26 replies to this topic

[kesmai]

For that endeavor Paul needs a new dartboard.
We need to donate!

[C E Dwyer]

Pointless to try I'm afraid.

Quirks don't make bad mechs good unless they are big heat generation ones or rate of fire, coupled with a moderate torso twist.something that P.G.I have stated that they are getting rid of, and reducing for more mobility and structure quirks.

Secondary to that are structure and crit chance quirks but they have to be pretty big like the atlas, or are worthless.

You will never see many Vindicators in normal play unless they get buffed to superman levels of power, because they are two slow, with low hard points

Another example of quirks doing nothing for a mech is the Pheonix Hawk a few feeble mobility and structure quirks and not a lot else the Blackjacks with its shape and high hard points is slowly being brought down to level mediocre but until then it remains the only meta choice, with the Cicada a distant second place for IS light mediums

[PhoenixFire55]

Requiemking, on 01 August 2016 - 01:37 PM, said:

Did I do a good job, or am I horribly wrong. Feel free to leave some feedback, and suggestions of your own.

Your suggestion is logical and probably the only reasonable course of action. However PGI and logic doesn't come in the same sentence. In order to balance the game they need to put effort into it, they need to play their own game on sufficient enough skill level and they just won't and can't. They are too lazy to address the pressing balancing issues for several years now, even tho changing quirks is as easy as changing one numerical value in an xml file, which could easily be done every patch if needed. They've recently made a rescale pass that was supposed to balance out mechs, but for each mech they've "fixed" this way there were 5 more they "fubared" even further, simply because their "logic" behind the rescale was broken to begin with.

Do not expect them to care about balance. They care about selling mechs, thats why certain mechs are sometimes suddenly quirked into OP only to be dequirked into trash shortly after enough people waste MC/c-bills on them (Dragon is a great example if you remember what it was like during the 50% AC5 cooldown quirk). Some Hero mechs also magically have quirks and/or hardpoints superior to their non-MC peers.

[Narcissistic Martyr]

Chuck Jager, on 01 August 2016 - 06:21 PM, said:

Also lights had to grow in size so that they would take the damage aimed at them. The agility was as much of a find the baseline plus the fact that the Jenner IIc has an enormous alpha. So instead of double nerfing the Jenner they met in the middle and will hopefully rebuff the others that are truly sub par. Remember sub par is below average not doing as well as the top1-3.

Thing is... the size increase really didn't affect my IIC very much. It still hits for 72 damage and at point blank range (where I typically fire) I will 1 or 2 shot pretty much any mech in the game from behind. If I don't give the enemy a chance to react... well me being bigger doesn't matter so much.

The Pew Pew versions... they suffered quite a bit but again they're better ridge humpers than knife fighters anyways due to the high mounted lasers.

I'm afraid this is very much a case of PGI ignoring the light mech pilots in the player base telling PGI exactly what we intended to do with the IIC (72 damage alpha to the butt) and giving it missile hard points to sell it. 3 would have been sufficient. It'd hit a little harder than the oxide and be a bit more durable due to the clan XL and have JJ, but it'd be far less accurate.

[Bishop Steiner]

Bilbo, on 01 August 2016 - 01:41 PM, said:

We started with no quirks, data was gathered, and we ended up with what we have now. What, exactly, makes you think repeating the process will get us to someplace different?

The Great Rescale had shifted the dynamic precipitously since. So a redo is sensible at this point. Whether or not it well be remotely well done, is of course a different story. But to do nothing is still worse

[Mawai]

You seem to forget that the purpose of quirks is to help balance IS vs clans.

IS generally has quirks while the clans generally don't (at least not much).

Removing all quirks simply moves the clans back to the OP relative to IS mechs (which some argue they have never left).

Removing quirks is not a solution ... especially not for months of automatically unbalanced game play while "data gathering" is performed.

No matter which way the problem is looked at ... it is challenging to come up with a balance for so many different mechs.

There are three separate balance problems that are intertwined which is part of the problem ...
1) Clan vs IS balance
2) Weight class balance (how should lights, mediums, heavies and assaults stack up against each other in general).
3) Balance between mechs of the same weight class or tonnage

Personally, I think the clan vs IS balance should be initially addressed by giving all IS mechs structure quirks. Perhaps start at 25% more structure across the board for all IS mechs. The point of this is to try to compensate for the longer range and higher alpha capabilities of clan weapons which tend to make them generally more effective. The value should be based on analysis of damage done by clan mechs and adjusted on a percentage basis to make IS more durable.

After this, quirks can be used to even out mechs in each weight class ... there are a variety of quirks that could be used.
- speed bonus quirks
- agility quirks
- weapon quirks
- armor quirks
- additional module slots

These need to add character, differentiate variants to make them more interesting, and make underperforming mechs a bit more effective.

However, the fundamental problem to mech balance are the outliers ... particularly clan mechs ... the Arctic Cheetah, Stormcrow, TImberwolf, Dire Wolf and Kodiak are particularly effective due to clan firepower and/or clan XL engines allowing for great maneuverability while carrying significant weapon loadouts. (As an example, look at the light mech leader boards ... the scores for the Arctic Cheetah are 10% or more higher than any other light mech ... presumably those scores are based on direct ingame performance ... demonstrating that the Arctic Cheetah [ and to some extent the Adder, Jenner IIC and Locust] are the most effective light mechs).

Positive outliers are the mechs players love ... and they are the ones that players howl about the most if they are nerfed since players tend to love the mechs that play the best. Everyone knows that it isn't impossible to defeat any of those clan mechs, especially on a team, but the leaderboard result seems to indicate that the Arctic Cheetah has at least a 10% and more like 20% effectiveness advantage over any other light (the 75th place arctic cheetah player's score is 20% higher (about 2400) than the similar score for any other light (about 2000 max with some as low as 1500) ... looking at the bottom end of the list gets around the issue of pilot skill vs mech since I don't think anyone can claim that every arctic cheetah pilot is simply better than the pilots of any other mech ... whereas if you only look at the top of the list it is easy to make that argument).

Anyway, PGI has trouble nerfing positive outliers due to the negative player reaction ... so they try to buff opposing mechs to compensate to some extent. The problem is that they can never buff all the rest to reach the level of most of these outliers.

Keep in mind that PGI can have ALL the performance metrics for every mech, every day, similar to the leaderboard numbers. That data can be mined to extract relative mech balance RIGHT NOW. You don't need to remove quirks to gather data ... PGI already HAS immense amounts of data.

The problem is easily shown in the light mech leaderboard numbers ... some of the mechs are generally 20 to 50% more effective that the others.

Here are the average scores for each light mech from the leaderboard ...


Mech AverageScore Tonnage Score/ton
"Arctic Cheetah" 2793 30 93
"Locust" 2429 20 121
"Adder" 2364 35 68
"Raven" 2319 35 66
"Jenner IIC" 2304 35 66
"Jenner" 2284 35 65
"Spider" 2126 30 71
"Kit Fox" 2080 30 69
"Wolf Hound" 2071 35 59
"Firestarter" 2060 35 59
"Panther" 1986 35 57
"Commando" 1868 25 75
"Urban Mech" 1860 30 62
"Mist Lynx" 1846 25 74

From this we can see what everyone more or less knows.

- the arctic cheetah (despite recent nerfs) is STILL the best light mech ... 15 to 20% better than the next mech.
- the locust is still better than many lights despite its fragility. Likely due to hit detection issues on fast targets.

- the first tier of lights consists of the Arctic Cheetah
- the second tier are the locust, adder, raven, jenner and jenner iic ... with the jenners at the lower end of the range
- the third tier are spider, kit fox, wolfhound, firestarter and panther
- the fourth tier are commando, myst lynx and urban mech

There aren't really any surprises in that list .. except that the locust is still far too effective for a 20 ton mech.

Anyway, that data gives you a rough idea of what needs to be done to balance those mechs ... no giant quirk removal and requirkening is required. What IS required, is the politcal will on the part of PGI to nerf the outliers. If you look at the score/ton ... there are two mechs that are extreme outliers ... the arctic cheetah and the locust. The rest of the mechs are actuallly almost balanced based on a score/ton evaluation.

Edited by Mawai, 02 August 2016 - 11:48 AM.

[jper4]

yet interestingly going by score per ton it's the commando and mist lynx that hold the #3 and #4 slots despite their bottom of the group average scores.

it's almost like they are in lore urbies, where they are the cheap, quick production lights designed to be more efficient in short bursts, but aren't going to last long enough to do serious overall damage in a long drawn out fight. i'm probably wording this badly though...