36 replies to this topic

[Mystere]

Nema Nabojiv, on 24 January 2018 - 10:44 AM, said:

It was. Before 12v12 we had close games and duels of heavily crippled mechs in the end. Since nothing been heard of switching back to 8v8 lately, I guess we're not seeing it again.

Good!

[kentares kodiak]

Lykaon, on 24 January 2018 - 12:27 PM, said:

The issue is quick play puggie teams are not good at recognizing these circumstances or even knowing how to adapt and play differently. Many players whouldn't know thing one about switching tactics mid game to compensate for a disadvantaged circumstance.

Some players are here just for the fun of shooting stuff. They wont bother with teamplay or tactics or thinking whats the best course of action in a given scenario (I even dare to think they dont worry about winning). Too many times I hear/read "stay together" and one or two guys just ignore it and do whatever they want. Too many times I see a couple of mechs pushing into the enemy and I think "cool!... Ill help too!" and suddenly Im charging the enemy alone because the other guys got cold feet at the last moment...

Edited by kentares kodiak, 24 January 2018 - 03:30 PM.

[Moonlight Grimoire]

Having been through ELO, through the beginning of PSR from T5 up and midway through T2 now. PSR has degraded in quality. Self sorting has occurred via this system and it has fundamentally made the game worse due to stagnation. This has led to it isn't about picking a good mech, but, picking a mech you can do well in without support of a team, as well as aiming to get as close to 250 match score as possible. Why? Break even point for on the loss for PSR. Even when I do get to T1, the quality of matches won't change as testament to what B33f shows off on his streams and from what others in my unit experience in T1.

What is needed is a change of how MM is handled in a few ways. Sure, player side you need to find a mech you can do good in, there are mechs most do well in (MAD-IIC), but, finding something you just generally do well in is priority in this day and age.

People talk about Zero Sum a lot so let's tackle something like that. Zero Sum sort of works. It would require seasons to exist to really end up working otherwise it falls apart, yes ELO is Zero Sum, but, the issue is that even with say Top 4 performers on the winning team go up massively, middle 4 go up moderately, and bottom 4 stay the same, enemy team top 4 stay the same, middle 4 go down a bit, and bottom 4 drop massively in 'standing' you still have inactive players filling ranks in the background. These inactive players would have to be culled per month if they played less than 8 matches or something arbitrary like that, not too low, but, something reflecting an hour or two of play a month. So it reflects someone not overly dedicated to the game but also regularly playing.

Tiers being broken up into Standard Deviations would be ideal. T1 is STD-3, T2 is STD-2, T3 is STD +&-1, T4 STD+2, T5 is STD+3. Though it might need to be tightened due to how big T3 would end up being and how small T1 and T5 would be (but T5 should be small and bleed into T4, same with T1 and bleeding into T2). This would instead of having everyone eventually filter up to T1 in PGI's idea, or as it has in reality made everyone slowly migrate to T1 or T5, the Extremes in a bi-modal distribution, we would get a bell curve for the distribution of population. Sure the beginning of each quarter would be rough as there would be no rank with a reset as resetting PSR is an absolute however I think it would be for the better as it is another way to purge inactive players. From here we have quarterly resets culling inactive players helping keep tiers more clearly defined, (could even have monthly purges of inactive players as outlined above to help keep that resolution).
Make Match Score tie into things other than damage to a greater extent, reward crippling mechs, reward kill assists, reward decapping heavier, reward stopping a cap by standing in it after an enemy started capping. Little things that yeah sure it could inflate one's match score but it could show they are capable of knowing how to WIN a Match not just how to put damage down range (as it is at a 1/3rd your damage is turned into match score now I believe, down from 1/2). While damage is important, it would give a bit more reward to those who are actively doing things to win the match for the team other than dumping damage into enemies dead sides. Also, possibly quick kill rewards for people who manage to efficiently take out mechs (so back stabs, XL kills) that is like 100XP-% left of mech before it died, and some decent match score to offset the 120-300 damage they are trading out doing to back stab to XL instead of farm. It is a way to also reward newer players (throw a big fat cbill bonus to quick kills) to learn how to identify where a mech is weak and shoot it there and try to identify what is an XL build and what isn't. While I know XL builds are a lot less common with the introduction of Light Fusion Engines, there is still a place for them, and knowing if a mech is a LFE or STD can also be just as valuable. While this would to a degree skew MS upwards in a way, it is trying to make MS mean more by rewarding knowing what is needed to win by killing and manipulating the mode to leverage things to your teams favor.

TL;DR, seasonal resets of PSR, change up how PSR is calculated so T3 is where majority of players are, not T5 or T1 over time as PGI designed it. Only so many people go up in PSR and go down in PSR per match depending on how well they did in the match.

[El Bandito]

KodiakGW, on 24 January 2018 - 11:11 AM, said:

You said your -B stats were pure solo. Hmmm...guess nobody in MS willing to work with him doing some two man drops with his MCII’s.

You are assuming wrong. There are plenty of people in (MS) who drop in GQ. Personally, I can show him pointers on how to pilot mechs better, but I prefer to drop solo in QP.

[KingCobra]

Well CPT this is a easy one to figure out =(they let groups back in solo play) duh.

[ShaneoftheDead]

I've said it elsewhere, and I'll say it again. Matchmaker is busted because it measures things all wrong. Balanced matches are rare and random because they are happy accidents.

Player Tier and PSR is not the right stat to use. Tonnange is not the right stat to use. Win/Loss is not the right stat to use.

Doing comparisons between players on Wins is amateur, at best. 23/24ths (95.83%) of a win is determined by other people and has nothing to do with that person's performance or skill. Yet it is a major factor in the current player rating system. This is obvious in the times when you die early and your team carries you to victory anyway. Also when you do awesome and yet lose.

Player ratings are well and good, but the main thing is what Mech are they in and what weaponry does it have? Put that together and you get a much better idea of its combat potential in a match. But, a general player rating is meaningless.

Tonnage and Weight Class are wrong because Mechs of the same weight have widely varying performance. Hell, Chassis of the same Mech have widely varying performance. The only possible way to do it is to have each load-out of each variant assigned a Combat Value. Base it off average performance of that load-out for start. But this would only be for the first few dozen games.

Past that, I would then take a Player's performance in a Mech with a specific load-out and assign them a performance rating in that load-out according to a calculation based on Damage Per Drop multiplied by their Kill to Death ratio. That will give a good rough estimate of their potential in that load-out. Lots of damage but few kills will get an average score. (think LRM boats) Lots of kills but little damage will get an average score. (think kill steals) Lots of damage with lots of kills will get a high score. (Good Players) Little damage with few kills will get a small score. (Bad Players)

Match Maker should build teams trying to balance that. And if it cannot get close, give the lower value team an advantage like how winning scouting gives your side advantage in Invasion. If it's a 12-man of a good competitive team vs. a few small groups of average players, why not give the average players Radar sweeps and Mech ID? Play around with the values where they kick in and you can tell how much the advantage is worth.

[justcallme A S H]

Racerxintegra2k, on 24 January 2018 - 08:09 AM, said:

I was complaining about this the other night and everyone in my unit was in agreement. Quickplay has become a terrible place. I've been solo'ing with a K/D of ~1.7 ish and have a .68 W/L. Something is very wrong with match quality and last night i went and did laundry instead of playing more.

Killing more than you die does not mean you should be winning more.

It just means you are killing stuff.

You might be:
1. Losing trades
2. Being out of positon
3. Not focusing targets

etc. The stuff that actually contributes to wins. Kills is not necessarily a proper gauge of performance as a standalone number.

Some games I only get 1 kill... Yet I cored 10 mechs and we win the game 12:0
Some games I kill 6 mechs and its 12:3

Which game did I contribute more towards? Game 1, as scoreline was better.

[KodiakGW]

Laser Kiwi, on 24 January 2018 - 11:01 AM, said:

Not to derail the thread, but what is a T2 performance, look at my stats, is that T1? I'm like 79th percentile, i'm literally the definition of second tier in a five tier system, yet here i am stuck in T1 with you lot.
*snip*
Which begs another question, is it too easy to increase pilot skill.....

That question is the basis of my comment. Do you think you are on par with those other T1s sitting at the 99th percentile on that list? Because, the current matchmaker believes you are, and it will use you to fill out a team matching you against them. Is that fair to you?

Pleanty of people seem to thing so. “Stomps will happen.”

The solo queue matchmaker may be a marvel to behold, pulling and placing players within milliseconds to keep teams near even with tonnage and a parameter on how well those player are expected to perform. But, if the parameter mentioned is flawed, then the matchmaker is therefore inherently flawed.

[Asym]

El Bandito, on 24 January 2018 - 09:20 AM, said:

Pick a good mech and make some difference in Solo-Q. My MCII-B currently has 8:1 WLR, 24:1 KDR, and 1078 average damage per match--all in Solo-Q. I aint complaining.

El, you should be because inflated and skewed "statistics" doesn't prove you are improving.......it all goes back to the farming analogies in that if you farmed a bunch of noobs and your W/L and KDR jump up, does that make you a better player? Is that success? Is that skill?

I would contend that if you played all of the same pilots of the same skill level you are, you'd not see those results and the statistics would say so with a lot more validity. HUGE errors are compounding now in-game it's sliding downhill at an expedited rate....

The OP has it right in his observations. We, the team I am on, aren't playing MWO at all. We had what, 8 players of 56 play in January. I'm playing for the 2x XP and it's been a terrible experience......more than the usual messes. Events are the only thing bringing players to play and once I am done mastering this mech, I'm off to sea again where I am having a lot of fun grinding ships and being with my team...... We're not completely gone but we are mostly gone and will be that way until MWO changes or falls apart. We love MWO and it is heart breaking to see it in such straights.

Edited by Asym, 25 January 2018 - 12:50 PM.

[Q]

How many legitimate tier 1 pilots do you think we have?
(Ignoring the xp bar warriors, how many of those LOOK like tier 1 pilots because they're matched with other tiers?)
How many of those are typically on at the same time?
How many of those have complimentary chassis choices at the right time?

What the heck do you guys expect from a 24/7 drop in pvp game that you can ignore/avoid comms/teamwork in?

Edited by Q, 25 January 2018 - 01:53 PM.

[El Bandito]

Asym, on 25 January 2018 - 12:47 PM, said:

El, you should be because inflated and skewed "statistics" doesn't prove you are improving.......it all goes back to the farming analogies in that if you farmed a bunch of noobs and your W/L and KDR jump up, does that make you a better player? Is that success? Is that skill?

I would contend that if you played all of the same pilots of the same skill level you are, you'd not see those results and the statistics would say so with a lot more validity. HUGE errors are compounding now in-game it's sliding downhill at an expedited rate....

The OP has it right in his observations. We, the team I am on, aren't playing MWO at all. We had what, 8 players of 56 play in January. I'm playing for the 2x XP and it's been a terrible experience......more than the usual messes. Events are the only thing bringing players to play and once I am done mastering this mech, I'm off to sea again where I am having a lot of fun grinding ships and being with my team...... We're not completely gone but we are mostly gone and will be that way until MWO changes or falls apart. We love MWO and it is heart breaking to see it in such straights.

You completely missed my point. Sure, my MCII-B stats will stabilize in the long run, but overall it will still have over 2:1 WLR because its is a damn good mech that can be easily piloted by anyone effectively. I am simply telling OP to run mechs that can affect the battle as much as possible, in order to tip the advantage to their side. Too many times the match is lost cause pugs bring crap mechs with crap builds.

[Racerxintegra2k]

justcallme A S H, on 24 January 2018 - 07:06 PM, said:

Killing more than you die does not mean you should be winning more.

It just means you are killing stuff.

You might be:
1. Losing trades
2. Being out of positon
3. Not focusing targets

etc. The stuff that actually contributes to wins. Kills is not necessarily a proper gauge of performance as a standalone number.

Some games I only get 1 kill... Yet I cored 10 mechs and we win the game 12:0
Some games I kill 6 mechs and its 12:3

Which game did I contribute more towards? Game 1, as scoreline was better.

I get what you are saying and you are right in that coring 10 mechs without killing them is basically a carry. I'm not able to carry hard all the time. I'm just average, and really dont have interest in playing the way it takes to be great.

My point is this if i can give my team a 1 mech advantage if the match maker actually evened out the teams right i'd be a lot closer to 1.00 W/L. When i drop i'm basically giving my team a mech advantage they just cant capitalize.

[ROSS-128]

If your K/D and W/L aren't tracking together well, you might want to work on timing. A kill, or even a component destroyed, near the beginning of the match can be worth a lot more than three kills at the end of a match. Of course, you don't always have much control of when opportunities pop up, but you can try to nudge things a little earlier by focusing components you know/expect to be weak, or double-teaming someone early on to try and secure a quick kill.

If you're often the last man standing and tend to pick up a lot of kills mopping up damaged mechs at the end, it means you might be playing too cautiously early on. Knocking out fresh mechs early will hurt your K/D ratio because you'll take a lot more return fire, but it should improve your W/L ratio because each kill will be worth more. And that should cause those two stats to track together better.

[Racerxintegra2k]

ROSS-128, on 26 January 2018 - 09:03 AM, said:

If your K/D and W/L aren't tracking together well, you might want to work on timing. A kill, or even a component destroyed, near the beginning of the match can be worth a lot more than three kills at the end of a match. Of course, you don't always have much control of when opportunities pop up, but you can try to nudge things a little earlier by focusing components you know/expect to be weak, or double-teaming someone early on to try and secure a quick kill.

If you're often the last man standing and tend to pick up a lot of kills mopping up damaged mechs at the end, it means you might be playing too cautiously early on. Knocking out fresh mechs early will hurt your K/D ratio because you'll take a lot more return fire, but it should improve your W/L ratio because each kill will be worth more. And that should cause those two stats to track together better.

Much to my own scores determent i'm a front line mechwarrior. I'd say 85% of the time i'm right in the brawl or trying to get into the brawl. I'm exactly the type of player you described but my K/D and W/L are not tracking at all.

[justcallme A S H]

Q, on 25 January 2018 - 01:12 PM, said:

How many legitimate tier 1 pilots do you think we have?

How many of those are typically on at the same time?

500-1,000 out of 28,000 currently active.

Even then the gap between #1 and #999 is massive to boot.

[Khobai]

if player skill follows a standard bellcurve then about 15% of players belong in tier 1

but 5 tiers isnt really enough to accurately seperate out the truly good tier 1 players from the bad ones

[Davegt27]

El Bandito, on 24 January 2018 - 09:20 AM, said:

Pick a good mech and make some difference in Solo-Q. My MCII-B currently has 8:1 WLR, 24:1 KDR, and 1078 average damage per match--all in Solo-Q. I aint complaining.

I was saying the other day the try hards have been blowing people to bits
with there Mad Cat MKIIs

some type of unbelievable UAC10 rate of fire (at least I think its a UAC10s I cant seem to last long enough to tell)

you can either suck it up or stop playing