88 replies to this topic

[Valhallan]

/shrug considering how the tuk stats have often been misrepresented too, hardly surprising. Needs more Statistics 101.

[Scout Derek]

Gyrok, on 19 June 2017 - 12:49 PM, said:

So, interestingly...

The MAL, VND, WHK, DWF, EBJ, BNC, AS7, ENF, STK, DRG, WHM, TBT and MDD are all better than MAD-IIC and NTG.

[Gyrok]

vandalhooch, on 19 June 2017 - 01:57 PM, said:

My favorite Gyrok post was during CW phase I.

He wrote this massive screed about IS being OP and PGI favoring IS over clans. The very first response post in the thread was just three words, "Shut up Gyrok." Made me laugh over and over again for days. I still smile now, thinking about it years later.

My favorite Gyrok post was the wolf humping the bear meme after we repelled MS from invading 2 of our worlds in a single NA primetime CW session while they were in Ghostbear...I still smile now, thinking about it years later.

As a side note, the player who said that (shut up Gyrok) was later banned for cheating.

I_AM_ZUUL, on 19 June 2017 - 02:12 PM, said:

Dude... I have to congratulate you on being such an amazing troll! there is no way that you can mean any of this because of what a giant piece of $hit that would make you just as human being, you are the best troll on the forum.

Should have known my biggest fan would come post with an ancient poorly designed meme.

You should change your sig to "The guy that lets Gyrok control his life".

I_AM_ZUUL, on 19 June 2017 - 02:18 PM, said:

Gyrok is the literal Dartboard of Balance that Russ uses... #confirmed

HL3 confirmed.

El Bandito, on 19 June 2017 - 05:37 PM, said:

Congrats Gyrok, you just played yourself.


Not that it is surprising.

My #2 biggest fan...I am not surprised in the least...

You should get a sig made with "Gyrok Fan Club" in it...

Edited by Gyrok, 19 June 2017 - 06:48 PM.

[VanillaG]

Bud Crue, on 19 June 2017 - 02:14 PM, said:

And yet, this is exactly how PGI balances as well. Think about it.
There are mechs that rarely get played, compared to others, yes?
Yet PGI must inevitably use the same sort of comparative...however skewed they may be...data sets when they make balance decisions.

Except PGI has the stats for every match, mech, and weapon used in all of the games. Just because you don't see specific mechs being played doesn't mean that they aren't being played. You can argue that PGI might not be able to interpret the data correctly but PGI has enough data that it is all statistically relevant unlike the OPs data from a small slice of games played.

[Deathlike]

VanillaG, on 19 June 2017 - 06:57 PM, said:

Except PGI has the stats for every match, mech, and weapon used in all of the games. Just because you don't see specific mechs being played doesn't mean that they aren't being played. You can argue that PGI might not be able to interpret the data correctly but PGI has enough data that it is all statistically relevant unlike the OPs data from a small slice of games played.

You can make some legit assertions about Night Gyrs, HBK-IICs, MAD-IICs and a few other mechs listed there to make a good decision off of.

[Gyrok]

VanillaG, on 19 June 2017 - 06:57 PM, said:

Except PGI has the stats for every match, mech, and weapon used in all of the games. Just because you don't see specific mechs being played doesn't mean that they aren't being played. You can argue that PGI might not be able to interpret the data correctly but PGI has enough data that it is all statistically relevant unlike the OPs data from a small slice of games played.

Yep...all those Tier 5 potatoes and their kitchen sink builds, with mouse set to 8000 dpi and in game sensitivity set to 1 with acceleration turned up all the way. You know...the ones you see running LRM mechs with no back up weapons, or laser builds with one of every kind on them firing from 700 meters (when more than half their weapons do not reach that far)...those guys are the players dictating balance for all of us, because they play the yolo queue more than anyone else...

Do you really want that determining what is strong and weak in this game? FFS, they gave the MAL-MX90 LRM quirks...I mean really...who the hell thought that was a great plan? Novakaine?

Seriously...consider what you are saying. Then consider that you also have to be gathering the correct data, and then you must also have someone who knows how to interpret that data.

Now realize that PGI made a bass fishing game that Bass Pro Shops said was crap.

I can see your face as you read this post:

[VanillaG]

Gyrok, on 19 June 2017 - 07:23 PM, said:

Yep...all those Tier 5 potatoes and their kitchen sink builds, with mouse set to 8000 dpi and in game sensitivity set to 1 with acceleration turned up all the way. You know...the ones you see running LRM mechs with no back up weapons, or laser builds with one of every kind on them firing from 700 meters (when more than half their weapons do not reach that far)...those guys are the players dictating balance for all of us, because they play the yolo queue more than anyone else...

Do you really want that determining what is strong and weak in this game? FFS, they gave the MAL-MX90 LRM quirks...I mean really...who the hell thought that was a great plan? Novakaine?

Seriously...consider what you are saying. Then consider that you also have to be gathering the correct data, and then you must also have someone who knows how to interpret that data.

Now realize that PGI made a bass fishing game that Bass Pro Shops said was crap.

I can see your face as you read this post:

If you do want you are talking about you are just balancing by "feels" because the data is the data. There are all kinds of things that you can do control for the various factors that you mentioned but the reality is that for every top tier comp player there are 10 or 100 potato tier players that put in more money combined than that one top tier player. You can argue that PGI might not be looking at the data correctly but they absolutely have the best picture of what is going on in the game.

[vandalhooch]

Gyrok, on 19 June 2017 - 07:23 PM, said:

Seriously...consider what you are saying. Then consider that you also have to be gathering the correct data, and then you must also have someone who knows how to interpret that data.

Irony meter just pegged out.

[Luminis]

KIT FOX 100% WIN RATE! KIT FOX OP PGI NERF PL0X!!!

[Scout Derek]

vandalhooch, on 19 June 2017 - 07:48 PM, said:

Irony meter just pegged out.

[Ultimax]

1st: Gyrok's interpretation of the stats aren't accurate.

2nd: There are a good number of mechs with a solid amount of drops that are very close to the performance, or better than, to the soon to be nerfed NTG & MAD-IIC (and KDK for that matter).

GRH, WHM, WHK, BLR etc.

They don't have as many drops = not as universally map/strategy friendly
They have good performance = do as well or better in the right circumstances

Edited by Ultimax, 19 June 2017 - 08:18 PM.

[LORD ORION]

NGR and TBR lol, no doubt the Gauss/Peep versions *eye roooooolllllooooll*

[pyrocomp]

Bud Crue, on 19 June 2017 - 02:14 PM, said:

And yet, this is exactly how PGI balances as well. Think about it.
There are mechs that rarely get played, compared to others, yes?
Yet PGI must inevitably use the same sort of comparative...however skewed they may be...data sets when they make balance decisions.

The crucial difference is that PGI uses same kind of tables that are padded with the significantly higher statistics. And the comment was about that. First, each mech is used, however rarely, but is used (I sometimes see IS Victors and Orions). Second, all scorn PGI for low pace of a decisions process. Sometimes they need half a year to make a conclusion. And sometimes that is the time to gather stats.
So, to be more literal, the comment was not about wrongly selected metrics, but about an attempt to compare chassis using data from 1 or 2 drops.
Total number of drops in that table is less than 1000. That's well below an hourly data sample PGI can get from their servers. And you probably seen numbers from Tyk.
So, yes, it's Gyrok. Should I add you here?

[pyrocomp]

Deathlike, on 19 June 2017 - 07:12 PM, said:

You can make some legit assertions about Night Gyrs, HBK-IICs, MAD-IICs and a few other mechs listed there to make a good decision off of.

To be able to do so the stats should show mech performance across different tiers (yes, no selfproclaimed elitism) on different maps and a general metrics for the loadout diversity (however you can do that). Right now even NTG has less than 100 per map. And an unknown number of real pilots in those drops. Really, some are good with lights, some with heavies.

[Deathlike]

pyrocomp, on 19 June 2017 - 09:53 PM, said:

To be able to do so the stats should show mech performance across different tiers (yes, no selfproclaimed elitism) on different maps and a general metrics for the loadout diversity (however you can do that). Right now even NTG has less than 100 per map. And an unknown number of real pilots in those drops. Really, some are good with lights, some with heavies.

The thing is, if a player and/or team used a particular mech as what is defined as "obvious duh" like in the case of the Night Gyr, it tells everyone that has a brain exactly what their preference is regardless of derp.

Like, we obviously can't tell much about the Vindicator about the lone 1 drop it was used it (you'd have to dig up context based on the match it was used in).. but if you surely could not tell why a Night Gyr is used in DivA MRBC (top tier) comp matches... then it means you're not very observant. Remember that MRBC drops are localized to lobbies and not pub play... and if you even had the fortune of dropping against any large groups of them in the pub queue, well... you'd be lucky if they weren't picking a super-must-get-goto meta mech.

You have to remember that public play that people that don't know any better will generate more random mechs, and the more "educated" about how mechs are popular and/or used, then they are more likely to use said mechs and gravitate to the better options than the mediocre/sub-par options that are available in MWO. If Tiers were actually "proper" measurements of actual skill, you'd find the really good players dictating what mechs are "the thing" pretty easily than whatever mass fail that is PSR and Tier 1.

I guess what it comes down to is this... do you want to listen what a comp player tells you (minus thread starter), or do you not care? I can give you all the answers you'd want... but the question is that do you believe what you're being told at all?

Edited by Deathlike, 19 June 2017 - 10:04 PM.

[Tarogato]

Really, the only stat I would look at for comp like this would be the number of times a mech is used. Because really, teams/players want to use the best mech they possibly can. Especially in Div A, you won't see a mech used unless somebody thought it was the best option for a certain role. And the more a mech shows up, the better it is at that role, or the more common and influential that role is.

Damage numbers and KDR mean little, because for instance people will ignore a MLX but hunt down a MAD-IIC, thus the MLX might perform better then the MAD-IIC because the latter has a big target over its head "heyo, big fatto powerful mech here"

Can't even trust WLR that much, because many mechs have a low sample size, you'd have to correct for that, and it would be an artificial benchmark.

So just look at the number of drops, that tells us a lot.

[PhoenixFire55]

What actually caught my attention ... Warhawks ... 1 kill, 48 assists ... srsly fkn NA div.A potatoes, learn to aim.

[Bud Crue]

VanillaG, on 19 June 2017 - 06:57 PM, said:

Except PGI has the stats for every match, mech, and weapon used in all of the games. Just because you don't see specific mechs being played doesn't mean that they aren't being played. You can argue that PGI might not be able to interpret the data correctly but PGI has enough data that it is all statistically relevant unlike the OPs data from a small slice of games played.

Consider the Victor for the three years it sat untouched. Yet PGI stuck to their guns that it was fine. How could they have possibly concluded that? For THREE YEARS. My guess is that they didn't look at frequency of play in their analysis. They looked only at its average damage, average this, average that. Its stats probably were within the mean of their target values based on the few hundred matches a month that they were looking at. That did not mean it was a good mech, but rather a few dedicated (probably good) players were playing it. As compared to the Dire, or Kodiak were EVERYBODY was playing it (see leaderboards throughout 2016) It wasn't until last December (when they finally hired new balance folks?) that suddenly changes were made to encourge Vics to actually get played again. I think that is because they finally started looking at frequency of play.

But even now, this is still a factor. If a mechs is determined to be crap by the players, its data set gets skewed due to relative infrequency of play compared to other mechs that are being played. All it takes is a few good players to play that crap mech exceptionally and now that mechs data set is also exceptional, because it is so limited. There is nothing PGI can do to change that except to quirk it in some manner and hope that the change gets it more play so that they can get a better data set thereby. If those quirks are not enough, they keep dumping em on until the lack of player interest finally breaks and the mech gets played. See for example the current state of the Dragon 1C.

Edited by Bud Crue, 20 June 2017 - 03:08 AM.

[pyrocomp]

Deathlike, on 19 June 2017 - 09:59 PM, said:

The thing is, if a player and/or team used a particular mech as what is defined as "obvious duh" like in the case of the Night Gyr, it tells everyone that has a brain exactly what their preference is regardless of derp.

Like, we obviously can't tell much about the Vindicator about the lone 1 drop it was used it (you'd have to dig up context based on the match it was used in).. but if you surely could not tell why a Night Gyr is used in DivA MRBC (top tier) comp matches... then it means you're not very observant. Remember that MRBC drops are localized to lobbies and not pub play... and if you even had the fortune of dropping against any large groups of them in the pub queue, well... you'd be lucky if they weren't picking a super-must-get-goto meta mech.

You have to remember that public play that people that don't know any better will generate more random mechs, and the more "educated" about how mechs are popular and/or used, then they are more likely to use said mechs and gravitate to the better options than the mediocre/sub-par options that are available in MWO. If Tiers were actually "proper" measurements of actual skill, you'd find the really good players dictating what mechs are "the thing" pretty easily than whatever mass fail that is PSR and Tier 1.

I guess what it comes down to is this... do you want to listen what a comp player tells you (minus thread starter), or do you not care? I can give you all the answers you'd want... but the question is that do you believe what you're being told at all?

See the point about elitism in the post you quoted. I've seen enough games chocked down by elitist. That group, however good at their abilities to read games mechanics and metrics is too small to sustain any game at minimally acceptable income (just enough to keep those servers running). So, while I'll listen to those who name themselves `comp.players', I'll take that with a significant doubt in regards to what is good for the game and it's future (and from the experience of encountering those in the GQ I'll doubt their undestanding of what this game really is). For the comp.scene the Solaris is a must and the only target. For the rest the maps, modes, PvM and coop.campaing. Completely different things to the point of mutual exclusion.
The only (yes, only) common attribute of each and every game is that it should be `easy to understand, hard to master'. But that is beyound the point of discussion in this thread.

And turning back to the original idea... You see, the comp.players are prone to others opinion and the some common understanding of what a meta is or how the game should be played. At some point I bought a DWF. Struggled with it a bit and came up with a build that I was comfortable with. That was 2xcGR + 6xcMLs. And it was almost half a year before that became a meta. And I checked, no changes for those weapon systems were made. Meta is a.. let's say this way, a todays fashion. Really, it changes with no more than a whim or belief. And it takes almost two month to settle.

So NTGs are heaviliy used. They are more suitable for the playstyle that is percieved as meta (right now that is a long range trade from staionary positions). The meta will get more spread, the players will get acustomed to it and the meta will shift as you will be more effective in other builds that exploit the habbits of playing with previous meta in mind (like a switch to the midrage fast skirmish and a NASCAR as it's simpliest form). The meta has it's seasons and cycles. So are the Mechs usages. As an example, the KDK-3 were hit with a mobility-engine decoupling. The percieved changes were that the KDK-3 was not nerfed but obliterated. And KDKs disppeared from the battlefield completely. Recently, in past two or so weeks the KDKs are back. You see them in matches and so on. And not in a random or joke builds. In reasonable builds. And the patch was not even here. So no, from the mech usage alone by a small players sample in a very narrow time window no reliable conclusion can be made. If we had graphed stats from 4000 players for a year - then there was basis for some conlusions. This is a main point. Any reasonable usage statistics cannot span less than a few months.

[El Bandito]

Comp players are the fastest people to find out and use "the latest effective mech/build/tactic", cause they tryhard the most. Common puggers cant even compare. Therefore the most popular mech/build etc they use are naturally the best.