[FatherCrixius]

A synopsis of GR2's impact on the game. Ty to all the homies who proof read this for me.

Some thoughts and feelings on GR2. 

Evade Viability 


I have mixed opinions on post gr2 evade. Its existence induces a skill-tax that needs to be taught to newer players due to it's obscurity and reduces the available class pool to 'Only things with hit buffs lest you deal 60 - 100% less damage forever. It is also very misleading as to how it works.

Currently, the game performs two-hit checks. One initially, another if the first misses.

Hitting the first gives you no damage loss, hitting the second gives a heavy 60% DR. 

In practice, this means you have a (HitChance) of doing full damage, a (HitChance * (1 - HitChance)) of doing 40% Damage, and (1 - HitChance)^2 of doing 0% damage. For those of you who get invited to parties:

Starting with a 50% To hit. 25% To glance, And a 25% to miss completely. (0.6 Damage or 40% Damage reduction)

A 30% to hit gives you 21% to glance and 49% to miss outright. (0.384 damage or 61.6% Damage reduction)

And a 10% to hit gives you 9% to glance and an 81% to miss outright. (0.136 damage or 86.4% Damage reduction)

The difference of 20 hit increases evasion from 40 - 61.6%, with 1000 HP you would go from 1666 Effective HP at 50% Hit chance to 2600 Effective hit points at 30% hit chance, and 7352 effective HP at 10%. (Hence the need for 250+ hit)

For people who really don't get invited to parties, here is a spreadsheet of 250 hit vs evade values. 



So what does this mean?

1. For a starter evade is numerically dogshit unless you are Giga-hardcapped or if the other guy does not have good hit.
   
2. You are not likely to see too many Evaded! procs, but the Damage reduction is there. 
   
3. Going below 250 hit is suicide without buffs.
   
4. Evasion dr is not linear. 
   
5. The +50 Cap should be hit as quickly as possible.
   

Mechanically in fights, this forces the 99% of classes that do not have active evasion debuffs (Soldiers flatfoot for example) to simply diceroll every hit check, and for the few classes that have the means to reduce evasion to do so every time it's off cooldown. This is good in that it forces gameplay changes, but in it's current state I feel hit checks should be more conditional with more classes having access to such. Such statuses should also be weaker, as flatfoot is a 15% max evasion nerf, reducing almost 32 evade from a 210 evade build. I feel this system could work much better with a little balancing and implementation. TLDR, The system has promise, but needs to be insanely fine-tuned balance-wise; and needs evasion / hit to be a more active system than 3M Buff -> +X Evasion, more systems like flatfoot and flanking bonuses to justify hit buffs rather than passives would make the mechanic interactable.

Weapon Accuracy - Many skills still do not use weapon accuracy, particularly demon hunters' matador, giving a base 90% when most weapons are pushing 100 + parts. In conjunction with the aforementioned more evade more better point, makes a lot of skills outright unusable vs evasion builds.

Racial Bases - Base stats matter a LOT more now. Average 4 Skill vs Nagas base 8 Cel is CRINGE. You cannot run any low skill (4 or below) without hit buffs from your classes, where higher base skill races are offered more freedom.

Conclusion - The system being as unintuitive as it leads to a very unhealthy mechanic, giving players very polarised opinions on the stat. I believe the system could be cleaned up a lot without gutting it by removing the glance check, and buffing weapons hit by 10 across the board to make up for evasion being buffed. This would keep the damage reduction the same, but make your displayed hit your effective hit. Some changes to general class or gear hit/evade buffs would be nice as well, as only a select few classes have access to enough evasion buffs to hit the cap, and a fair few classes lack hit buffs entirely forcing a black knight subclass crutch. I feel it's current state is misleading as being below the +50 evasion buff leads to players feeling like evasion is MORE useless than GR1 evasion, and being low hit feels like you are shoehorning your build or auto losing vs a relatively inexpensive (But very restrictive) defensive stat.



Class Viability

As of right now, BK is THE subclass of all time.

• Fortitude (Soldier)
  
• 24 Crit evade
  
• 10% Physical DR (20% Feared)
  
• 6 Defence
  
• 15 Hit (Parry)
  
• -15 CE (Feared)
  
• 15 Hit (FF or Fear)
  
• Parry Reduction
  
• Steel Series (Aura)
  
• Board Shaker 
  
• Crescent Rook (V. Good Scalings)
  
• Checkmate (Autohit)
  
• Hanging (Hit bonus)
  
• Forced Move
  

I am of the opinion that BK is a fine class that synergises with itself and has utility. That utility is it's strength. I am referring to it here as a problem because every other class is not BK. Many classes, I feel, need to be brought up to BK. This is most clear with it's hit buffs, as many classes post GR2 legitimately are left without, and without a hit steroid subclass. Some setups outright need BK, Bonder, etc in order to hit a minimum amount of hit (Which in current meta is around 230, but 250 is far more reasonable.) This, in turn, really damages build variety as many people are shoehorned into picking more in-meta classes to be able to damage a percentage of the playerbase -> Dodge is an unequal system and the burden is on classes to remedy this.

Many classes either lack hit or evasion buffs, causing BIS subclasses to be required should you want to do either. For evasion, ST / VA / BOX / SS / BON are almost required to hit the +50 cap reliably, even if you include gear such as windswept gi. With hard-capped SKI and Blessed (Current Meta Enchant), a weapon can be expected to hit 240~ hit, 220 without, where 250 is considered to be a standard. It is, even then, not ideal numerically. Evasion typically reaches 155 - 160 without buffs and gear, and 205 - 210 with buffs. The +50 system only seems to restrict class variety, for both parties. I feel like this could be remedied either by 

• Lowering evasion / hit buffs across the bored 
  
• Giving evasion / hit buffs to far more classes
  
• Reducing the evasion cap and base hit by an equal amount
  

Classes with built in free damage (such as Firebird, Ghost and Bonder) are favoured in this meta due to the SKI TAX, but this will be covered shortly.


Weapon Scalings 

Separate from GR2 as a whole, the weapon scalings update I feel was a mixed bag. On the one hand, GR1 weapon variety was very stagnant, and this change definitely changed what was being used, but I do not think the issue leading to stagnation has changed-- Rather that the current pool is temporarily cleaner. For contrast, GR1 was largely defined by Rebelling 1* weapons for Str based autohitters, 10* or 9* weapons for basic hitters (Largely due to their generally higher Acc / Crit / Power) and Avalon Dynamic tomes for mages. In current GR2 meta, 1* weapons, Rebelling enchants, and virtually 90% of the weapons are outright outclassed by 10*s -- Be it due to higher Accuracy, Power, Scalings, Effects or etcetera, leading to weapons such as kingslayer being 'The best sword' for almost every circumstance. Other weapon types like fists and more uniquely scaling weapons like eternal solitude have suffered the most, fists originally having higher scalings to make up for their inability to be two-handed and Eternal solitudes multiple wack scaling stats to enable shit like VABK Dag / Shield. 

I am not opposed to the greater shift towards 10* equipment, I actually enjoy the sentiment, but as it stands 10*s are not rare. Most weapons below 10* are directly outclassed by the 10* of their respective category. Contrasting an Anchor Edge or a Spine leash (Both basic hitting weapons) to almost any other axe is simply not a competition, In between anchor edges 5% Higher hit, 10% higher base crit, or it's +8 base power on the average axe. Spine leashes' range, 10 base crit or bonus to ice attack. Trying to stat any fist build without a hands of the giant will give you like 115 - 120 swa in an era where Axes can hit 150 pre-claret with relative ease, and now in ski-tax era affording less numerically efficient stats like Strength specifically tends to cost you elsewhere in the build. 

TLDR -- Build variety hampered by lack of diversity in useful equipment, Balance affected by standardization of scalings not accomodating for differences that existed in GR1, sudden importance of hit etc. More forms of 10*'s (Promising from what we've seen of Kovara) or just better existing equipment would be the remedy to this.

This has caused a drift from direct swa scalings towards additive bonuses like claret, crit, and elemental attack scalings (Non-mage) from things such as crescent rook and lift off- Or most heinously substituting damage through summoner. Back when these were implimented, Elemental attack was generally less gouged. Will was less common as were bonuses to elemental attack, but now every class' elemental attack seems to scale better than their raw swa scaling attacks (See Crescent, Asura's fist especially, Lift Off, etc.) 

Attacks that utilise 100 120 scalings, IE 'better spells' as the community sometimes lovingly call them, tends to 55% SWA based 45% elemental in damage. Lift off for example, with 140 swa and 90 Ele atk will end up with a 248 raw, whereas a falcon strike will only net a 198 raw, making physical damage and swa gouging in general far less rewarding. 140% Swa scaling is very standard, and partly why you see physical autohits such as demon hunter's reaver so rarely nowadays -- It's just not viable to hit the 270 swa required to break even with lift off if damage is your goal. Not only that, but physical damage is more mitigateable (Due to armour stat availability, General frequency of % Physical DR's in classes, etc) than elemental attacks unless you are using that elements specific resistance gear.

TLDR - Moves that do not utilise split scalings are falling behind due to numbers creep and general lowering of swa. I believe this could be remedied by either 

• Raising the base SWA % Scalings on non-split moves by 15 - 20%, with exceptions to moves that have multiple sources of damage (Firebird, Ghost etc) and additional focus on typically underperforming damage classes (Such as Demon Hunters Reaver, Monks non-utility skills)
  
• Adjusting the 120 100 skills and allowing SWA to take a larger portion of the scaling. Skills such as BK's Crescent rook or Boxer's Asura's fist completely outperform the majority of their kit by just having higher raw numbers. Lowering each of these by 40%~ and raising the swa scaling on them by 20% would promote SWA as a damage stat, rather than the tangential elemental attack which is typically raised to 70 - 80+ on builds that utilise them.
  

The last point I have to make on scalings is the tragic loss of Eternal Solitude's identity, the scalings being the most interesting part of the weapon in that it was a dagger that needed to be built around. The loss of SWA from fist weapons, as they have maintained identical scalings to sword weapons despite being unable to be two-handed, combined with the general inability to gouge strength without losing all hit / defensive stats have rendered them 25 - 35 SWA below what they used to be, and upwards of 30 swa below two-handable martial weapons due to part itemization. I believe this, in combination with the fact that MA classes are very GR1 Autohit heavy combined with the SKI TAX, has lead to the classes decline in popularity and capacity.


SKI TAX

I have mixed feelings on this too. As a consequence of evasion's changes, skill is now a mandatory stat on 99% of builds -- There simply aren't enough autohits to warrant going lower than 60 scaled: The consequence of this however is every build is effectively 15 Levels down in terms of stat points. This predominantly affects swa based 'autohitters' of GR1 the most, as basic hitters build skill regardless and mages have the stat points available to invest without compromise. This has had the greater effect of reducing most peoples defensive stats by 10 - 15 each, the abandonment of luck and san as stats and Blessed (enchant + talent) to become the staple weapon enchant-- Granting 20~ hit to virtually every build. 

The loss of san in favour of skill and more primary defensive stats has also limited san races greatly, usually only to classes that already utilise stat-less damage (Akin to old-time doomwalls)  or stat-san benefiting races such as the Instinct Kaelenesians and wyverntouched. I don't... really know what to do about this. Or how to react to San being relegated despite being one of the most thematic stats.

I believe the main consequence is classes building skill-- Hexer for example benefit outwardly from every setup being 60 stat points lower on average, which is partially why they are considerably stronger post GR 2 (As with the Guard / DOT change). 

Classes that utilise statless damage; Summoner / RM etc, Are also in a far better spot than previously low ski classes, being able to focus on defences that some class combinations can simply not afford anymore. Summoner is particularly egregious right now as every third person is a Bonder / Something, as while Seiryuu has shitty hit, everyone who has not built to evade has had to skimp out on resistance and defense, and can be increased through synchro summon and fight as one, on top of 60 evasion ignoring damage per target per turn.

I don't particularly know how to fix this, without sweeping nerfs across the board for overperforming classes, as this new system is opposed to most forms of autohit. Frankly, I would just be happy to see less bonders [1 , 2].


RIPOSTE 

Most moves are now riposte-able, some are more riposte-able than others and should be bug fixed I feel, but the fact that skill became mandatory and virtually everything triggers it, Armor of nails, Fang Faced shield etc warrants some damage nerfs on said retaliation sources I feel.


Turn Order

Riposte critting, or any status that ends the enemy's turn but does not end yours causes said enemy to go first. This is basically abusable but is generally considered to be an unfortunate coincidence. Summoners dying cause their minions to two turn the other party, and boxer's geist allows you to send a high initiative boxer to the bottom of the ladder. I definitely feel the new system requires more rigidity, as the current meta and frequency of two turns is unfun to try and workaround.

[Lolzytripd]

This is why I wanted a logged curve for evade vs hit formula, that way no matter what the "actual" numbers players would build their hit and evade at you could forcibly correct it the intended balance point by weighting the log either way....

As for swa scalings in single vs multi stat scaling, that's sadly always going to be a victim of the stat diminishing return caps. It ramps up too harshly in my opinion which punishes the use of single stat scalings.

[Fern]

Turn Order

Riposte critting, or any status that ends the enemy's turn but does not end yours causes said enemy to go first. This is basically abusable but is generally considered to be an unfortunate coincidence. Summoners dying cause their minions to two turn the other party, and boxer's geist allows you to send a high initiative boxer to the bottom of the ladder. I definitely feel the new system requires more rigidity, as the current meta and frequency of two turns is unfun to try and workaround.

To elaborate a bit more on what's being talked about here, here's an example that's more common than you'd think in a 1v1 scenario:

1. Someone with Riposte goes first and normally swings at their enemy, does not land critical hits.

2. Enemy's turn, they swing at the Riposte user a couple times. Riposte finally goes off and lands a critical hit.

3. However, it got that critical hit when the enemy just finished their turn, meaning the Riposte user is the last turn in this round. They're left with 1M (or 2M if they have Fleur) and usually that means ending turn.

4. Next round begins, and the enemy goes first, essentially having a "double turn" because the Riposte user had a (basically) useless turn previous round by sheer coincidence of life.
 
Thing is, this doesn't just happen with Riposte, it happens with anything that can make you gain momentum mid-fight. Sometimes people will unequip or just completely avoid things that give you momentum in that manner for the reason that it can be extremely detrimental.
 
On the other hand, being able to gain momentum mid-fight is a cool extra mechanic and entirely disabling that would just be a sad solution to that particular problem IMO. Maybe there should be a minimum momentum requirement for a turn to "count" in the round robin, so minor momentum gains like that don't suddenly get you killed? I'm not sure, but that's not the main point of this highly informative thread, so I won't go further with that train of thought.

---

I'm still of the opinion that the current game state is too polarizing to the point it's not fun. My problem with the SKI tax is not "I can't disregard hit entirely and still do good damage to everybody", it's "I need high hit for my character to have any meaningful impact in a fight". It's a huge part of why I preferred the idea of autohits not being generally full-dodgeable, because then you could just make Evasion! a high damage reduction against autohits while potentially still having it ignore status effect inflictions. At least then the person with low hit rate is at disadvantage but not entirely useless in the battle, and not feeling -obligated- to be stacking hit rate.
 
Because as it stands, the game makes you feel forced to get as much hit as possible by penalizing you for not doing it in a manner that's not so much "fair" as "too harsh" if that makes sense. This means that there's less incentive to go out and try other setups that could be more thematic or, I guess, "more original" for your character, and there's more encouragement for doing cookie-cutter meta. The bread and butter setups, because they're the ones that are considered to "work".
 
And generally the builds that are considered to "work" in people's eyes (to my knowledge) are the ones that can have at least some meaningful impact in a fight, some kind of usefulness, rather than being doomed to missing every hit and not contributing anything whatsoever to the battle or even the roleplay itself.
 
So I still think we should go back to having a damage reduction for "dodging" autohits, but making it fairly high (i.e 45% or so) and keeping the ability to dodge status inflictions. The second part alone, in addition to still being able to fully dodge basic attacks would still give Evade its own identity rather than just being alternate tanking, because after all tanking doesn't just let you "dodge" a game-changer poison when you get lucky (obviously not accounting for builds that stack up status resist.) 
   

TL;DR

I think the penalty for not building enough hit rate is too harsh to the point it gets very polarizing and makes you barely able to contribute meaningful things to a fight or even in the roleplay of it in some cases. I believe evade should go back to being a damage reduction (preferably somewhere around 40-45%) against autohits while keeping the ability to dodge statuses, which will help it keep a different identity from just normal tanking. Glancing Blows is a system that sounds cool on paper, but in practice skewers things too much in one side's favor and in a way that detracts from the general fun of the game, since you'll see people doing the same things over and over to -really- try and avoid the part of it that's considered very unfun.
 
But that's just my opinion/two cents. I'll take anything else at this point! I appreciate all the work into showing the math behind things, it helps put a more 'objective' and informative look on things!

[InsainArcaneBirdbrain]

Alright, since I have a lot of my own opinions and such on the subject matter here as someone who spends an unhealthy amount toying with this game's systems, I'll go in order based on the original post's main points. Just remember these are opinions, as is every discussion on this sensitive topic, as mine might differ from yours.

Evade Viability 

The numbers being as skewed as they are is my primary reason overall for my distaste for the new evade system. Outside numbers retuning on existing hit/evade buffs, very little has been done to the classes in a way to balance out the hit/evade race. While even back in GR1 people gravitated towards certain classes for whatever was meta at the time, GR2 has been something more of an extreme case. This is due to how the Hit/Eva imbalance allows properly supported kits such as FB, BK, and Bonder to name a few the ability to have more free-form stats, SKI not being as important. I like to personally dub these classes "Hit" classes, or classes that provide the proper hit support needed to function in such a system comfortably whether that be hit numbers, or evasion ignores.

As stated in the topic post, this creates a system where a large majority of classes can do nothing but roll (often rigged) dice, there's not much interaction going on here. The wild variance in these values also create a game that quickly spirals into a completely one sided matchup, in the fully buffed dodgy's favor if the opponent has sub-240 hit, or in the attacker's advantage if they have 250 hit or more. Speaking of that, the numbers presented in the table assume FULL EVADE BUFFS, and even then, you're never reaching more than 215-220 evade with just buffs alone, you need debuffs to go past that (and there's even less support for that in this game currently). Meanwhile, most hit buffs are PASSIVE, meaning there's no real momentum upkeep cost. Flatfoot is a 0-1m application, blessed/bloodhunt are free, class hit passives are free, etc. Most evade buffs however are a 3 momentum investment in most cases for a 3-5 turn buff typically. Other evade buffs such as Miragewalk and Unarmored Prowess can be applied passively, but only last a single attack or require select classes or circumstances to be used.

What this means is that a dodge build will always face a momentum tax in order to "play the game", and even at that highest potential, they still lose to a hit stacker (Which is just about everyone in the current meta). This along with the sheer investment needed in CEL in the first place makes it very limiting to build for. Using the above numbers, let's assume a fully buffed dodge averages 210 evade, and the average attacker is using 250 hit. This nets roughly a 50% net DR according to the math, which seems fine at a glance (pun totally intended), but in practice results in a net loss for the dodge build as they have to often spend 3-6 momentum every 5 turns to keep that up vs the passive nature of that 250 hit. In many build experiments I've done over the past few months to get a taste for these numbers in practice, the amount of investment needed is often straight up not worth building over DEF/RES in many cases (For PvP). Before I get lynched for saying this, hear me out:

Dodge in its current incarnation is simply a "stat check" in a sense that it just effectively automatically wins against setups unable to reach the hit numbers needed to win, or have adequate amounts of evasion ignores to deal with it. It's a concept I've seen in other games as a "win more" stat in the sense it only really enables you to bully builds that would have been likely ill prepared to deal with you in the first case. The major exception to this trend is Youkai, who bully tank setups quite heavily due to the sheer momentum advantage summoners hold. I cite this very momentum economy as to why dodge feels so miserable to play, you lose a lot of momentum and stat investment for something that often straight up doesn't matter in many matchups. A well stacked armor tank (it's own issue of balance) matches or outperforms on the DR front in nearly every normal situation using the average 250 hit this meta sees, as no dodge will ever get above 215 evade without niche shenanigans. This is further a issue when CEL requires nearly the same investment as a DEF/RES user stat-wise (If not more!) in order to have a DR. Evade is unreliable to the point where running absolutely no DR option results in oftentimes superior results, as at least you can then try your hand at completely winning the damage race instead of hoping your evade gamble pays off in a system rigged against it. 

TLDR: Current class balance is ill suited for this kind of system, leading to both dodge players to be upset at how their choice of DR is largely meaningless much of the time, and for everyone else because dodge if not kept in check is so utterly broken that you HAVE to play the hit/evade game, ruining build diversity as a whole.


Weapon Accuracy

Being stuck a 90 base hit is indeed a death sentence in such a system in many cases against a dodge, especially when you factor in that you also lose your hit enchant for said attacks. This creates not a 10-20 hit deficit, but a 40+ one. With the provided math in the topic post, it's not hard to see why this is often damning. 


Racial Bases

Not a terribly large issue compared to other points, but it can definitely make or break some setups due to scaling, usually for dodge builds. Largely a consequence of buff caps, as it means base values from stats and gear are even more important than ever.


Class Viability

If you read my thoughts on the evade viability section, you'll probably know what to expect from me here. Ideally, EVERY class should be a hit/evade class, or enough in a sense that every class has at least some means of playing the hit/evade game. I do agree that BK is actually a good benchmark of what a class should have in terms of mechanical support for such a thing. If class hit/evade balance gets fixed, we'd immediately start seeing classes other than the same five or six being used.


Weapon Scalings 

Weapon scaling changes are one of the reasons I don't call GR2 entirely bad. The generalization of scaling along with general nerfs to overall scalings was a good change that makes building a lot simpler. The main cause of concern here though is the lack of weapon identity, since base weapon stats and effects are so much more important now than scalings as topic creator pointed out. This makes it so many weapons are still the immediate go-to in many cases (mostly 7*+s), such as the Anchor for axes, since it's a case of "Oh, this axe has extra range and more crit while having the same stats as nearly every other axe, no reason not to use this". Logically, makes sense if rarity is to = power, but personally I find it boring. Rebelling used to be the main thing to justify lower rarity weapons, but in GR2 even after buffs, it's an absolute joke of an enchant (You not only give up a hit enchant, but you actively LOSE hit). The solution to this isn't so clear honestly, but I think a nice start is to make the no-effect weapons more desirable stat-wise (more hit/crit/etc) compared to weapons with effects as to make them even considered. 

As for some slightly related personal thoughts on weapons, STR weapons could stand to see slightly more crit mod built into weapons as to not be completely overshadowed basic attack-wise by guns/daggers. Also stressed this before in several balance-fus, but bows all being 100-70% STR is AAAAAAAAH in combination with the other system changes made. Personally would like to see bows make use of some of the other scaling modifiers if STR is to stay as to give some diversity, like the element + dextria combos tomes see a good amount of. Heck, that could be a interesting way to make the bland weapons in other weapon types more interesting too.


SKI Tax

Good idea in concept, icky in practice largely due to the prior evade issues. Stat taxing is just like taxes in real life to a degree, it doesn't do much to affect the rich (BKs, Bonders, etc), but severely punishes the poor (DH, Pure Lightning Crit Mages, etc). This only serves to widen the gap between the meta picks from the other classes in this system, as having more options in building is more fun to people. SKI taxes discourage exotic setups on non-meta classes simply because it means you can't afford other key stats otherwise. SKI is not an option, it's build in high amounts or die to the first dodge build you see. SAN builds have consequently fell off the face of the planet as stated, only really being made use of if said racials give stats to make up for it....  or you're playing a "hit" class that can carry one statwise-  hence my rich/poor tax comparison. 

On the topic of taxes, luck as a stat is kind of one of these now, for basic attackers/lightning critters I mean. The removal of hit/evade was very necessary to avoid a double universal tax, but the stat as a whole now feels very weak for those still forced to build it. It gives 1 crit and 1 crit eva, and that's it for actual battle application. Crit is cheesable by other means (Hexer says hi), so people will often rely on those external mods and minimize the luck they can build. Crit evade is nice, but FAI completely and utterly outclasses Luck here as I pointed out in another balance-fu post, so you will almost never see a non crit build use Luck. 


Riposte

Turn order and riposte with the round robin system has always been a bit screwy. This situation only really often applies with youkai/bots due to them always moving last, and builds with high initiative (man, dodge can't catch a break).

Gonna go off on a side point I didn't see brought up that this circumstance reminded me of, and that is how initiative is often detrimental rather than helpful. Riposte on a speedy char puts you at risk of getting double turned, which is why many dodge duelist unequip the skill. Being fast also means that many threats can't be responded to until it's too late. Two big dangers dodge builds face constantly is the evasion ignoring Explosion and Dancing Water. Both these skills are too effective at smacking dodge because the dodge build in question CAN'T REACT TO IT. They just have to eat the full damage.... with no DR... which then stacks said opponent's bloodhunt hit. Another point against dodge, which is silly because slower tanks can react and counterplay these things.

[FatherCrixius]

This is why I wanted a logged curve for evade vs hit formula, that way no matter what the "actual" numbers players would build their hit and evade at you could forcibly correct it the intended balance point by weighting the log either way....

As for swa scalings in single vs multi stat scaling, that's sadly always going to be a victim of the stat diminishing return caps. It ramps up too harshly in my opinion which punishes the use of single stat scalings.

Logged curve would be far more base but very awkward to explain, and potentially even harder to balance around hit/evade buffs. (But far less snowbally)

I do not feel that is the case on your second point, on the grounds that swa gouging has historically been fine in the past even with relatively similar swa's, but in current meta GR2 you will not have the points available nor the ratios we originally had to make such sources meaningful, barring additive/math sources like claret. Matter of fact even with 100 base strength (Near 210 swa @ 100% str scaling) landing a 140% standard autohit will net you 294 damage, this 40% SWA scaling contributes only 84 damage -- Equal to 70 fire attack on a 1.2x scale. With 100 str you would have nearly double that amount of fire attack. Numerically, these auto hits are STRONGER than the physical ones, even with post-hard cap eventmin stats. The differences only become more apparent with lower SWA values.

To elaborate a bit more on what's being talked about here, here's an example that's more common than you'd think in a 1v1 scenario:

1. Someone with Riposte goes first and normally swings at their enemy, does not land critical hits.

2. Enemy's turn, they swing at the Riposte user a couple times. Riposte finally goes off and lands a critical hit.

3. However, it got that critical hit when the enemy just finished their turn, meaning the Riposte user is the last turn in this round. They're left with 1M (or 2M if they have Fleur) and usually that means ending turn.

4. Next round begins, and the enemy goes first, essentially having a "double turn" because the Riposte user had a (basically) useless turn previous round by sheer coincidence of life.
 
Thing is, this doesn't just happen with Riposte, it happens with anything that can make you gain momentum mid-fight. Sometimes people will unequip or just completely avoid things that give you momentum in that manner for the reason that it can be extremely detrimental.
 
On the other hand, being able to gain momentum mid-fight is a cool extra mechanic and entirely disabling that would just be a sad solution to that particular problem IMO. Maybe there should be a minimum momentum requirement for a turn to "count" in the round robin, so minor momentum gains like that don't suddenly get you killed? I'm not sure, but that's not the main point of this highly informative thread, so I won't go further with that train of thought.

---

I'm still of the opinion that the current game state is too polarizing to the point it's not fun. My problem with the SKI tax is not "I can't disregard hit entirely and still do good damage to everybody", it's "I need high hit for my character to have any meaningful impact in a fight". It's a huge part of why I preferred the idea of autohits not being generally full-dodgeable, because then you could just make Evasion! a high damage reduction against autohits while potentially still having it ignore status effect inflictions. At least then the person with low hit rate is at disadvantage but not entirely useless in the battle, and not feeling -obligated- to be stacking hit rate.
 
Because as it stands, the game makes you feel forced to get as much hit as possible by penalizing you for not doing it in a manner that's not so much "fair" as "too harsh" if that makes sense. This means that there's less incentive to go out and try other setups that could be more thematic or, I guess, "more original" for your character, and there's more encouragement for doing cookie-cutter meta. The bread and butter setups, because they're the ones that are considered to "work".
 
And generally the builds that are considered to "work" in people's eyes (to my knowledge) are the ones that can have at least some meaningful impact in a fight, some kind of usefulness, rather than being doomed to missing every hit and not contributing anything whatsoever to the battle or even the roleplay itself.
 
So I still think we should go back to having a damage reduction for "dodging" autohits, but making it fairly high (i.e 45% or so) and keeping the ability to dodge status inflictions. The second part alone, in addition to still being able to fully dodge basic attacks would still give Evade its own identity rather than just being alternate tanking, because after all tanking doesn't just let you "dodge" a game-changer poison when you get lucky (obviously not accounting for builds that stack up status resist.) 
   

TL;DR

I think the penalty for not building enough hit rate is too harsh to the point it gets very polarizing and makes you barely able to contribute meaningful things to a fight or even in the roleplay of it in some cases. I believe evade should go back to being a damage reduction (preferably somewhere around 40-45%) against autohits while keeping the ability to dodge statuses, which will help it keep a different identity from just normal tanking. Glancing Blows is a system that sounds cool on paper, but in practice skewers things too much in one side's favor and in a way that detracts from the general fun of the game, since you'll see people doing the same things over and over to -really- try and avoid the part of it that's considered very unfun.
 
But that's just my opinion/two cents. I'll take anything else at this point! I appreciate all the work into showing the math behind things, it helps put a more 'objective' and informative look on things!

Fern

Muy Bien, thank you for doing the full explanation on how double turning functions like that, it was something I probably should've gone into further detail on but didn't want to drag out an already long and mechanic heavy post -- Especially with something people complain about so frequently.

One of the things I was worried about when making a post critiquing GR2 evasion in some ways was that I do not make basic hit builds. Those classes have never been fun for me, and so most of the points I could make felt tilted by my own perspective on the mechanic, the inclusion of goose patented spreadsheets™️ was more to display it's functionality as an alternate tank stat, it's cost effectiveness and how hard you are punished for being just 20 hit below a viable build, which is typically one class passive hitbuff plus a few. It was also to display polarisation of effective DR to hit's dodged -- In that evading! an attack is rare but numerically functional as DR, so even with a 50% + Effective DR you are still more likely to be hit than to evade the attack outright, leading to the misconception that the stat is underperforming. (40 % to hit, 36% to dodge, 24% to glance) I have not fully formulated a sentiment for or against the system currently, but I see it's weaknesses and hope it would be improved before other things are tried.

The issue GR1 Had regarding increased evasion DR (From 30) was that cel was a relatively cheap stat to hit the bare minimum of evade on, evasion being calculated linearly caused dodge-tanks and stat-walls in general to outmath the other guy were they not to build hit, and usually beat basic hit builds by virtue of superior numbers (IE, mage.) and I believe this would still be the case now. If people no longer need to pay the Ski tax, I think a lot would need to change, the issue would be of classes being too strong rather than some anemic kensei being too damage resistant if that makes sense.

Alright, since I have a lot of my own opinions and such on the subject matter here as someone who spends an unhealthy amount toying with this game's systems, I'll go in order based on the original post's main points. Just remember these are opinions, as is every discussion on this sensitive topic, as mine might differ from yours.

Evade Viability 

The numbers being as skewed as they are is my primary reason overall for my distaste for the new evade system. Outside numbers retuning on existing hit/evade buffs, very little has been done to the classes in a way to balance out the hit/evade race. While even back in GR1 people gravitated towards certain classes for whatever was meta at the time, GR2 has been something more of an extreme case. This is due to how the Hit/Eva imbalance allows properly supported kits such as FB, BK, and Bonder to name a few the ability to have more free-form stats, SKI not being as important. I like to personally dub these classes "Hit" classes, or classes that provide the proper hit support needed to function in such a system comfortably whether that be hit numbers, or evasion ignores.

Zerg (chad) P1

100%. 

As stated in the topic post, this creates a system where a large majority of classes can do nothing but roll (often rigged) dice, there's not much interaction going on here. The wild variance in these values also create a game that quickly spirals into a completely one sided matchup, in the fully buffed dodgy's favor if the opponent has sub-240 hit, or in the attacker's advantage if they have 250 hit or more. Speaking of that, the numbers presented in the table assume FULL EVADE BUFFS, and even then, you're never reaching more than 215-220 evade with just buffs alone, you need debuffs to go past that (and there's even less support for that in this game currently). Meanwhile, most hit buffs are PASSIVE, meaning there's no real momentum upkeep cost. Flatfoot is a 0-1m application, blessed/bloodhunt are free, class hit passives are free, etc. Most evade buffs however are a 3 momentum investment in most cases for a 3-5 turn buff typically. Other evade buffs such as Miragewalk and Unarmored Prowess can be applied passively, but only last a single attack or require select classes or circumstances to be used.

What this means is that a dodge build will always face a momentum tax in order to "play the game", and even at that highest potential, they still lose to a hit stacker (Which is just about everyone in the current meta). This along with the sheer investment needed in CEL in the first place makes it very limiting to build for. Using the above numbers, let's assume a fully buffed dodge averages 210 evade, and the average attacker is using 250 hit. This nets roughly a 50% net DR according to the math, which seems fine at a glance (pun totally intended), but in practice results in a net loss for the dodge build as they have to often spend 3-6 momentum every 5 turns to keep that up vs the passive nature of that 250 hit. In many build experiments I've done over the past few months to get a taste for these numbers in practice, the amount of investment needed is often straight up not worth building over DEF/RES in many cases (For PvP).  []

P2

One thing I didn't cover was the 'ease of buff' privilege some classes get versus others. Priority one is hitting your +50, I do this through hunters wind to proc Windswept -> Chaos Reflex -> Feather of Mercury for +48 in 3 momentum. Other classes like Rogue end up with several 2 - 3m skills they need to use and upkeep in order to not be made of paper, whereas one bash from a BK or proc of a stalemate will counteract 3M of the dodge's buffs. The upper limit on a regular setup is around 210, 215 if rogue main class and hardcapped, and as you said that's where the hit debuffs come in. Fear / Locking guard / Ken passives are all additive here, but do not reach the evasion cap, allowing for some disgusting stacking. I do not feel this is an issue of viability but most evasion buffs are straight up bad, and the class pool of viability is like you said incredibly niche.

I do not feel cel is expensive as a stat, however. Being able to maintain 40 or less hit rewards more DR and the oppertunity to glance than Cel split into defence or res respectively, of course without the blessings of our lord and saviour Armour. In my own pvp experience, the biggest counter is me mispositioning and getting hanging'd onto or being foolish enough to be hit by a bash. I believe it is stronger than meets the eye, but some sources such as flatfoot are very overtuned in the new system, and many evasion buffs need to be functionally useful. 
My initial suggestion was to lower both the hit and evade cap to 25 or 30, reduce some of the sources so that it'd only take 1 -2 evasion or hit buffs to reach, and increase base weapon accuracy by ten-- accomodating the loss of the glance check. It seems incredibly dumb to me that evasion is suffering BEFORE the hitcheck.

Before I get lynched for saying this, hear me out:

Dodge in its current incarnation is simply a "stat check" in a sense that it just effectively automatically wins against setups unable to reach the hit numbers needed to win, or have adequate amounts of evasion ignores to deal with it. It's a concept I've seen in other games as a "win more" stat in the sense it only really enables you to bully builds that would have been likely ill prepared to deal with you in the first case. The major exception to this trend is Youkai, who bully tank setups quite heavily due to the sheer momentum advantage summoners hold. I cite this very momentum economy as to why dodge feels so miserable to play, you lose a lot of momentum and stat investment for something that often straight up doesn't matter in many matchups. A well stacked armor tank (it's own issue of balance) matches or outperforms on the DR front in nearly every normal situation using the average 250 hit this meta sees, as no dodge will ever get above 215 evade without niche shenanigans. This is further a issue when CEL requires nearly the same investment as a DEF/RES user stat-wise (If not more!) in order to have a DR. Evade is unreliable to the point where running absolutely no DR option results in oftentimes superior results, as at least you can then try your hand at completely winning the damage race instead of hoping your evade gamble pays off in a system rigged against it. 


TLDR: Current class balance is ill suited for this kind of system, leading to both dodge players to be upset at how their choice of DR is largely meaningless much of the time, and for everyone else because dodge if not kept in check is so utterly broken that you HAVE to play the hit/evade game, ruining build diversity as a whole.

WIN MORE is a very apt way of describing it's current function. It feels like this is a result of one; The numerical difference in hit / evasion being a very small percentage of the total amount (Subtractive) and Two; that evasion does a lot more than just reduce damage. It suffers from being capable of too much, Nullifying youkai, Glancing inflicts what have you, and in a system where nobody wants to do 0 damage the natural consequence is to be in an investment race that is ultimately decided by stacked RNG in one persons favour. It will never feel good. 

Again, I have hard summoner privilege and have to deal with like 1/3 of the weaknesses evasion has, but I fear that buffing evasion across the board only makes the abuse cases stronger. Numerically it's mostly fine I feel, barring a few evasion debuffs, but many sources of evasion suck balls and should be made usable.

Not too much to say on the rest of your post as I agree with just about everything, but I did forget to bring up luck and this has pissed me off greatly.

[Druby]

Seeing as a bunch of nerds above me did all the math, and I agree with basically like 99% of what everything everyone said above, I'll try to keep my contribution to the discussion on the shorter end of things.

I think one of the big issues with GR2 comes down to consistency. I remember talking about this once on the Discord, where I mentioned that potentially missing all the time wasn't fun, and Dev replied saying that, in return, the character who's dodging is probably enjoying themselves, which is a perfectly fair argument. In competition, the one who's winning is probably generally the one enjoying themselves, and that's hard to move away from. And to be entirely fair, at the end of the day, dodging is basically another form of DR, as 50% chance to avoid something is effectively 50% DR over a long period of time (please refer to Goose's fat ass chart for actual numbers I'm just simplifying). Everyone talks about how polarizing the Hit vs Evade game is, which I absolutely agree with, but I think one thing a lot of people haven't quite pinpointed on why the game feels the way it does is just how fast fights can go.

To explain a bit further, it feels to me that fights in GR2 have been, in general, shorter than in GR1. The hit to Def/Res that came to GR2 drastically reduced the value of those stats at high values, which certainly helped to reduce hyper tank builds, but I think it has had a further unintended consequence. Between the stat tax of needing to build Skill at a minimum, the reduced value of Def/Res at high values, one of the places people can pull points from is their Def/Res which even further reduces the average staying power in fights. Not to mention dodge builds, who are recurringly some of the point stretched builds out there, used to generally shuffle some of their free points into Def/Res no longer have that luxury anymore, and now often rely almost entirely on their evade to survive. This has, essentially, turned a lot of PVP, even for tanks, into a game of rocket tag. Everyone has less staying power, and fights are much quicker as a result. This, in of itself, is not a bad thing, as there's something to be said about matches being resolved in very short times, but this is part of what I feel to contribute to that consistency problem.

Dodge DR is an AVERAGE. You only really get to see averages over a long period of time. When your average PVP match lasts for probably 5 rounds or less (unless you are in a match up against someone who kites, heals, or a build that counters your damage type), you really don't get much chance to see this and fights where you rely on evade seem much more like a game of Russian Roulette where you're gambling on whether or not the next hit is what blows your brains out. 

In GR1, the hit/evade game was there only for people who WANTED to play it. Otherwise, you could be slapped by an auto-hit and have a reliable 30 DR from evasion in the vast majority of cases. A bit on the low end, and evade builds were still regularly considered a bit underpowered, but this seems like a simple enough fix by upping the evasion DR as someone suggested above, maybe to 40 or something. However now dodge builds often times get nothing at all save for the RNG check for when they get tagged, and that sort of RNG swing is felt much harder over a smaller amount of rolls compared to a larger average.

I will continue to stand by the idea that GR1 was a healthier version of the game (though not all the changes in GR2 were negative). However, if the intention is to stay with GR2, one potential fix I would suggest would be to increase the amount of HP every player has. Perhaps increasing the amount of HP investing a single point gives you from 1 to 2, or maybe even 3. Since players have 240 stats to invest at level 60, this increases HP in 240 increments, and making fights draw out a bit longer might help with that "feeling" of consistency I mentioned.

[Fern]

I agree in that increasing overall survivability might be an option too. Fights end way, way, -way- too quickly compared to GR1 fights. The rocket tag was kind of funny in GR1 because it wasn't as common, but now it's a thing for a lot of fights because of the damage reduction nerfs and evade being so easy to hit combined.

It's actually a pretty realistic occurance in PVP for an evade character's fight to end in about 3 rounds nowadays. And that's not even an exaggeration.





ah close ggs

I don't think this is because damage is too high across the roof, I think this is because of the overall DR nerfs (including Evasion being made much easier to hit) making damage seem a lot higher than it used to be. And I feel that if the route to nerf damage across the board was taken, you'd see Zerg's example of "the rich are barely affected, the poor suffer" in that the meta classes end up just fine and the non-meta stuff is screwed over more.

Imagine a Black Knight versus Evade matchup in a world where damage is overall lower with the current system. Seems pretty 10-0, IMO.

[sadbot]

There are many walls of text in previous posts, so I'm not going to repeat everything that's been said previously just to add my two cents onto the end. What I will say is that I agree almost entirely with Crixius' talk on the current issues that exist and the potential solutions. My only difference in opinion is with the solution to the problem with current evasion mechanics where I'm of essentially the same mind as Fern, so I'll focus on that instead.

I've been vocal about my sort of 'dream solution' to the evasion problems now since the GR2 dust settled. Something to the effect of reverting the system to GR1 evasion, but with a higher evasion DR to support what was lackluster evade build survivability at the time. Secondary effect avoidance would also stick around in some capacity, to maintain a separate identity to tanking for dodge builds as Fern mentioned. This punishes people not choosing to dump skill while still making it a valid option, something very important for build diversity which is unfortunately entirely lost right now. The orbiting 'secondary parts' of the system such as evasion ignore and hit/evasion changing effects and their frequency and availability would of course be reverted/changed to better fit something similar to the GR1 system.

I'll explain now why I think this would be the best solution. I think that the current evade system is virtually impossible to balance effectively due to the sheer number of moving parts in this game. On top of this, even in a dream land where hit and evade are well balanced between all class and build options I think that the system is inherently unenjoyable to interact with. Forcing basically every build to hard dump skill is very unhealthy for build diversity for a start, and squeezing every last little bit of hit and/or evade out of your build only to lose that number roll anyway feels so bad for the loser, even if the odds were fairly balanced. Having something as major as almost all damage you deal be heavily mitigated or even nullified entirely by a numbers roll is also mechanical torture, as fights can be made or broken by which side gets lucky even when hit vs evade is well balanced, ESPECIALLY when it's well balanced. It takes control of the outcome out of player hands more than anything else of its kind by a large margin, such as status infliction rolls for example. The numbers stacking game also makes pvp even more noob unfriendly than it already was, making an unoptimized build essentially incapable of so much as touching any half decent evade build or triggering so much as a single glancing blow against any build with half decent hit, making for a very discouraging new player experience with pvp.


The best way I can paint a picture of how the current evasion system plays out in practice is with my verglas build. It's my only build that refuses to engage with evasion mechanics and crutches entirely on evasion ignore against evade builds while putting almost no points into skill. It's also my most fun build to make and play as a result. It has better build options and diversity than my other builds by far. It has a large amount of sanctity which is fun and thematic to build and play, something I've found almost unjustifiable even in small quantities on almost any other build post GR2. It also has enough free stats without paying the skill tax to have decent survivability without being forced to wear heavy armor and gouge armor as hard as skill, allowing me to play with some unarmored synergy without being forced to play evade or be some kind of no DR glass cannon.

This results in a very clear picture being painted every time I go against an evade build. Against non evade builds, I get to have the kind of build diversity I could never have on my other builds with the now required skill tax, while still getting the full flexible, mix and matching skills verglas playstyle. Against evasion however, the game plan shrinks down to essentially just 'initiate with icicle spear then hit them with something else while they're frozen', with a side of zoning people out with the threat of evasion ignoring setting sun. The most interesting thing that ever happens with this build against evasion is icicle spear into kadouha into face stomp, that's the one 'combo' I get to do if I'm lucky compared to the full flexibility and enjoyment of chimera style shenanigans. Now don't get me wrong, this is still incredibly effective and wins the match against evasion more often than not, but it's just not very fun and often repetitive due to the narrow playstyle. I'm sure that a lot of people can agree that icicle spear spam isn't very fun to play against on the flip side either. Not having to deal with all of my damage coming down to essentially just rig-able rng is very nice, but it comes at the cost of my playstyle being forced into being narrow and less fun against evasion. I tried every which way to make a version of this build that builds skill so I could play the fun way against evasion too. I mean come on, it's an ice focused class, if anything should be able to do it well you'd think it would be verglas. The unfortunate reality is that even taking every reasonable bit of hit I could (At heavy sacrifice to every aspect of the build), it would never be enough for the old icicle spear strategy to not just be the better and safer option against evasion regardless. This could be used as an argument for better numbers balance or evasion ignore nerfs, but as I mentioned above I think this system is just inherently unfun to interact with even in an imaginary world of decent balance due to its inherently random nature and the various negative effects it has on the landscape of pvp and build diversity.

TL;DR - Crix makes good points. I think GR2 evasion mechanics are virtually impossible to balance and inherently unfun to engage with even in a dream world where they are well balanced. I think skill tax annihilates build diversity to the point of making certain builds essentially 'unmakeable', and we'd all be better off with evasion mechanics akin to GR1 with buffs for evade.

[Kiss]

Live reaction of a Evade only player actually hitting an evade on a menial attack in GR2 (colourized)






Then the next attack piercing through your full optimal ghost passives +50 bonus evade + san scaling stat race max capped evade bonus (colourized)

[Sawrock]

just make GR3: combine skill and cel