Guild Wars 2

alain.1659:

snip

If you read the post all calculations were done AGAINST each other. This means a warrior was hitting the “tank” and still doing double the DPS and with zero traits more than 33% more burst damage than the Mesmer. You use roughly 33% more stat points in damage stats but do more than 33% more damage than the “tank” with two damage stats. So, all in all, this argument is 100% flawed. Against someone with just base toughness the warrior would do double the DPS AND double the burst listed in the calculations.

Essentially stats still do not compensate for the lack of damage of conditions even after the buff.

I’m going to compare the burst a bit more as I think there’s some interesting analysis to be done there. The Mesmer has to ramp up to three clones before attempting the double shatter. This means when fighting a Mesmer, if you see them generating scepter or staff clones (or really, any clones in general) you need to keep them suppressed. Mesmers are GOING to shatter now. It’s not an if, but a when. To counter their burst before it happens, make sure you drop AOE’s to keep their clone generation down.

Also, the Mesmer is likely now to offensively blink or roll into a shatter. Shatters happen on the Mesmer now too, so offensive shatters are much more likely. The ranged clones have to run in, but the freshly generated ones will pop right on top of you. If you see a Mesmer with 3 clones roll or blink into you, pop defensive cooldowns or dodge.

Lastly, if you do get hit by the shatter bomb full force you still have options. Your best bet is to pop a two condi cleanse. This will remove a HUGE portion of the burst. Try to save your cleanse for this moment. A good Mesmer will try to bait out your cleanses and dodges so when they shatter you go down. Be aware of this. If you can cleanse, you can survive. Also, when the new boon Resistance becomes more common you can just pop that for a few seconds to mitigate about 10-15% of the burst damage per second of resistance. Again, if you can wait try to save your resistance to counter the burst.

By contrast, the Warrior burst has little counter play. They are going to root or stun you into a Hundred Blades. It’s a very simple formula. Save your stun breaks and condi cleanses for this moment. If you eat a stun or immob you’re probably going to eat a Hundred Blades chop or two. It’s pretty hard to eat the CC, insta cleanse and eat zero shots. Don’t get too low as two or three strikes could put you in the dirt. Even if you survive the first Hundred Blades it’s on an 8 second cooldown. Warriors can fish for good Hundred Blade shots quite easily, but they need CC to land more than a hit or two. Shield is common for Hundred Blades Warriors as they advance under the block stance and aim for shield bashes. If you see a Warrior shield up, back up.

NOTE – I said COUNTER PLAY when referring to the Warrior. Yes, I know there are COUNTERS and that is not what counter play means. Counters refer to things that outright trump other things. Counter play, by contrast refers to the complex decision making tree during actions. Look at the Mesmer and her targets decision making during the burst. There are options for both sides on when to strike and how to use their cooldowns. Now look at the Warrior. There is little decisions to be made as you simply try to CC into the Hundred Blades. The difference in decisions to make is astronomical. The Mesmer duel decision process is a great example of counter play.

With lots of talk about the condi changes in sPVP, lets take a look at some numbers to see how imbalanced conditions really are.

Mesmer Scepter in sPVP rabid amulet with Undead runes (and the 10% toughness to condi trait) will be used for this calculation. Torment sigils were used to test how many stacks of Torment I could keep up while auto attacking, but are otherwise unused in the calculations. No other talents or sigils are taken into account.

Warrior Greatsword in sPVP zerker amulet with Hoelbrek runes will be used for this calculation (but their condi reduction is NOT taken into account). No other talents and no sigils are taken into account.

All calculations are done assuming one class is hitting the other. Torment DPS is calculated as if the target was always moving.

Alright, set up aside, let’s look at the auto attacks. We’re assuming both classes standing still eating each other’s auto attacks. With this in mind we’re assuming three clones hitting the Warrior along with the Mesmer. This can reliably keep up 5 stacks of Torment.

The total DPS for the mesmer scepter auto attack is 2,804 DPS. 194 DPS per stack of Torment per second. 1,048 damage for the first and second swings on a .5 second swing timer, and 1,834 damage for the last swing in the chain with a 1 second cooldown.

The total DPS for a Warrior Greatsword is 5,881 DPS. 2,685 for the first and second swing on a .5 second swing timer, and 3,452 damage on the third swing in the chain with a .5 second swing timer.

Greatsword Warrior does over double the DPS. That’s not too bad, considering he’s melee and can be kited.

Now lets look at a burst example. Both Mesmer double shatter and Warrior hundred blades are easy to dodge and block, so I feel it’s a pretty apt comparison.

Mesmer is assumed to have generated 3 clones. She will shatter with Mind Wrack, pop mirror images and a second clone generating skill, and shattering with Cry of Frustration for 3 clones. All shatter calculations also include the Mesmer since shatters go off on the Mesmer now too base.

The Mesmer will apply 10 stacks of Torment (with the torment on shatter trait) for 6 seconds. The Mesmer will also apply 15 stacks of Confusion (with the blind on shatter + confusion on blind traits) for roughly 8 seconds.

With a DPS of 194 on Torment, and 73 DPS on confusion, and 2,286 direct damage on the two shatters this comes to 22,686 total damage over 8 seconds. 22,000 of that damage is able to be cleansed. Since Mirror Images is mandatory for the quick double shatter this can be done on a 36 second cooldown utilizing 4 skills. Note this calculation does not take into account any swings while under confusion.

Hundred Blades will deal 16,148 damage over 3.5 seconds to the Mesmer. Hundred Blades has an 8 second cooldown, and requires only one skill cast to obtain the printed damage.

Both burst damage values, if everything hits, would be enough to kill the other class.

So, a zerker melee Warrior has over double the auto attack DPS, and does 71% of the burst of a Mesmer on one skill cast with no setup on a six second cooldown, and can kill the Mesmer with that skill alone.

The Mesmer has half the DPS of a Warrior on the auto attacks, but is ranged. They do roughly 40% more burst but it requires 4 skills to be cast and a major part of the burst chain is a 36 second cooldown. 90% of the burst damage can be cleansed as it is done over 6-8 seconds.

I don’t see where the issue is with conditions (aside from burning) at all. Dodging and blocking stances affect both the burst and auto attack DPS of the attacks equally. Protection reduces direct damage by 33%, but condi cleanse removes roughly 90% of a Mesmer burst (about 10% less per second it’s left on, so I’d say on average you can catch 70-80% of the condi duration). Mesmer clones can also be killed to reduce their auto attack DPS substantially, and make it hard to get the perfect double shatter off. Overall, I’d say conditions still need a bigger buff (again, burning aside). If you have other arguments like stealth and so on and so forth, then remember your argument isn’t against conditions – it’s against stealth. If stealth + conditions is an issue, then stealth needs to be tuned down, NOT CONDITIONS.

Note – Weapon coefficients are taken from the GW2 wiki. Condition DPS rates were taken from the patch notes. The damage formulas were taken from the GW2 wiki as well. Attached is my math. I won’t explain the excel. If you don’t understand it, read the wiki.

EDIT: Looks like I forgot the last strike of hundred blades. Add in another 1500 or so, not too much, to the Warrior burst. Regardless, if he gets a full hundred blades off the 15,922 HP Mesmer is dead.

Overall, the Thief needs to be looked at in Guild Wars 2. The style of play it brings to the table reduces complex decision making for the thief and to counter stealth as their class mechanic leads it to the best burst in the game while simultaneously making them almost unkillable at the press of a button. As it stands now, Thieves don’t get kills through skill, it’s simply mechanics that allow them to pick up kills. Players play PVP because they don’t want to play simply against mechanics. PVPers want skill to be the defining characteristic of their fights. Bring depth and skill to the class and to people reacting to the class. Of course, changing the stealth durations will require more balancing with class survivability, damage, and traits. As I have outlined, the class reduces complex decisions for all players involved and that simply is not good game design. Please go back to the drawing board and take another look at the Thief class.

If you read this far, thank you! You’re a trooper. I know it was lengthy but game balance is very complex and understanding the fundamentals is important. Thanks so much for your time and I hope you learned something.

Guild Wars 2 could go to this model, but it would take massive rebalancing. I don’t think it’s viable as every class is meant to be able to do everything. As you can see, lots of emphasis is put on deciding WHY, WHEN, and HOW to implement stealth in your game. As Anet seems to have bolted on stealth to jump on the bandwagon, they missed out on a great deal of this planning phase. So what could Anet do?

Of the two things that Anet could do, normalizing and lowering stealth durations is the best option. In my opinion the Mesmer is very well balanced as far as stealth. It simply acts as a way to throw off your enemy for a second, reposition yourself a bit, and reengage while possibly dodging a spell. If you engaged something you shouldn’t have then it probably won’t save you. Their stealths sit at around 3 seconds (the best amount of time in my opinion, they should all be exactly 3) and their elite is 5 seconds and does nothing but stealth. They also get no bonuses when stealthing which keeps it from becoming even more unbalanced. Remove the ability to increase stealth duration and it would be perfect. The low time means that you can’t use stealth to endlessly to analyze potential targets and completely ignore fights that you can’t win. Low duration also reduces the amount of time you’re able to wait out your burst potential. Abusing stealthed movement is at a minimum with low duration and makes it only useful for slight repositioning not outright quitting a fight. It still gives the first strike advantage, but as I’ll discuss below allows for more tactical decisions and skill from both players. This makes short untargetability, first strike, and repositioning the benefits of stealth and still makes it very powerful.

Increasing the number of untargeted spells is a great way to mitigate stealth’s benefits to a simply deceptive role. Smite is a MOBA with only untargeted spells. Everything is a skill shot. They have one character, Loki, who has a relatively short stealth. He can really only use it for escape juking and quick engagements from the jungle because if you try to use it in full view of your enemy they will just CC where you disappeared from and hit you with everything anyways. This brings a great tactical element to stealth and AOEs as it gives you the ability to make more decisions in response to a stealth. In Guild Wars 2 there are two things preventing this from being viable at mitigating stealth. Dodge rolling in stealth and delayed AOEs. Very few AOE spells are instant cast making the ability to react to a stealth and hit them quickly impossible. On top of that dodge rolling makes it so that you can just avoid the AOE and stay stealthed. The best way to balance this is to make dodge break stealth. This plays out very nicely on the battlefield in regards to short stealths.

Let’s say a Mesmer decides to hit her torch 4 and enters a 3 second stealth. The enemy has a few options. Blow a defensive cool down and absorb the burst, try to hit the Mesmer with AOE to deal damage or CC to keep pressure on, if they have stealth they can stealth second and just get to have the first strike advantage when the Mesmer leaves stealth, or do nothing and absorb the re-engagement or let the Mesmer leave combat but have their cool downs up for the next time they need them. Let’s say the enemy decides to blow an AOE and drops it right on the Mesmer’s head. Now the Mesmer has a decision to make. Do I blow the dodge roll, leaving stealth and allowing the enemy to have the first strike advantage while I finish my roll animation or do I eat the damage? With long stealths the defensive buff duration won’t last long enough, the CC / damage will be impossible to follow up on, and the enemy can just walk away regardless. There’s no skill or branching decision making to be had. I could easily give more examples of decision making that long stealths eliminate but good skilled and complex decisions that short stealths add.

As you can see this branching decision making adds a lot of skill and complexity to PVP, something that is very good. However, this assumes that stealths have low durations and dodge rolls be adjusted to only allow the dodging of one spell (in this case the AOE) and not complete invulnerability during the entire roll. However, even if the dodge animation makes them completely invulnerable like is currently the case the enemy still baited out a dodge roll making that one less available to them later. I have several issues with dodge as currently implemented however but that’s another story. Let’s leave it at I think all buffs to endurance pool size and regeneration should be removed from everything and it should be normalized amongst all classes and dodge should only stop one attack when you hit the button choosing amongst the first to hit you.

As you can see, stealth can easily reduce the skill level of classes and reduce meaningful skilled choices players can make to adapt to it. Especially in Guild Wars 2, stealth can make it so you win fights by mechanics, not by skill. So how do you balance stealth in a game with all these massive advantages it gives? It’s very logical and quite simple. I’m going to refer to three examples I feel are extremely well balanced stealth mechanics in PVP games: The Mesmer, EVE online, and Smite:

• Remove stealth or do not include stealth
• Remove the ability to do anything useful after stealth
• Design your game with rock, paper, scissors balance
• Keep stealths short in duration
• Increase the number of untargeted spells

The first is the most obvious. Just don’t have stealth in your game. Stealth is such a powerful mechanic and designers really need to see the role very clearly for stealth before including it. Don’t just include stealth because it’s the hip thing to do. Including stealth in a game is by no means going to increase the reception of it. Quite the opposite is true if it unbalances the game in some way. Also, make sure if you do decide to include stealth make sure your engine can handle it properly and it’s not buggy. There’s a lot of complex things code wise that stealth needs to do in a game. If your game simply cannot handle stealth from a base perspective, think of another way to fill the niche your game needs. This is rather impractical in Guild wars 2 as the design team has already committed to using stealth. However, they could make stealth simply turn the Thief shadowy and make him untargetable but still visible. This would make them still susceptible to AOEs but still give them the ability to mitigate most kill strike burst damage. This could make for a very interesting and balanced alternative in my mind, but still shouldn’t have more than a 5s duration on non-elite skills.

Removing the ability to do anything useful after stealth means quite simply you only want it to be used for scouting purposes. Trying to make a high damage low survivability class stealth is completely impractical. Make it so they can burst a target quickly and it becomes a one sided fight of “can the stealth character choose their targets properly.” Make the stealth class do too little damage and the enemy simply turns around after being engaged and guts them like a fish. The balance line is so razor thin it’s best to avoid it altogether. Having stealth characters do no damage isn’t practical for a lot of games but games with big open worlds, sandbox games, and other creative uses can often make use of a pure scouting or support role. It could be tweaked to make sure the class that can stealth can only use support functions like CC or buffing friendly players. This has the added benefit of increasing communication and team play as the stealth class generally cannot solo kill anything but can be a force multiplier for their team. If you want a great example of well balanced stealth support and scouting, EVE online does is splendidly with 3 different classes of stealth craft. One class can only really scout enemy systems, the other can do fairly decent CC but very low damage, and the third is only good for fleet logistical purposes. This is impractical for Guild Wars 2.

Another from the design perspective, balance your game in a rock, paper, scissors fashion. Make it so stealth characters counter something, but something hard counters stealth characters. You see this in MOBAs frequently and some RPGs. I think it’s good balance as it emphasizes team composition, positioning, and sticking together much more but other players dislike it. It promotes strategy at the cost of reactionary tactics. What you’ll often see is warrior characters having decent CC and physical mitigation, but low maneuverability and mage characters doing high armor ignoring damage but with low escape and survivability. What ends up happening is if a rogue engages a warrior the warrior brushes off most of the rogue’s damage and caves his skull in with CC and damage. If the rogue stealths up to a mage he can gut the mage in a matter of seconds leaving the mage with little defense. Mages by contrast do heavy damage to the warrior as most of their damage is unmitigated. You can throw in other things like support characters and objectives to make sure the team is focused on supporting each other instead of running around mindlessly TDMing each other. This has the added factor of increasing balance. If the rogue gets a huge buff for some reason teams need only to increase the number of warriors to compensate which then makes mages more viable etc.

Movement and positioning can often be the key between winning and losing a battle in most game types be it PVP, PVE, what have you. When you stealth the enemy has no idea what direction you’re moving in but you can still see theirs. This creates a huge disparity in what the non-stealthed player can do to mitigate what you can do. This highly favors mechanics instead of skilled gameplay. At best, they can see you off in the distance, watch you stealth, and move away from you preventing you from gaining much distance on them. At worst, during a melee combat the stealth character gets a huge advantage to peel from combat if they are losing since they can just move in the opposite direction the enemy is or makes it so they can stick to a fleeing target with no hope of the fleeing target landing CC because the stealth character can’t be targeted. It also could allow a stealth character to stealth out of view of a player and close the gap between you two without the player being able to do anything to mitigate this gap closing. Again, being unable to mitigate this means the mechanic not only requires less skill to use but promotes less skilled gameplay from opponents as well.

First strike advantage is absolutely gigantic in any game. Entire classes and spells are built around mitigating first strike advantage for the simple reason that classes like thieves and rogues exist. The element of surprise gives even the most skilled opponents a couple seconds of disparity which can be devastating. If the stealth class has any CC this becomes even worse. The stealth character can strike, CC, and strike a few times more before the enemy gets any chance to do anything reasonable. If the stealth is short this is generally negligible as the opponent can predict when you will come out of stealth more accurately and take appropriate measures to counter your opening. However, it does still provide a fairly solid benefit even with short stealths. Being able to press a button and gain the first strike advantage also mitigates skilled usage however it does allow for the defending player to make skill based decisions which could then take adapting to. I’ll discuss this a bit more later.

Untargetability is obviously a gigantic advantage. This is what allows stealth characters to have the first strike advantage, gives them the ability to avoid CC, and allows them to escape things other classes are simply incapable of. Couple this with the ability to avoid being seen and it makes even untargeted AOE spells incredibly hard to land. Without the invisibility part untargetability is still strong but it can be mitigated by AOE CC and AOE damage. Untargetability itself allows for very skilled play with short durations to interrupt spiking combos, dodge or anticipate important CC from enemies, and mitigate ambush burst on squisher characters. With long durations and invisibility it basically provides invulnerability for the entire duration of the stealth and again, reduces the impact on skilled decisions and instead emphasizes mechanics.

Scouting undetected can be gigantic in certain games. Games like EVE online have entire squads dedicated to reconnaissance because in war, knowledge is power. In MOBAs stealth generally gives a huge advantage to scouting as well as knowing enemy movement can make the difference between lost assets or a defended push. The lower the sight lines, game importance on knowing enemy movements or compositions, and size of the combat maps can make or break undetected scouting as a mechanic. At short durations of stealth, undetected scouting becomes almost impossible. From a game design perspective at the inception of the game developers need to analyze what emphasis they want to put on scouting. Measures like EVE online’s inability to target enemy players for several seconds after stealth or stealth ships’ inability to do substantial damage can put heavy emphasis on balanced scouting and supporting without impacting balance in combat situations. This is one of the few instances permastealth can be easily balanced.

Stealth is a relatively new addition to gaming. It seems like every MMO from Shooters, Space games, RPGs, MMOs, and MOBAs all have to have stealth these days. In this post I’m going to analyze why stealth is an inherently unbalanced mechanic and why should generally be avoided in games in the future unless there is a specific purpose for it instead of being bolted onto any genre and trying to balance it afterwards.

First off, let’s visit stealth as a mechanic. We’ve all seen it. You enter stealth becoming invisible for a time and then reappear under set conditions be it time, action, or something else. However, stealth gives a TON of benefits, both tangible and intangible that need to be considered during balance. The longer the stealth is, the larger most of these benefits become and often the more uncounterable the benefits become. This uncounterability of players while in stealths lends itself to use of mechanics instead of skilled decision based gameplay. So let’s analyze these pieces one by one:

• Time to analyze your enemy
• Cooldown reduction
• Movement closing
• First strike advantage
• Untargetability
• Scouting undetected

When a PVP battle starts several things happen in a very short span of time. What a player does the first few seconds of combat can determine how good they are as a player. Let’s say you see your enemy off in the distance. You now have a few seconds to determine their class, look at their armor and weapons, analyze their buffs, and from there try to determine how their character is built and if you could handle their build in your current build. This can make or break a good player. Engaging an opponent when they are built to counter your build is suicide, but if you guess wrong at their build they can send you a haymaker in combat that you weren’t expecting and force you to adapt.

The challenge and fun of PVP begins at first contact. The less information that is outright given to you but clearly displayed on the enemy the better, but that’s a topic for another time. Stealth breaks the initial period of analyzing, theorizing, and preparing for the coming battle. One party gets to analyze the other often for large amounts of time while the other can have no idea of the existence of the other. The stealth character gets the ability to almost perfectly select targets they know they can kill. It skews the initial planning phase of the battle drastically in the stealth characters favor with no more than the push of a button. This removes a huge skill aspect of PVP and greatly favors mechanics over skill-based gameplay.

Along the same lines, during combat you often have to adapt to enemy tactics and strategies. Giving yourself a big time out during combat is a gigantic advantage. However, with stealth, this also applies to your enemy. This can be a wash in instances where there’s not a mechanical advantage to having a slight breather. However, in games where a stealth class has lots of front loaded damage this is a dramatic advantage. Not only do you get to take a breather and plan your next move but you also get to sit out all your cool downs and apply more front loaded damage again. It essentially removes the detriment of burst damage which is generally high cool downs.

The longer stealth is the more it lessens the number of decisions opponent can make for how to deal with it as well. So if the enemy has no important decisions to make while the character who is stealthing does, this creates an even greater informational advantage. I’ll say this multiple times, but again this points to a mechanical advantage instead of skilled gameplay.