152 replies to this topic

[JernauM]

Tarogato, on 11 July 2016 - 01:38 AM, said:

In order to verify that MM is working properly, PGI should check the tier/PSR of all pilots in an "unbalanced" match. If the PSR appear to be balanced, then the problem is PSR, not matchmaker.

PSR is indeed a problem, and the community has known it from the beginning and discussed it at great length. It is pointless to use what is essentially an experience bar for the purposes of matchmaking.

The issue with PSR is that the "R" stands for "Rating", when it should really be "Ranking" instead. A player can only be judged as good or bad in relation to other players. To classify a player as "top-tier" because their experience bar has reached an arbitrary threshold is meaningless.

Any solution to matchmaking should involve tiering players according to a ranking, and not according to some abstract scoring system.

[Ex Atlas Overlord]

Mazzyplz, on 11 July 2016 - 01:46 AM, said:

so winning teams always have better kdr?

why not match kdr average between teams when matchmaker creates match?

b/c KDR varies between Tiers

I can hold a 2.5-3.5 in T1...

But if played in T5 all the time it would be 10+

IMO the main problem is the Tier spreading used to save time....

Back to the same problem that got us here...

You either get good close games and STFU about the wait times
Or get fast games and STFU about bad match making

Edited by Ex Atlas Overlord, 11 July 2016 - 03:00 AM.

[C E Dwyer]

having made a few of my own records since this thread came out, using just ranking in the leader boards, I'm of the opinion that using these stats proves nothing.

This morning had a classic we won 12-0 so I thought this must be ripe for dissecting.

Wining team

19,954
19,038
4518
3827
30,882
6343
2462
2095
21,641
17,726
26,284
3328

Total rank score 158,098 averaging 13,174

Losing Team

15,668
7783
16,952
8423
5407
7114
18,925
6338
14,800
13,530
13,190
4114

Total rank score 132,244 averaging 11,020.

The winning team had 4 players higher ranked than the losing team.
The winning team had 5 players Lower ranked than the losing team.

The winning teams 30k was the lowest scoring guy on the winning team but still managed 120
highest damage score on the winning team was around 580 ish .

in theory, the losing team should have picked off the alleged scrubs and then with numbers pushed over the higher ranked guys of the winning team.

losing team had 3 kodiaks two or them KDK-3 winning team had one Kodiak KDK-3 one Highlander IIC-C

Winning teams kodiak was the 4518 guy the Highlander was ranked 19,038 they both got one kill, the highlander did one less assist and 20 points less damage, it wasn't equiped with any missiles, a lpl's and Uac10's build.

The only thing I can guess at from the comments some of the other team made, at the end, and what was happening in our team was communication we had a guy calling out where they were. We had a very well placed UV showing where 10 of them were, and how they were moving for about 5 seconds before it was shot down, and the guy on our team called got to push now, and we did.

The end result was 12-0 .

You can throw numbers as much as you want to, but the real deal maker or breaker is calling, and a team willing to do the right thing at the right time.

You can have a trytard using viop to rant and rage, have the rest of the team switch off, and ignore the fool.
Or you can have a good caller that talks calmly, gains people trust and win.

Edited by Cathy, 11 July 2016 - 03:41 AM.

[Hodor72]

DemonRaziel, on 11 July 2016 - 12:41 AM, said:

That's the funny thing about statistics - everyone can interpret them differently

"There are three kinds of lies: lies, dammed lies, and statistics" Mark Twain

[VictoriaSeymore]

Belacose, on 10 July 2016 - 10:38 PM, said:

Not is he only using the Kodiak 3 for each and every one of those 285 matches, judging by that curb stomping 4.16 W/L ratio I'd also guess he's solely playing on an All-Star super team as well. He averages 2.8 kills per game.

Even with the KDK-3, wonder how his numbers would look if every game played was a QD pug?

BTW, if it turns out that his games were all QD pugs I'd be shocked, eat my words and want an autograph.

EDIT: Actually those numbers are impressive regardless of anything else. Wonder if he's wearing one of those new VR headsets which allow you to look out the side of the cockpit without having to turn your torso?

Leaderboard statistics on this account are 100% solo queue. It's the primary reason I'm using two accounts at the moment. My main account, kaffeangst is for group/community warfare/competitive practice/etc. VictoriaSeymore is for solo dropping only. I wanted to have separate group/solo numbers. A 4 W/L in solo queue is great. A 4 W/L in group queue isn't the best. Let me pull up some numbers to explain.



^ This is an example of my main account's (kaffeangst) numbers of primarily group queue. Notice the similar K/D, yet vastly improved W/L. If I were playing in the group queue for leaderboard with my unit, House of Lords, my W/L would be closer to the 12 shown and not a 4. Biggest indicator is the Average Match Score though... 633 is averaging 1075 damage/match with closer to 3 kills/match. This isn't possible to maintain if you're stomping group queue with an all-star team.

Edited by VictoriaSeymore, 11 July 2016 - 03:56 AM.

[iLLcapitan]

ok you waved your e-p33n in everybodys face excessively now, topic anyone?

[Belacose]

VictoriaSeymore, on 11 July 2016 - 03:54 AM, said:

Leaderboard statistics on this account are 100% solo queue. It's the primary reason I'm using two accounts at the moment. My main account, kaffeangst is for group/community warfare/competitive practice/etc. VictoriaSeymore is for solo dropping only. I wanted to have separate group/solo numbers. A 4 W/L in solo queue is great. A 4 W/L in group queue isn't the best. Let me pull up some numbers to explain.



^ This is an example of my main account's (kaffeangst) numbers of primarily group queue. Notice the similar K/D, yet vastly improved W/L. If I were playing in the group queue for leaderboard with my unit, House of Lords, my W/L would be closer to the 12 shown and not a 4. Biggest indicator is the Average Match Score though... 633 is averaging 1075 damage/match with closer to 3 kills/match. This isn't possible to maintain if you're stomping group queue with an all-star team.

wOw! You actually have maintained a 4.16 W/L ratio in pure pug matches and with 285 games under your belt proving it's no fluke. Simply unbelievable. I'd have to say that thus far I know of nobody else near as good as you.

Are you wearing a VR headset and/or running wicked macros? Any videos? Would love to see some actual footage of your game play!

[Dread Render]

best thread i ever read on this forum... Great work[color=#B27204] Mister Blastman[/color]

[Mister Blastman]

Cathy, on 11 July 2016 - 03:35 AM, said:

You can throw numbers as much as you want to, but the real deal maker or breaker is calling, and a team willing to do the right thing at the right time.

You can have a trytard using viop to rant and rage, have the rest of the team switch off, and ignore the fool.
Or you can have a good caller that talks calmly, gains people trust and win.

Well, consider that chronicbads don't listen to drop calling and thus when you analyze their stats, will show it. Folks who cower and let teammates soak damage to increase their KDR may have a decent KDR but an average to bad win/loss ratio. Chronicbads, on the other hand, who bumble around and hide or clip their toenails while watching Young and the Restless and reruns of the Spongebob Squarepants are going to have a below 1:1 KDR and below 1:1 win/loss ratio.

I think it is important to consider that leaderboard rank alone doesn't give us enough information. We need to know how the match score (component of damage), kdr AND win/loss relate to one another and then look at the games played to determine if they're new, seasoned or just bad at the game. Someone who plays 1k games and has a 0.7:1 win/loss ratio is a chronicbad.

[Blue Pheonix]

VictoriaSeymore, on 11 July 2016 - 03:54 AM, said:

Leaderboard statistics on this account are 100% solo queue. It's the primary reason I'm using two accounts at the moment. My main account, kaffeangst is for group/community warfare/competitive practice/etc. VictoriaSeymore is for solo dropping only. I wanted to have separate group/solo numbers. A 4 W/L in solo queue is great. A 4 W/L in group queue isn't the best. Let me pull up some numbers to explain.



^ This is an example of my main account's (kaffeangst) numbers of primarily group queue. Notice the similar K/D, yet vastly improved W/L. If I were playing in the group queue for leaderboard with my unit, House of Lords, my W/L would be closer to the 12 shown and not a 4. Biggest indicator is the Average Match Score though... 633 is averaging 1075 damage/match with closer to 3 kills/match. This isn't possible to maintain if you're stomping group queue with an all-star team.

Victoria,

Congrats on those numbers.

However, I have seen you play and you primarily play in your Kodiak. While those are impressive stats for any player regardless of the mech, because of "power creep" this new "cash only chassis" (the Kodiak) is currently overpowered. It would be interesting to see you hold those numbers when PGI inevitably nerfs the Kodiak to put it more in line with other assaults that are available for C-bills. I would put money on your stats would go down, significantly.

I have played quite a few matches. watched quite a few of the final score/stats screens and watched quite a few players play. I am also confident that around 50% - 70% of the mechs in the top 5,000 of the leader boards are from people running Kodiaks. If PGI implements what I am requesting, watch everyone's rating who isn't piloting the Kodiak to skyrocket.

Also, on doing a search of your name in each of the weight classes I see you have no rating in any of the classes except in the Assault (Kodiak's) class. Again, congrats on those numbers but again, this shows me that you just primarily play the Kodiak on this "VictoriaSeymore" account, not any other weight class. Again, this is very telling of power creep in new pay/cash only mechs. It will be interesting when you play the other weight classes too see if your numbers will remain as high. I see you have played over half of 500 games only in the assault class. How come you haven't played any other weight classes on this account?

I also did a search of your main account kaffeangst" and could not find your stats on the leader board. Did you mistype that name?

Thanks

P.S. PGI, please keep MWO from being Pay to Win regarding the leader board.

Keep everyone on the same playing field with the same tools for success by keeping new "cash only" chassis from contributing towards the leader board stats of a player. Let "cash only chassis" players continue to have their fun dominating in the matches but keep the integrity of the leader boards by making sure everyone has access to the same tools at the same time for success. Not everyone has the cash to buy the "new shiney" mech. There are a lot of good pilots in this category. Please don't hamstring them because of power creep. Lets keep everyone on the same playing field regarding the tools available to them. Let talent, skill and effort be the determining factors. Not cash.

Edited by Blue Pheonix, 11 July 2016 - 08:45 AM.

[VictoriaSeymore]

The Kodiak is strong, yes. However, there's always going to be certain mechs that are optimal. I don't believe that the Kodiak would suffer from changes; it could handle negative quirks. You cannot nerf the UAC/10's because the Kodiak would still be powerful with other UAC's, Gauss or any number of possible loadouts. It could also cope with movement and agility quirks because you could compensate with a bigger engine in your build. My stats wouldn't drop as much as you believe. Remember that it would affect everyone and those with skill would still remain at the top. Years ago PGI kept trying to nerf my Highlander, and yet I still won all the leaderboard challenges. Adapt to the changes, and continue.



^ Select "optimal" / META mech numbers from archived/current stats. Kodiak isn't really better than any of these (at their peaks). The only difference is the average damage per match. Great mechs perform and allow good players to carry matches - solo, group, or competitive.

Edited by VictoriaSeymore, 11 July 2016 - 10:22 AM.

[topgun505]

When they intro'd the psr system I started in Tier 2. I expected to see an improvement to games after moving up to tier 1. Not so.

Frequently seeing matches where 6 or more of the losing players didn't break 200 damage (and often don't even break 100). Not sure how MM pairs people up but yeah, it can certainly use some tweaking.

Edited by topgun505, 11 July 2016 - 09:55 AM.

[Hit the Deck]

VictoriaSeymore, on 11 July 2016 - 09:44 AM, said:

WTF

Especially the number of matches. Most I've got is like 300 matches in my HGN-733C since the end of 2014.

[wanderer]

Quote

Calling someone out as good or bad due to their stats isn't right.

I don't care why you're playing. I only care about your effect on my game and the ten other people by extension, too.

If your stats are bad, you're a drag on your team and people can and should legitimately call you out on it.

Quote

It is also a perfect example on why "cash only whole chassis" should not contribute towards stats on the leader board.

Which, by the way is saying "pay to win". Which PGI won't do, so deal with getting mauled by good players in good 'Mechs.

Quote

BTW, if it turns out that his games were all QD pugs I'd be shocked, eat my words and want an autograph.

I hope they taste good with milk. That's a QP-only player that happens to be master-class at carrying.

Edited by wanderer, 11 July 2016 - 12:00 PM.

[wanderer]

Quote

Or you can have a good caller that talks calmly, gains people trust and win.

I'd point at the four players ranked 4K or higher on the winning team as the big reason why they won. The odds of one or more of them carrying hard enough to start a stomp are incredibly high, given their stats- and clearly, they did.

Quote

ok you waved your e-p33n in everybodys face excessively now, topic anyone?

People didn't understand why an account could be a huge outlier and also were mistaken in thinking the stats came from group play. It's not waving your epeen, it's just dropping the sausage on the table and letting the facts speak for themselves.

[Blue Pheonix]

wanderer, on 11 July 2016 - 11:58 AM, said:

Which, by the way is saying "pay to win". Which PGI won't do, so deal with getting mauled by good players in good 'Mechs.

Yes, it is another way of saying "pay to win". Cash should have no influence in leader boards. At all. Only skill, work and effort. I hope this is PGI's mindset. If not, hopefully that changes and soon.

I am a decent player who has probably beaten at one point almost every single player in this game (who regularly plays). I have also probably got beaten by almost every single player. So it has little to do with "getting mauled by good players in good mechs". Ive been on both ends of the spectrum "mauler, getting mauled".

Again, cash should have no influence in leader boards. At all. Only skill, work and effort.

Edited by Blue Pheonix, 11 July 2016 - 12:48 PM.

[Mawai]

On interesting statistic that could come from this analysis is the average number of games played reported for each team ...

Is it a function of the tier that you play in? Do games with higher tier players contain players with on average more games played? This might offer some insight into the PSR as an experience bar meme

[Mystere]

Blue Pheonix, on 10 July 2016 - 07:59 PM, said:

This is a perfect example on why matchmaking should not be random. It is also a perfect example on why "cash only whole chassis" should not contribute towards stats on the leader board.

Actually this entire thread is a perfect illustration of why PGI spending countless work hours on changing the matchmaker is just countless work hours not spent on actually working on the missing and incomplete parts of the game and why they should just have stuck at the original random matchmaker.

[Mystere]

Akulla1980, on 10 July 2016 - 05:03 PM, said:

Calling someone out as good or bad due to their stats isn't right.

...

Not everyone cares about others epeen.

Well, it's too late now. PGI was negligent enough -- or just no longer cares a flying **** -- to have finally just published everyone's stats for all to see. People will now start using the still very much flawed leaderboards for ePeen measurement.

[Jake Peril]

Here you go, for your research:



Score KDR W/L Games
200 0,86 0,68 160
192 0,89 1,01 230
162 0,58 1,11 114
151 0,47 0,8 185
179 0,81 0,85 636
106 0,06 0,79 25
111 0,37 0,64 165
74 0,2 0,71 127
203 1,54 0,96 414
176 0,65 0,99 411
204 0,85 1,14 94
183 0,58 1,19 116
161,75 0,66 0,91 223,08

Score KDR W/L Games
183 0,81 0,67 137
164 0,53 0,79 321
96 0,18 0,63 13
108 0,23 0,67 50
160 0,6 1,33 14
144 0,65 0,78 66
157 0 2 4
160 0 1 2
243 1,43 0,84 160
202 1,02 0,97 614
208 0,96 0,97 150
201 1,11 1,02 384
168,83 0,63 0,97 159,58

This was from a group drop, I left the clan tags in there.
It probably skews the numbers a little bit, since there were not totally random players in the game, but might be interesting nonetheless.
Sorry for the strange number formats, my Excel severely dislikes commas as decimal separators and interprets them partially as dates, so I had to convert them manually.
Otherwise I would have had to change the system settings, which I could not be bothered to.
This forum also seems to hate inserted tables, so you get plain text instead...