2016 - 05:02 AM.

[justcallme A S H]

And just did 270 damage in an assault, one kill - and a WIN.

PSR up. Should've gone down IMO.

Very much unrelated to damage and all about the win, for the most part it seems.

[adamts01]

Navid A1, on 06 September 2016 - 04:37 AM, said:

PSR variation should be based on how much more you scored above team average in a specific weight class regardless of win or loss...Score higher than average and go up. Score lower and go down.

I still think W/L is all that matters. Computers can't judge who actually contributed and who padded their stats or cheated the system. You've got a 1.94 W/L so your scores are obviously well deserved, but many other's scores aren't.