692 replies to this topic

[CeeKay Boques]

Be Rough With Me Plz, on 08 August 2014 - 08:09 AM, said:

Just out of curiosity, do you believe the majority of matches being played are in the Group Queue or Solo Queue?

If the majority of the matches being played happen in the Solo Queue, then the way PGI implemented their tests is accurate because it captured the randomness from PUG matches. If you were to hold a test in a controlled environment, not simulating actual conditions of how the majority of the game is actually being played then you would have sampling bias.

...
That's bad logic for a video game. I'm going to say this, as a PuGGer, PUG matches don't matter, especially for testing.

If my mom can get an account today, at play 3pm and influence the study... that's not the metric we want. The PUG playerbase is so wild, so crazy, so full of "lets try this" or "I want to try this new trial mech" or "Daddy let me drive!" or "Where's the rest of my elite group? Ah well, lets pug till they get on" that it makes NO sense to balance around those people play abilities. Am I pugging? Hell yeah! Is my son on my lap? Hell Yeah! Is my 9 year old? Hell yeah! Don't test us for weapon viability though, jesus.

Only in the "I've invested so much energy to MWO that I even joined a team" environment is it even remotely possible to BEGIN to test for things like Weapon and chassis ability.

I'm not disputing the results, they are probably accurate. The test itself is dumb.

[Revis Volek]

Ultimatum X, on 07 August 2014 - 09:01 PM, said:

And nearly every clan team is almost guaranteed to have at least 1 ECM unit.

Every clan team is guaranteed to have an assault that is 85 tons or higher, and very likely to have multiple 100 ton assaults.

Every clan team is highly likely to have 75 ton heavy and at minimum 70 ton heavy, and guaranteed at minimum 50 ton mediums.

Every clan team is guaranteed to have 30 and 35 ton psuedo mediums.


No 60 ton heavies, no 25 ton lights, no 40 ton mediums with "machine gun crit seeker" nonsense builds, no trial mechs.


There's more to the story than "clan weapons", hopefully this weekend will be cleaner and more useful testing.

HOLY SNIKES! Somone who can look at the WHOLE picture and take it in for what it is worth! I commend you Ultimatum X! You are the first person i have seen (includes PGI) who can make a rationale assumption based on the data we have....

Until we have things like Clan trials, Lemons like RVN-H's and such, and more then just the ELITE (META PLAYERS) who have the clan mechs i will say this is an over reaction on PGI's part. They will be dialing these nerfs back in a few weeks im sure....once they see ERLL on the shelves across the board.

[Kyrie]

I find it pleasantly ironic that the "Clan Invasion" is producing stats that actually track the lore of the game despite PGI's best efforts to the contrary.

The truth of the IP behind the game triumphs over Paul's evil ways.

[Be Rough With Me Plz]

Technoviking, on 08 August 2014 - 08:19 AM, said:

...
That's bad logic for a video game. I'm going to say this, as a PuGGer, PUG matches don't matter, especially for testing.

If my mom can get an account today, at play 3pm and influence the study... that's not the metric we want. The PUG playerbase is so wild, so crazy, so full of "lets try this" or "I want to try this new trial mech" or "Daddy let me drive!" or "Where's the rest of my elite group? Ah well, lets pug till they get on" that it makes NO sense to balance around those people play abilities. Am I pugging? Hell yeah! Is my son on my lap? Hell Yeah! Is my 9 year old? Hell yeah! Don't test us for weapon viability though, jesus.

Only in the "I've invested so much energy to MWO that I even joined a team" environment is it even remotely possible to BEGIN to test for things like Weapon and chassis ability.

I'm not disputing the results, they are probably accurate. The test itself is dumb.

I agree to a point. When you're trying to determine something general like how effective Clan Mechs are against IS Mechs then the main thing to look at would be Win/Loss. In this sense the randomness in PUG matches will give a better idea since you're including every player in the Elo Spectrum and also including Trial Mechs. You're getting the entire picture. If you were to do that in a controlled environment then you're leaving out low Elo players and Trial Mechs. You're also adding in communication that doesn't normally exist in PUG matches.

Those three variables aren't something you can just disregard through your controlled environment tests because you're not going to get an accurate representation of the game when you have sampling bias. If you don't test in a way that reflects the majority of how matches are being played out then it's not a significant test.

If you were to talk about specific things like weapon effectiveness then you would have to do that test in a controlled environment where the majority of the variables are locked down.

There are 2 different environments within MW:O. You can't determine anything until you figure out which Queue is a "general" representation of the game.

Edited by Be Rough With Me Plz, 08 August 2014 - 09:09 AM.

[Sephlock]

Mcgral18, on 07 August 2014 - 08:55 PM, said:

Well, 62% of the players did fail to get the pathetic damage of 200...

I'll be trying again the 15th. Private matches and gimped cERLLs should make it much easier.

How exactly is the Private match thing supposed to work? Are we just supposed to play in Private Matches on our own and trust that they will be counted? How much do we need to play to get whatever the "prize" is ?

Jman5, on 07 August 2014 - 09:03 PM, said:

Sorry but this was debunked by PGI.

They accounted for Elo differentials and it was still insanely out of wack. Going by skill it should have been 60/40 worst case scenario. The fact that it was 90/10 means there is imbalance.

Games in MW:O tend to snowball, badly. That tendency was there since Closed Beta. Any tiny differential in player skill is going to have a huge end effect in 12 v 12.

[Mcgral18]

So, any actual numbers released yet? Or will I actually have to head to Twitter and ask.

[Hillslam]

To the OP: you'll never get an apology.

You'll get arguement, rationalizations, subterfuges, feints, strawmen, etc.

Its not the gear! Its never the gear! Its our skillz! It must be! Protect our gear at all costs so we can keep saying its our skillz!

Deploy the words!
Cast Aspersions!
Muddy The Water!

*data arrives*
Uh oh. Combat the data!



The Forum Lawyering goes good with a bag of popcorn, lightly salted, with a little parmesan cheese sprinkled on, a dash of garlic powder, and some white chedder.

*sits back and enjoys*

[Scendore]

Stats can be interpreted to fit someone's point of view. I haven't seen it mentioned but perhaps he tactics and the load outs need to be re-thought for IS. Having done some internal testing most of the posts here are anecdotal at best.

I propose looking at IS instead of just throwing around Clan nerfs. IS has gone through a lot of change in regards to weapons perhaps some tweaks should be added.

[Revis Volek]

Be Rough With Me Plz, on 08 August 2014 - 08:26 AM, said:

I agree to a point. When you're trying to determine something general like how effective Clan Mechs are against IS Mechs then the main thing to look at would be Win/Loss. In this sense the randomness in PUG matches will give a better idea since you're including every player in the Elo Spectrum and also including Trial Mechs. You're getting the entire picture. If you were to do that in a controlled environment then you're leaving out low Elo players and Trial Mechs. You're also adding in communication that doesn't normally exist in PUG matches.

If you were to talk about specific things like weapon effectiveness then you would have to do that test in a controlled environment where the majority of the variables are locked down.

But there is not Trail Clan mechs so IMO that jut gives the IS side a disadvantage out of the box thus skewing the results even more. I'm not saying nerfs to clans are unwarranted....i knew they were coming but this seems extreme. We will see though, my real concern is that it will be horrible and no one at PGI will care/want to change in back for fear of looking like another mistake on there part.

and still no concrete numbers....we are all just going off hearsay at this point. "Russ said this, so and so said that"....show me some DAMN DATA! I'm a nerd i crave it!

Edited by DarthRevis, 08 August 2014 - 08:41 AM.

[Mcgral18]

Hillslam, on 08 August 2014 - 08:32 AM, said:

To the OP: you'll never get an apology.

You'll get arguement, rationalizations, subterfuges, feints, strawmen, etc.

Its not the gear! Its never the gear! Its our skillz! It must be! Protect our gear at all costs so we can keep saying its our skillz!

Deploy the words!
Cast Aspersions!
Muddy The Water!

*data arrives*
Uh oh. Combat the data!



The Forum Lawyering goes good with a bag of popcorn, lightly salted, with a little parmesan cheese sprinkled on, a dash of garlic powder, and some white chedder.

*sits back and enjoys*

I asked Russ for some actual info, we'll see if we get any.

Individual Elo scores for each team should be a good indicator if Elo averages actually work.

Wanting some actual info is a good thing.

Sephlock, on 08 August 2014 - 08:28 AM, said:

How exactly is the Private match thing supposed to work? Are we just supposed to play in Private Matches on our own and trust that they will be counted? How much do we need to play to get whatever the "prize" is ?

I don't know anything about it yet. They said they'd release some more info later on. I'm debating on bringing the WubShee or not. If we get to choose Frozen City, I'll definitely take it.

[Be Rough With Me Plz]

DarthRevis, on 08 August 2014 - 08:37 AM, said:

But there is not Trail Clan mechs so IMO that jut gives the IS side a disadvantage out of the box thus skewing the results even more. I'm not saying nerfs to clans are unwarranted....i knew they were coming but this seems extreme. We will see though, my real concern is that it will be horrible and no one at PGI will care/want to change in back for fear of looking like another mistake on there part.

and still no concrete numbers....we are all just going off hearsay at this point. "Russ said this, so and so said that"....show me some DAMN DATA! I'm a nerd i crave it!

It only "skews" the results if the majority of the matches are being played in the Group Queue. We don't know in which Queue the majority of the matches are being played so you can't have "skewed" results. All we have are results.

[Revis Volek]

We dont have any results, we just have the aftermath sir...

Results would be data, conclusions and few decisions on where to go from here.

[Be Rough With Me Plz]

DarthRevis, on 08 August 2014 - 08:46 AM, said:

We dont have any results, we just have the aftermath sir...

Results would be data, conclusions and few decisions on where to go from here.

So just completely disregard the averages presented by PGI "because"?

I see.

Results are in PGI's database. The conclusion is that Clan teams won 90% of the matches. The decision is that changes need to be made. You know, everything said in the tweet...


I really don't understand how you can deny what PGI says without anything to support your denial. I understand the desire to view the data collected that lead to the averages presented, but without any proof of your own to use to refute their claims simply because you don't want to believe Clan Tech provides an advantage isn't logical.

Edited by Be Rough With Me Plz, 08 August 2014 - 08:53 AM.

[Hillslam]

Honestly, they don't need to share anything or ask us for our analysis. They should look at their own data and make the call, then live with the consequences.

(Although they are asking *some* players, hee hee)

Opening it up to the forum for design is like opening up the liquor cabinet to a bunch of drunks. All you'll get is trash.

[Faith McCarron]

Be Rough With Me Plz, on 08 August 2014 - 08:47 AM, said:

So just completely disregard the averages presented by PGI "because"?

I see.

Not just "because". Because it was gathered from a willy-nilly sampling that did not properly control for hardly ANY variable at all and draws conclusions that are unwarranted by the methodology used.

[Hillslam]

You have little to no idea how the data was collected. The numbers involved. Etc. All you have is your opinion and 3rd hand anecdotal stories and group think yes-manning.


But yeah, clearly we should leave the whole design up to the opinions of the neckbeard clan customers. They obviously have it all sorted...




...Forum lawyering *rollseyes*

Edited by Hillslam, 08 August 2014 - 08:53 AM.

[Apnu]

I must be playing on a different server from everybody else. I drive Hunchbacks and Jagers. My W/L is even and my KDR is 1.56 as of right now since clans came out. I average over 200 pts a game.

Last night on Alpine I had 2 kills, 7 assists, 722 damage, match score 97 in a HBK-4J w/ 2xLRM10s and 4xMLs and TAG. Our team won and we only had 3 clan mechs (WHK, TBR, NVA) against the other team possessing 4 clan mechs (WHK, TBR, SCR, KTX).

Over all, in the hunchies last night I went 4 for 6 and averaged over 300 pts per game.

[Oderint dum Metuant]

Because they are averages from a flawed test.

[Enigmos]

Did anyone consider that the vaunted ELO might be a bit wacky itself?

[GoManGo]

Your all on a island and I cannot hear you!!!!!!!!