Jump to content

Combat Team Metric Derivation


dipstik

Recommended Posts

Just wanted to start a thread to brainstorm how the combat team derives their metrics. a few thing stick out for myself:

 

they do not calculate/simulate hybrid builds unless we poiont them out

-"We never intend for hybrids to outperform the 36-point builds" taken from vanguard answers

however they also said "The 36 point builds should outperform the hybrids now." in the same response, implying that they determined the hybrid performance and according to their metrics, the full build should now do better

 

-they account for sub 30% and armor debuff in their derivation

"Please keep in mind that Pyrotech will perform considerably worse on a target dummy than it does in a real combat situation. There are two primary reasons for this. The most obvious reason is that the Burnout skill does not get its full effect on a target dummy, because a target dummy never falls below 30% health. The less obvious reason is because only around 48% of an Pyrotech’s damage ignores armor, while for Advanced Prototype this number is around 65%. This means that the Pyrotech has more to gain in situations where an armor debuff is present on the target, as should be the case in an operation."

"Only certain specs (Rage/Focus is not one of them, but Carnage/Combat is) can use execute abilities on a target dummy. It is also unlikely that you are testing on a target dummy with your own personal armor-debuffer (which you should definitely have against a real raid boss). So that 9% gap you may be seeing on target dummies is closer to approximately 4% in reality."

 

time to kill/die is a standard metric used for balance:

"These upgrades should make Vanguard damage dealers more competitive in PvP, with faster TTKs than before, and intrinsically more survivability because of that."

 

dps for any DD spec should be within 5% and have a certain amout of consistency

"We agree that Rampage/Zen Strike can fall on hard times due to random number generation. We are not opposed to fixing this issue by increasing the chance that Rampage/Zen Strike occurs, but it likely means that the rate limit will also need to increase."

Im not sure where the 5% quote is (buried on the forums i think). from what i have read, it is very common to use scripts to determine performance of builds, and from iterations of the script, or even output from one iteration, you can see the variance in some kind of rolling average. this could be used as a form of variance by which they try to avoid RNG effects of a build and how these affect the DD performance.

 

skill intensive classes may be deemed so by the complexity of their script, or proc dependence in the rotation.

"Sorcerer’s are an “easy to play, hard to master” class, in that most of their abilities are rather straight-forward and easily understood, but getting the most out of those abilities may require a higher skill level than many players are comfortable with."

 

-each ability has some metric for which they are shooting for that is an abstraction of the properties of the ability

"A scoring check reveals that, for its energy cost, it should deal more damage than it does, and we can fix that in a patch by properly rebalancing the ability. Hopefully, rebalancing Quick Shot/Overload Shot will make it into the competent filler attack that it was meant to be."

"We currently have no plans to add any extra Force management abilities, skills, or mechanics to Corruption or Madness, though we may consider lowering the cost of Static Barrier (but that would likely come at the price of lowering the absorption it provides). "

"Our metrics assume that Shoot First/Hidden Strike is only used once per encounter, so if you are using Disappearing Act/Cloaking Screen to perform additional Shoot First/Hidden Strike attacks in a single encounter, then you are probably performing above our expectations."

"It could very well be the case that the sustained damage gained by adding five energy to the Scoundrel’s/Operative’s resource pool is too small. We will look into this further, and we may at some point make a change to this particular set bonus if we find that we are not completely satisfied with its performance."

So they have metrics for encounters as well as scoring for each ability. the scoring seems to include damage/heal, cost, cooldown, cast-time etc. to produce a score. specifically, a ratio of cost to damage seems to have some balancing point they are shooting for. it looks like they never even tried to score the set bonus for scoundrels. Specifically, look at this quote:

"Because of how our ability damage is scored, XS Freighter Flyby/Orbital Strike puts out some of the best DPS in the game – assuming all of its damage ticks hit a target – and this is without the set bonus. While we do intend for XS Freighter Flyby/Orbital Strike to be an excellent choice against three or more enemy targets in an area, the single-target effectiveness of the ability is not intended to be so good, with a possible exception for the Saboteur/Engineering Gunslinger/Sniper."

So the scoring indicates that it gives good dps... maybe they mean the % damage it does in relation to the rest of the abilities in the encounter, or the damage/cost ratio (akin to damage per force/energy etc.)

So they talk about changing a few variables and getting a new score:

"First, we will be lowering the effectiveness of XS Freighter Flyby/Orbital Strike by reducing some combination of its cooldown, activation time, and energy cost. Then we will rescore the damage it does. "

So perhaps the score is just the % of overall damage with some relation to its cost.

 

 

-many of the survivability notions seem to be more visceral, such that time to die is not the cornerstone of this metric.

"Scoundrels/Operatives are meant to be shady, slippery characters that survive by escaping rather than sticking around to face the brunt of a foe’s attack. For Scoundrels/Operatives, both their ability to heal and their ability to stealth count toward their ability to survive. "

 

-here is a quote that provides evidence that they do not some sort of algorithm for determining max dps rotations (an algorithm to edit the script to find best possible script)

"That said, we are not completely satisfied with Lethality at the moment. We specifically don’t like that using Orbital Strike and Explosive Probe rotationally as a Lethality Sniper increases your single target damage, but that’s more of a problem with Orbital Strike and Explosive Probe than Lethality. "

sounds like they added orbital and explosive probe to the script, ran it, and it came out higher. it seems off to me that they can include explosive probe in their rotation but regular lethality folk cannot use explosive probe without popping adrenaline probe or target acquired. However this quote indicates that players may wish to forgo using series of shots to use explosive probe instead:

"Series of Shots and Snipe could probably use some boons to make them stronger abilities in the Lethality rotation – as it stands, they are in the rotation, but not strong enough to keep players from using Explosive Probe and Orbital Strike."

 

-They do end up getting feedback from actual players for some specializations:

"We are not completely satisfied with how Engineering currently plays in general. When compared to Lethality or Marksmanship, Engineering feels rather clunky. "

 

-for tanking they have stated that they use time to die as a metric of import. so this would require some kind of damage profile and perhaps even an event list which deploy unto the tank being analyzed.

 

-they may use encounter duration variation of their script/simulation/calc to find burst/sustain performance:

"As a result, the top damage over an extended period should go to Annihilation/Watchman. The top damage in a short fight should go to Rage/Focus. Carnage/Combat should hold its own in both situations, while not being the best at either – think of its strength as versatility."

 

It also seem slike they can alter the number of combatants in an encounter to determine aoe performance:

"While we do intend for XS Freighter Flyby/Orbital Strike to be an excellent choice against three or more enemy targets in an area, the single-target effectiveness of the ability is not intended to be so good, with a possible exception for the Saboteur/Engineering Gunslinger/Sniper."

"Rage/Focus is the top burst spec for a Marauder/Sentinel, and that is by design. Right now, it is also the top AoE damage spec for Marauders/Sentinels (by design) – but that is something we would like to get away from in the future."

Edited by dipstik
Link to comment
Share on other sites

I think you are right, in that many do not pay much attention to this group of threads. Not sure if i want to clone this thread just yet, in general or elsewhere.

 

i remember a dev saying that they check all changes in warzones, fps, leveling and operations. this is most likely QA running content, but Kitru found out that some new content was never tested with a shadow tank. It is also possible that tanks and healers have fp and operation scripts to measure performance. :

"Finally, the perception of a specific class being not desirable can also be affected by the desirability of other classes. For example, Sage/Inquisitor healers are currently able to exceed our intended healing performance at times by affecting multiple heals with the same Conveyance/Force Bending buff. Game Update 1.2 will remove the ability to do so."

"All specs for all roles have a target performance. This is what drives the balance of the game: soloing, Heroics, PvP, Flashpoints, Operations... everything. When those targets aren't hit, we can't just ‘bring everyone up’ to the highest performer without negatively impacting the balance of the game and creating unsustainable inflation in our combat system. "

"Based on the feedback brought to us so far from testers playing on PTS along with metrics and combat logs gathered from our guild testers, we are going to make additional adjustments before Game Update 1.2 is promoted to the live servers. "

"The fact of the matter is that Scrapper DPS is closer to target in 1.2. Understandably, if you were already struggling with an encounter, you may view this as an undesired change. For clarity, we don't agree that this is "unfair" for "pure PVE" players as the changes are not meant to (despite popular perception) specifically target PVP over PVE."

 

"(b) In response to your feedback, we've re-tested all classes to ensure they are falling within our desired DPS targets and found one issue with Demo Round / Heatseeker Missile which was getting increased DPS, not just from other people's stacks but also from an unwanted interaction with other skills. That issue was corrected in the 1.2.0c patch this week.

 

© There has been some level of inflation in the overall DPS budget of the game as result of Legacy benefits and other changes. Over time, these add up and as it stands, we have established that the global burst potential across the game is slightly higher than we are targeting. We will likely take some minor to moderate action about this in the near future by adjustments to the magnitude and duration of offensive relics (longer duration, reduced magnitude, identical power amortized over time)."

 

 

It seems like they use ratio of overall damage of an ability to shoot for targets:

"That is correct. However, I'll repeat, the reduction in Grav Round damage was made to address how much damage Grav Round contributes to your total rotation - not to address how many times you press the Grav Round button. Players have run off with the idea that we're trying to eliminate Grav Round spamming, which is not the case. My goal has been to eliminate how effective Grav Round spamming is if it's all you do.

 

An effective Gunnery rotation still utilizes Grav Round as both a "build up" ability and a "resource dump" ability. That hasn't changed, nor was it my goal to change that. What has changed is that competitive DPS now comes from utilizing all of your key rotational abilities instead of leaning so very heavily on just Grav Round. I hope that's more clear now."

"For clarity, the purpose of the change was to address the issue that Grav Round, on its own, was responsible for too much of your overall DPS. In other words, using totally made up numbers, if all you did was mindlessly spam Grav Round, you could still achieve something like 85% (again, totally made up) of your total damage dealing potential.

 

Not only was it bad, it wasn't fun (for the target or the caster). When things are bad, not fun, and yet they're still effective, they become high profile candidates for tweaking.

 

What we did was lower Grav Round's damage and increase Demo Round's damage. I don't know what the net result is off the top of my head, but we feel it's much, much closer to a "rebalance" than an outright "nerf."

 

And in case you're worried, if it turns out to be more of a nerf than a rebalance (which we don't want to be the case), then we'll fix it. But even then that'll probably be in ways that don't directly affect the damage dealt by Grav Round."

 

Here is a direct quote talking about a script:

"Hey guys, I'm not seeing this internally, but after looking at the script, I did notice that the Barrage buff is removed prematurely when you are NOT dual wielding."

 

In regards to hybrid targets:

"nakomaru: You have mentioned DPS, healing and tanking "targets" as objective methods of balancing the roles in the game. Are there hybrid targets (e.g. healing & dps spec) as well? It seems some hybrid specs are viable on certain classes and out of the question on others. I suppose this has to do with objectively balancing lower tier skills across the classes.

 

Austin Peckenpaugh (Senior Designer): Yes and no. In your example of a healer and DPS hybrid, by and large, whatever damage you sacrifice you should pick up in healing. For example, if you gave up 10% damage, we'd expect you to pick up 10% healing. Sometimes that's the case, but sometimes you lose or pick up synergies that move you above or below an even shift. Because "hybrid" in our skill trees means "virtually any combination of points," it's certainly not the case that we have targets for every combination. However, we do have expectations, and we're happy with hybrid builds so long as they don't radically stray from our expectations."

 

PvP Metric: they looks at win ratio and composition:

"For example, teams with all Sages and Guardians do not even appear in the top 100 win ratios and the composition with the highest win ratios include a fairly even distribution of Advanced Classes. That, of course, is not an absolute conclusion, especially considering full team compositions rely on matchmaking. We’ll pay close attention to the variance in composition win ratio considering rating when full team queuing is enabled."

 

This is probably my favorite quote found yet:

"Austin Peckenpaugh (Senior Designer): The damage boosts to Snipers and Gunslingers that you're referring to was actually a targeted weapon damage boost. We found an error in the assumptions we use for the mathematical interactions between weapon damage and defense chances. Without going into the boring specifics, we found that weapon damage (especially in abilities with high costs, high cooldowns, or high activation times) could be somewhat underappreciated. When we change our assumptions and come up with a new model like in this situation, we have two diametrically opposed goals: 1) to propagate all necessary changes to the game so that it's using our most recent math, which opposes the second goal 2) maintaining balance without blindly following our math. We altered the abilities that would pick up a significant enough change to be worth making without having a negative effect on the balance of the game. The list you're referring to is the result.

 

Class change prioritization is a bigger topic. Without getting too long-winded on the subject, what we're aiming to do with any class change is get a class or spec as close to target as we can. These "targets" are objective math-based goals that apply identically to all like-roles. By that I mean that all DPS specs have the same target because they are all the same role: damage dealing. Tank roles have their own targets, and healing roles have their own targets. When someone is off target (too high or too low), we try to find non-invasive ways of bringing them closer to target. We don't try to bring classes closer to target because we're mean - we do it because the global game math assumes everyone hits their target. When someone doesn't hit their target, it has an adverse effect on the game at large. And since rebalancing the entire game to accommodate an outlying spec is completely infeasible, we address the outlying performers with class changes, individually, as appropriate. It's an ongoing experience that we'll never be done with, so prioritization is based on what we feel is the most egregious in any given patch. How we identify outliers and how we determine which is the most egregious is... another bigger topic."

 

Here is a quote that speaks to skill level and meeting targets:

"Hargan: Will the Commando Gunnery spec be rebalanced after the changes in Game Update 1.2 so that we are back on par with the DPS output of similarly geared Sages/Gunslingers/Sentinels in Operations?

 

Austin: I think I'll confirm a lot of suspicions with this answer, but the bugfix that addressed Demolition Round scaling had a large enough impact on Gunnery DPS that it surprised us, too. Although Gunnery and Arsenal had been hitting our targets, it became harder to do so than we were comfortable with. Another way to say that is that the "low end" of our test results was hit too frequently by too many people. The changes you're going to see are mostly in resource management and usability, which will make it easier for you to deliver the considerable damage you already wield. We'll have more detailed information for you soon."

this suggests that they analyze histograms from testers to see how many are hitting what dps, and in this case, the low end had higher frequencies.

 

Here they show that the sum of passive plus active mitigation has a specific target:

"We are also striving to make all tanks hit the same survivability targets for Game Update 1.3. Testing shows that the self-healing generated by Shadow/Assassin tanks is too powerful after the armor adjustment they receive via Combat Technique/Dark Charge. This armor adjustment should have brought Shadow/Assassin tanks to lower passive survivability levels than the heavy armor tanks, with the self-healing they provide making up for the difference. However, this armor adjustment was making them passively just as good as the heavy armor tanks, with the self-healing taking them a bit beyond our survivability targets. Rather than hit armor or self-healing too hard, we’ve opted to adjust both by a much smaller amount."

this is the spike problem shadows have been talking about.

 

Here is a rough outline of the process and philosophy of class balance/design:

"Class Changes

I think it’s important to first communicate that not all class changes are meant to address balance. In fact, I try and treat usability, quality of life, and balance all as equal parts of good class design. To that end, when class changes come to a server near you, it doesn’t necessarily mean that your class was “too good” or “too bad” prior to the patch. In fact, it’s just as likely that these changes are meant to address kinks or even pains in your day-to-day play, ability functionality, and rotations.

 

However, many changes are made with the express purpose of changing balance. This is an ongoing process, but that doesn’t mean that we’re locked in a never-ending tug-of-war between the classes. In actuality, many components of classes are balanced against “hard targets.”

 

For example, some components of class balance have a “give and take” quality that is directly related to what other classes can do. If we lowered the cooldown of a Guardian’s Force Leap, the kiting and anti-kiting dynamic with ranged classes will have shifted. If, however, we increase the damage Vigilance and Focus Guardians deal, that moves them closer to or further away from our targets – the same targets shared by all DPS specs.

 

These targets are rigid in that all specs must hit them, but these same targets are also flexible that once a spec is within range of them, there’s some flexibility as to where that spec ultimately ends up. Is being 2% above target acceptable given how immobile you are? Is being 2% below target acceptable considering your off-healing and control capabilities? This is where our balance gets subjective and why, in spite of any mathematical accuracy, we’re constantly reevaluating class balance. It’s through this iterative process that we test our assumptions, using our own experiences and those of our players to identify the issues and, more importantly, identify non-invasive fixes."

 

Speaking to sin tank performance following 1.3, and hitting tank targets:

"Daitenzin: Does the team have any concerns about how the 1.3 changes will affect how Assassins/Shadows will perform as tanks compared to the other tanking options in an Operation situation?

 

Austin: No, not especially. I know a lot of players feel like this was a PvP fix that glossed over the idea of Operations survivability, but that just isn't the case. In reality, Shadow/Assassin tanks were slightly over-performing prior to Game Update 1.3, but some of that is obfuscated in the current climate of boss encounters. What some players have correctly identified as an issue is that some Operations bosses deal significantly more Force/tech damage than weapon damage, which favors tanks with high mitigation over those with high shield/avoidance.

 

Therefore, it'd be more accurate to say that I'm more concerned that we're currently overemphasizing armor and health pools (as the most valid channels of survivability) in Operations encounters. This may be the case because bosses aren't using enough weapon damage or because too many tank defenses only work against weapon damage. Not all channels of survivability can be balanced for every boss encounter, which is why tank survivability is measured against a norm, and why we're going to be pushing harder to hit that norm in the future - either through tweaks in boss damage or tweaks to the way shields work."

 

I found the 5% quote, when reffering to deception expectations.

"Deception should epitomize "hit-and-run" and "lone wolf" gameplay. Obviously that's less the case in Operations boss encounters, but if this is a question of sustained DPS, the short answer is that they hit within the same 5% "grace window" targeted by every DPS spec in the game."

 

Speaks to PvP versus PvE and how to balance for both (also sounds like the original heal to full in a few ways lol):

"Coldin: Every class who fights primarily in melee range (Assassins, Warriors, Powertechs) gets some kind of ability they can use to close the distance outside of 30 meters. That is except for Operatives. What's the reasoning behind the Operative's apparent lack of ability to quickly get within range to deal their most powerful attacks?

 

Austin: In PvP, the short answer is a combination of stealth (including Cloaking Screen), the ability to self-cleanse, and the ability to self-heal. Operatives also benefit from a good deal of control with snares, Sleep Dart, Debilitate, and Flash Bang, with additional roots, snares, and knockdowns available in skill trees. However, those things don't offer much benefit in boss fights that demand a high level of mobility and target switching. The question for us becomes one of how to best introduce a "fix" for a few encounters without dramatically impacting gameplay in other game modes or environments. That's a much longer answer, and something of an ongoing discussion. High mobility fights and encounters with a lot of target switching are proving to be fun ways for our Operations designers to challenge players. So with that theme keeping up, this issue is quickly floating to the top of our priority list."

 

this is funny: Why is Accuracy Back on Tank Armor? Senior Game Balance Designer Jason Attard talks about new gear and why accuracy is useful for tanks.

the link is broken: http://www.swtor.com/community/showthread.php?p=5155531#edit5155531

hah

 

This whole meet the developer blog seems very PvP centric. Peckenpauh seems more even handed when speaking about pve and pvp: http://www.swtor.com/blog/meet-developers-rob-hinkle

 

It seems that with 2.0, they changed the model by which DoT abilitiees are scored:

"DoT Improvements: We’ve changed the way we evaluate damage over time effects, so it should now be less costly to keep your DoTs up and to spread them around."

 

This is a good example of how they might shift around variables of an ability to reach target scores:

"Slow Time: Now has a 9-second cooldown (up from 7.5), costs 20 Force (down from 30), and deals slightly more damage."

 

Another quote that hints at how scoring and balancing is done:

"Changes to Smash: Smash hits for significantly less in 2.0, reducing Rage’s on demand burst. However, Smash is now responsible for less of Rage’s overall damage output, so these changes don’t negatively impact Rage’s sustained damage."

"Redistributed Damage: Damage has been redistributed from Smash into other abilities, reducing Smash’s contribution to your overall damage while increasing that of other key abilities."

Link to comment
Share on other sites

This is an outstanding thread, dipstik. I get a bit nostalgic reading over these quotes. Thoughts in no particular order…

 

Regarding HoT and DoT scoring (since they're one and the same). Judging by the changes they made (reduced cost on nearly all HoT and DoT effects), I think that they're now assuming a significant amount of clipping and/or overkill. The latter isn't an issue so much in PvE, but it's a pretty big problem in PvP. The changes have been pretty balanced in that realm (see: rise of the Lethality Sniper in warzones). Unfortunately, these changes have also had some pretty dire balance consequences. The most significantly example of this is the cost on Slow Release Medpack/Kolto Probe, which is obscenely low relative to its potential output. A skilled Scoundrel/Operative can achieve unbelievably high raid healing numbers just by being extremely tight on HoT management. This, combined with some other survivability changes in the healer trees for all classes except Commandos/Mercs, has made Scoundrels/Operatives kiting machines in PvP and HPS monsters in PvE. Hopefully they see this and readjust things soon.

 

I consistently get the impression that their tank "norm" is fairly naive. I think they're basically just balancing mean mitigation and time-averaged cooldown values. Guardians/Juggs have the best cooldowns by far if you time-average the effects, and they also have the worst mean mitigation on NiM bosses. We (as a community) may be overestimating the amount of analysis Bioware puts into the "spikiness vs efficiency" question, since if that were a primary design goal, vanguards/powertechs would have the worst mean mitigation across all content.

 

Their DPS balance philosophy seems to work along three axes: AoE, on-demand burst and sustained damage. Classes with long setup times and no appreciable AoE (*ahem* Watchman/Annihilation) should have far-and-away the best DPS when adjusted for raid buffs and executes. Classes with long setup but decent AoE (such as Tactics/AP and Balance/Madness) should be middle-of-the-road in terms of sustained damage. Classes with low setup times and decent-to-good AoE (basically, just Focus/Rage) should be significantly behind in terms of sustained damage. (incidentally, the best Focus PvE parse that I can find is 2770, which adjusts to 3102 on a boss; so Focus/Rage isn't as far behind as once was thought)

 

Bioware definitely pays attention to how a spec "feels" and how "fun" it is, but they seem to miss a lot of things. The Scrapper/Concealment filler period is a good example. Also, Balance Shadows/Madness Assassins feel both clunky and spammy. I wish they would spend a bit more time on this somewhat-subjective issue, because it has a pretty profound effect on player engagement and (more broadly) the class balance metagame.

Link to comment
Share on other sites

after thinking abouut it all for a bit i think i have some ideas:

 

scoring: i think they do something like fraction_time/fraction_damage , fraction_energy/fraction_damage etc. so when they change the cast time and cooldown, it gives them a new score, and as long as they tweak the values such that overall damage is within balance (aka 5% of target), then they are not being too "invasive"

 

i think they have goals for what fraction of damage comes from abilities x_1 through x_n for each spec, and if you use 20% of your energy on an ability that does 10% of your damage, that would be a bad score.

 

as a lethality sniper, i am still hung up on the "use explosive probe rotationaly" comment. unless you dont use Series of shots, how can u get explosive probe in the rotation without adrenaline probe or target aquired? im thinking of them tweaking the action script to something like:

 

higher priority stuffz

if energy >80, use explosive probe

if energy >70, use SoS

lower priority stuffz

 

and that got them a higher dps. if you ever played around with simulation craft action scripts, you know that minor changes will gives drastically different results.

 

-----

 

dot/hots: i think they were scoring it using dmg/cast_time at first and changed it to fraction_abilities/fraction_damage or something along those lines. things like mind crush have great damage per cast time, but the actual damage over time of the ability is rather lackluster.

 

part of me wants to start an Austin P. glossary:

 

ability functionality: life is more than numbers

aoe: who needs single target?

balance: prioritization is based on what we feel is the most egregious in any given patch. Is being 2% above target acceptable given how immobile you are? Is being 2% below target acceptable considering your off-healing and control capabilities?

burst: who needs sustained?

easy to play, hard to master: learn to play our script

hit-and-run: im sure someone else will kill them

lone wolf: if they dont die in 5 gcds, run away

perform/outperform: maths != button mash

quality of life: 0 deaths 30 kills

rate limit: WTB more unloads

rng: 30% chance to proc your rotation

rotations: action script

scoring: something over something that you are no where near

streaky: i know where to shove your flying fists

TTD: IM STILL ALIVE!

TTK: WHY ARENT THEY DEAD YET?

survivability: heal (your tanks) to full

usability: my mara can tank, just get another dps.

Link to comment
Share on other sites

this is funny: Why is Accuracy Back on Tank Armor? Senior Game Balance Designer Jason Attard talks about new gear and why accuracy is useful for tanks.

the link is broken: http://www.swtor.com/community/showthread.php?p=5155531#edit5155531

hah

 

:( I would have loved to read the reasoning behind that

(edit: now that doesn't let me rest I'm missing something really Important if I ignore Accuracy I knew it all along I'm just not able to find what I'm missing *verdammt*)

 

:D

 

thanks for gathering all that wisdom in one place.

Edited by DarthSpekulatius
Link to comment
Share on other sites

balance: prioritization is based on what we feel is the most egregious in any given patch. Is being 2% above target acceptable given how immobile you are? Is being 2% below target acceptable considering your off-healing and control capabilities?

I just want to know where that +-2% point is tbh, I mean if they say scrapper can go over it what does that mean for other classes?

Link to comment
Share on other sites

i remember when they said they were going to nerf sages so my combat medic would be more desirable. they made our AOE 4 targets instead of 3!

and then nerfed everything else into the ground! ):<

+3% inc healing with kolto residue (down from 5%), -5% inc damage with charged shield (down from 10%), only 1 cell generated with supercharge (down from 2), adding an ammo cost to trauma probe, reducing the ammo reduction from field triage from 2 to 1 while keeping it a 3-point skill (the absolute biggest change imo)

 

and then they gave us a battle rezz that was basline for scoundrels and sages (even dps). thanks, zoeller.

'we buffed your AOE and nerfed everything else, so stop complaining about being broken'

good time.

 

 

 

anyway, another comment. i have a guess that them not testing shadows with the new content was probably a result of shadows being so OP for so long and the fact that they still do the best in TTK metrics, so they just thought 'well shadows are still better, so i guess we should test the other tanks'

i also have a guess that kitru wasn't entirely honest about his interactions with the community members. the only takeaway i really got from his conversation was that he was really annoyed no one knew him by his forum handle.

 

 

the PvP quote reminds me of zoeller's stunlocking operatives comment from his interview with darthhater.

On the topic of nerfs, it’s obviously controversial if we decide we have to nerf something. I don’t expect people to throw us a parade. Ultimately, people are going to be very unhappy. Practical matter is you need a complete bird’s eye view of all eight classes — how they interplay with each other, how the entire health of the game is — to fully understand how things go. There is the small-minded approach of “so many people play Sorcerer, BioWare is never going to touch them.” No, that is not how it works. If we leave the Operative the ability to stun lock and kill people — yes, there aren’t many Operatives — but over the long term, that means people will quit the game cause it’s not fun. We have very measurable statistics that tell us if people lose a certain number of Warzones in a row being stun locked by a team of Operatives, then that might be part of that, and they will be not as likely to re-subscribe.
it also goes against this quote

"The fact of the matter is that Scrapper DPS is closer to target in 1.2. Understandably, if you were already struggling with an encounter, you may view this as an undesired change. For clarity, we don't agree that this is "unfair" for "pure PVE" players as the changes are not meant to (despite popular perception) specifically target PVP over PVE."

 

 

 

there's another quote too, right after, where he talks about balance philosophies

The stun lock was part of it but we actually identified the Operative at launch as problematic, and we had to watch it. We had to wait at least until we had data to make sure it was actually a problem. It’s very hard to balance a game on theory. We like to take a measured approach, watch what is actually happening in the game, and not meddle with you every week. Because everyone will have a feature of the week and, in many cases, the counter to that develops organically in the community in no time. There’s no reason for us to interfere, but if we see something that warrants interference, then we will have to act. 1.2 is one of those cases where we can go across the board, we can look at things, and we can rebalance.

 

 

and here's another 5% quote

I do reject that there is a “well, Operatives don’t have the sustained damage in Operations.” They do. They are within 5% of a Marauder.

 

 

here's the whole interview http://www.darthhater.com/articles/swtor-news/19979-guild-summit-interview-with-georg-zoeller

Link to comment
Share on other sites

:( I would have loved to read the reasoning behind that

(edit: now that doesn't let me rest I'm missing something really Important if I ignore Accuracy I knew it all along I'm just not able to find what I'm missing *verdammt*)

 

:D

 

thanks for gathering all that wisdom in one place.

 

Quick google search gave me this: http://www.bbcguild.com/showthread.php?4920-2012-09-13-DevTracker-AllisonBerryman-Why-is-Accuracy-Back-on-Tank-Armor

 

Basically it was saying 10% accuracy is good for not just DD's, but also tanks, because of threat generation. However, if people wanted more mitigation by default instead of more threat then they would be open to swapping it back, and until then they can still swap accuracy enhancements for shield ones. Bear in mind this was back when shields only absorbed weapon damage, and a majority of raid bosses damage came from force/tech attacks, so shield was useless there (most of the time) as well.

Edited by TACeMossie
Link to comment
Share on other sites

good catch.

 

i really wish we still had simulation craft (korrig updated it for sage dps, but thats the only one). that program allowed us to play around with action scripts and see if inserting one ability over another made a difference, and would also provide energy use and damage fraction statistics, even show typical energy levels during the fight.

 

unless we get a tool that is capable of analyzing the specs in a manner similar to that of the devs, we are basically just going off of parses, which have huge variance even for the same user, with the same gear, with the same rotation.

 

specifically, i want to look at the fraction_energy/fraction_damage of overload shot. of the value is greater than 1, then it costs too much energy and should have more damage or less energy. if it less than one, that could be seen as overperforming. when they talk about orbital stroke being one of the best abilities in the game i think they are looking at fraction_time/fraction_damage , fraction_energy/fraction_damage and seeing values of less than 1.

Link to comment
Share on other sites

  • 4 months later...

good thread. I gotta admit, some of these quotes just make me shake my head. There is something to be said for metrics, statistics and modeling, but it should never be solely relied upon. Makes me wonder if they even play the game. I think the player base could tell them a lot more about class balance than their precious metric could ever hope to.

It should be a requirement for the combat design team to play in a good pve raiding guild and a good pvp guild anonymously. Talk to their guild and other players and actually learn what's going on in the game.

Link to comment
Share on other sites

  • 5 years later...

thought this should go in here:

 

from http://www.swtor.com/community/showthread.php?t=522210

 

in regards to accuracy of boss abilities:

 

I asked Senior Game Balance Designer Jason Attard about the Accuracy on attacks from Operation bosses, and he let me know that the default boss accuracy is 100% for special attacks and 90% on basic weapon attacks. Any accuracy debuffs applied to the boss (or defense buffs applied to yourself) should reduce the boss's chance to hit. We do have the ability to increase accuracy dramatically for a special attack if we want to guarantee that it hits, but typically that only occurs for abilities that are critical to fight mechanics.

Link to comment
Share on other sites

so you want me to start a new thread for each post? thats not how forums werk

 

here is a good one in regard to how gear stats are generated:

 

I talked to Jason Attard (Senior Game Balance Designer) and he was able to shed some light on modifications in endgame items and how stats are distributed.

 

Our goal is that a highly-skilled player can complete content without needing to swap mods on their gear, but spending the time and credits necessary to optimize the mods should still make a noticeable difference.

 

When we generate mods for endgame gear, we generate all the mod variants at a given item level and then run a process that combines the mods with endgame shell items to create the items needed for an Operation. This process helps us ensure that the stats are useful for Advanced Classes and roles, but it doesn't generate gear that is completely optimal for every spec. We always manually check the results and tweak items where necessary.

 

We have had some problems in the past with gear that stacks a single rating too much and with incorrect armorings being assigned to some items (high Endurance gear should go to tanking items, Aim/Cunning/Strength/Willpower should go into DPS and healing items), and we're continuing to refine the processes we use to make more useful gear. In future tiers, you'll see many issues with mod allocation corrected!

Edited by dipstik
Link to comment
Share on other sites

×
×
  • Create New...