849 replies to this topic

[Z Paradox]

Brauer, on 02 June 2020 - 07:32 AM, said:

Meta mechs are fun mechs. What's fun about nerfing yourself?

wining in narc rvn with glass armor and getting some kills is fun, wining in FLE, SDR, JN7, WLF... with 4+ kills is just wining...

Edited by Z Paradox, 02 June 2020 - 09:50 AM.

[Brauer]

Z Paradox, on 02 June 2020 - 09:49 AM, said:

wining in narc rvn with glass armor and getting some kills is fun, wining in FLE, SDR, JN7, WLF... with 4+ kills is just wining...

Wait, are you claiming a spider is a good mech? Really they're only meta for the capping spider last I knew.

[Z Paradox]

Brauer, on 02 June 2020 - 10:28 AM, said:

Wait, are you claiming a spider is a good mech? Really they're only meta for the capping spider last I knew.

[Brauer]

Z Paradox, on 02 June 2020 - 10:40 AM, said:

That's certainly better than I'd expect any SDR to do, well done!

[Dionnsai]

Kosomok, on 02 June 2020 - 07:29 AM, said:

Win/loss is not really a measure of player skill. It is largely beyond the individual player's control. It should not figure in the calc at all.

This is a terrible and completely wrong statement that can be easily picked apart. A player's win/loss rate is a direct reflection of their personal impact on matches. Now, in any individual match, and across a sample of say 10 matches, that number might be meaningless.

However, if we look at a subset of say 1000 matches:

If the player's Win/Loss is 1.0 you could say that the player on average neither helps nor hinders the team towards the eventual outcome.

If the player's Win/Loss is 2.0 across 1000 matches, it becomes obvious that this player's mere presence on a team increases the odds of a win significantly, indicating that either they are extremely deadly, or very good at influencing a team towards a win, see that player's k/d & kmdd to infer more.

If the player's Win/Loss is 0.5 across 1000 matches, this player is generally a detriment towards the team and any team in which he is a member is more likely to lose because of him. You'll probably find that the same pilot also has an extremely low survival rate and k/d.

Obviously for any individual pilot, there is a bigger picture in the statistics that you will have to look at, but in general where they fall in the win/loss curve will at least give some indication of their ability across a large set of games.

[Anomalocaris]

Horseman, on 01 June 2020 - 09:53 AM, said:

Which is why it is folly to try using Steam's data as exhaustive. You need to wait for Jarl's to get May and June data at the very least.

EDIT: May data is in: https://leaderboard.isengrim.org/stats

• 17K players
  
• 2K new players

Essentially, the numbers are back to values from last August. With the lockdowns ending soon, we'll get to see how much of that influx remains.

And again, note the trend line. Slope is almost unchanged starting from March onward. Same goes for new player adds. Would normally expect a major game change to alter the slope of the curve somewhat.....

[DAEDALOS513]

Dionnsai, on 02 June 2020 - 01:47 PM, said:

This is a terrible and completely wrong statement that can be easily picked apart. A player's win/loss rate is a direct reflection of their personal impact on matches. Now, in any individual match, and across a sample of say 10 matches, that number might be meaningless.

However, if we look at a subset of say 1000 matches:

If the player's Win/Loss is 1.0 you could say that the player on average neither helps nor hinders the team towards the eventual outcome.

If the player's Win/Loss is 2.0 across 1000 matches, it becomes obvious that this player's mere presence on a team increases the odds of a win significantly, indicating that either they are extremely deadly, or very good at influencing a team towards a win, see that player's k/d & kmdd to infer more.

If the player's Win/Loss is 0.5 across 1000 matches, this player is generally a detriment towards the team and any team in which he is a member is more likely to lose because of him. You'll probably find that the same pilot also has an extremely low survival rate and k/d.

Obviously for any individual pilot, there is a bigger picture in the statistics that you will have to look at, but in general where they fall in the win/loss curve will at least give some indication of their ability across a large set of games.

That would be true if every player was dropping solo all the time.. where players are randomly selected.. ie. random teamates and random opponents.. otherwise you can game the system immensely by dropping strictly with say some tier 1 buddies where you can easily carry each other to astronomical w/l ratios. In this soup queue environment where solos are dropping mixed in with grouped players, psr should not be calculated by match wins but rather by contribution regardless of outcome. If anything, match outcome should play only a small part in psr calc.

Maybe using match score would be a good start as that takes into consideration many things.. correct me if i'm wrong but doesn't match score take into consideration missiles shot down, flanking, damage done, tag locks, uav spotting, etc? I'd be curious to know if anyone knows the calculation for match score..

Edited by DAEDALOS513, 02 June 2020 - 02:21 PM.

[Brauer]

Dionnsai, on 02 June 2020 - 01:47 PM, said:

This is a terrible and completely wrong statement that can be easily picked apart. A player's win/loss rate is a direct reflection of their personal impact on matches. Now, in any individual match, and across a sample of say 10 matches, that number might be meaningless.

However, if we look at a subset of say 1000 matches:

If the player's Win/Loss is 1.0 you could say that the player on average neither helps nor hinders the team towards the eventual outcome.

If the player's Win/Loss is 2.0 across 1000 matches, it becomes obvious that this player's mere presence on a team increases the odds of a win significantly, indicating that either they are extremely deadly, or very good at influencing a team towards a win, see that player's k/d & kmdd to infer more.

If the player's Win/Loss is 0.5 across 1000 matches, this player is generally a detriment towards the team and any team in which he is a member is more likely to lose because of him. You'll probably find that the same pilot also has an extremely low survival rate and k/d.

Obviously for any individual pilot, there is a bigger picture in the statistics that you will have to look at, but in general where they fall in the win/loss curve will at least give some indication of their ability across a large set of games.

Good description Dio. I'd just add that now with groups in the mix WLR in QP is also heavily influenced by A) how much the player drops with groups, and the overall impact on WLR of the groups they drop with. If four players with pre-merge WLRs of around 2.00 group up they can relatively easily start putting up double-digit WLRs.

DAEDALOS513, on 02 June 2020 - 02:15 PM, said:

That would be true if every player was dropping solo all the time.. where players are randomly selected.. ie. random teamates and random opponents.. otherwise you can game the system immensely by dropping strictly with say some tier 1 buddies where you can easily carry each other to astronomical w/l ratios. In this soup queue environment where solos are dropping mixed in with grouped players, psr should not be calculated by match wins but rather by contribution regardless of outcome. If anything, match outcome should play only a small part in psr calc.

Maybe using match score would be a good start as that takes into consideration many things.. correct me if i'm wrong but doesn't match score take into consideration missiles shot down, flanking, damage done, tag locks, uav spotting, etc? I'd be curious to know if anyone knows the calculation for match score..

If people are consistently dropping in a strong group it just shows that they understand the new merge queue meta well and are using it. Technically from a metagame point of view their WLR is accurate as they are doing quite a bit right to drive wins. People might not like it, but that's the new meta that some people apparently wanted.

[Nesutizale]

Wouldn't those Meta-Groups move to the top tier and play among other Meta-Groups over time anyway so that the non Meta players are then again fighting among themselfs?

At least that it what should happen, when there are enough players for both sides online at the same time. Witht he player numbers we have that might be problematic. Still shouldn't it work out, theoreticly?

[DAEDALOS513]

Brauer, on 02 June 2020 - 02:33 PM, said:

If people are consistently dropping in a strong group it just shows that they understand the new merge queue meta well and are using it. Technically from a metagame point of view their WLR is accurate as they are doing quite a bit right to drive wins. People might not like it, but that's the new meta that some people apparently wanted.

But why not make psr do its' job and at the same time actually have it mean something, Players are fond of comparing stats and this is a big driving force for many players to grind and do their best.. using match outcomes as a sole or major part of the psr calculation would render psr meaningless except for sub-optimal matchmaking purposes.

Edited by DAEDALOS513, 02 June 2020 - 03:20 PM.

[DevinMace]

I really like that you can be able to group with friends and I don't even do it, but I also think there is merit in wanting a "solo" ladder I guess. One issue of coarse with the 8v8 was the QP queue fell off a lot, which was a problem. Can we not make FP more accessible, which I also know would not be a very popular idea.

[OneTeamPlayer]

Brauer, on 01 June 2020 - 03:55 PM, said:

afaik there is no talk of increasing group size. Drop as a 3 and a 2. You'll likely often get dropped in the same matches and even on the same side on occasion.

Which is another reason matches have been the way they have been lately.

Imo if they're going to have a limit of 4 mixing the solo and group queue, they should also specifically hard code a limit on separate groups of the same unit dropping together to circumvent the system.

Pretty bold to go on the official subforum and point out ways people are cheating the system which is already generous enough to allow groups to enter solo play.

Edited by OneTeamPlayer, 02 June 2020 - 05:18 PM.

[Brauer]

DAEDALOS513, on 02 June 2020 - 03:12 PM, said:

But why not make psr do its' job and at the same time actually have it mean something, Players are fond of comparing stats and this is a big driving force for many players to grind and do their best.. using match outcomes as a sole or major part of the psr calculation would render psr meaningless except for sub-optimal matchmaking purposes.

I don't think you can get a truly consistent PSR with groups and solos mixing as it removes the randomness that actually makes measuring individual performances in a team game halfway valid.

[Brauer]

OneTeamPlayer, on 02 June 2020 - 05:08 PM, said:

Which is another reason matches have been the way they have been lately.

Imo if they're going to have a limit of 4 mixing the solo and group queue, they should also specifically hard code a limit on separate groups of the same unit dropping together to circumvent the system.

Pretty bold to go on the official subforum and point out ways people are cheating the system which is already generous enough to allow groups to enter solo play.

I think they should prevent more than 4 grouped up players from dropping on the same side of a match, but they don't. All I've done is explained how the group system currently works. I have screenshots of over 90 percent of my groupq matches and have not been attempting to sync drop (there was one unintentional sync to the same side with a couple other unit members a few nights ago, but that's the only same-side sync I've personally experienced). The reality is you definitely do not need to sync drop to put up a double digit WLR with the queue as it is. Groups of 2 and 3 getting together on the same side is not good, but it's not the reason that the matchmaker and queue are a mess. You can lay that at PGI's feet imo.

Anyway, I don't think there's much point arguing that anyone is "cheating the system" when the system in place is such a half-baked mess.

[DAEDALOS513]

Brauer, on 02 June 2020 - 05:19 PM, said:

I don't think you can get a truly consistent PSR with groups and solos mixing as it removes the randomness that actually makes measuring individual performances in a team game halfway valid.

Sorry but ya lost me there..

[Brauer]

DAEDALOS513, on 02 June 2020 - 07:12 PM, said:

Sorry but ya lost me there..

In old solo queue while group queue was dead the fact that teams were random meant you could assess a player based on all of their stats. Now that groups are involved, in my opinion, you either accept the stats as they stand and merely as a demonstration of how well a player leverages all possible advantages (including grouping up) or the stats are basically garbage because teams are no longer random. Any individual stat is going to be heavily influenced by whether or not a player is in a group.

For example, dropping with other strong players likely lowers average match score while raising WLR, KDR, and survival rate. Dropping solo has a chance of severely depressing WLR and to an extent KDR, average match score, and certainly survival rate, but really the impact of dropping solo is hard to predict because it depends so much on what groups are in the queue when you drop.

TLDR, imo you either take the stats at face value and all are to be considered valid, or they're all garbage.

[Aidan Crenshaw]

Actually, a win gives you exactly 14 points matchscore over a loss. The rest of the points is gained via your other actions in a match. Due to the snowball effect, the losing side has less of these, of course.

[MrMadguy]

I O O percent KongLord, on 02 June 2020 - 07:40 AM, said:

I Think you can be viable for quick play in any mech (which means not playing meta) but I think you misunderstand what fun is, fun is not a bad build fun is whatever you like to play.

Also when u say missiles mean more match score... Mate u are 40% on jarls (which means you are worse than 60% of the people this game has to offer)... Ur opinion regarding that holds no value...

I could be doing this, but I don't want to. Because my "DMG record" race - is what actually caused my MM rating to overinflate and ruin this game for me at the end.

About Meta 'Mechs. May be some of them are fun. MKII for example. But not all of them. When people pick ER-LL/ER-PPC Mediums, like ugly Shadow Cat, in order to pew-pew and steal kills from safe distance - this just isn't fun.

And I love that "Your rating is low, so you're wrong" fallacy. First of all, Jarl's isn't right. In reality, when ELO->PSR translation was done, I was seeded at exactly 50% Tier 3, i.e. in a middle of normal skill distribution. May be I'm not right, but I'm majority, I guess. And yeah, that screenshots from old times, when this game was good:

Edited by MrMadguy, 02 June 2020 - 10:33 PM.

[RRAMIREZ]

*unproductive post*
good old time when everybody wanted to have Legolas nickname ^^

[Voice of Kerensky]

pseudofiction, on 02 June 2020 - 06:09 AM, said:

I’m not sure if you are being serious or sarcastic when you suggest that balancing changes drove away more people than the imbalances. I can only speak for myself but no amount of nerfing could make me leave game if it was for a good reason. Besides, the nerfhammer would hit both sides and not even that hard. None of the few changes I suggested would invalidate any of the lameboats (for instance the ghost heat on 5+ ac2's would be only minimal, just enough limit the dps by a reasonable amount). They would still count among the power builds. And for those few players who would leave because their favorite lameboat was nerfed, probably good riddance since they likely drove away many more. On the other hand the imbalance that the groups brought both directly and indirectly by sometimes seriously uneven team tonnages and messed up spawns did a bloody good job driving me away and judging by some other comments here, I'm not an exception.

I don’t understand why you started a conversation about balancing weapons / mechs? Do you understand that this topic is about how to balance opposing teams?
You say that the nerf hammer will hit both factions (Clans and IS). What, then, is the point of breaking the balance of weapons / mechs again? A balance has been established for the moment, let it work. People got used to it. Why bend and break the system, which already works with difficulty again?
Leave it. Let me remind you once again that we are now talking about how to balance the opposing teams (into account solo players and premade groups) so that the matches would be interesting, rather than a game in one-gate with a score of 12:0 - 12:4.