Sunday, 27 April 2014

Weapon Scaling In Dust 514

So anti vehicle balance has been a hot topic recently. I decided to try out all manner of different means to destroy my enemies. Some saw success, some felt lacklustre. After getting some different opinions on the matter, I decided to take a break from playing tonight to see if I could crunch some numbers.

The Weapon Scaling Formula

So there was one constant that I found within the majority of the different weapon types I looked at. The following formula appeared

y = zx-x

x = Standard weapon damage
y = The difference between weapon ranks Standard->Advanced and Advanced->Prototype
z = 1.05

There are several notable points we can gather from this.

  • A fully skilled commando using a standard light weapon does more damage than another dropsuit with a prototype light weapon.
  • An advanced weapon with a complex damage modifier is slightly better than prototype weapon without one.
  • A standard weapon with maximum specialization skills is slightly better against it's specific offence type (shields/armour) than a prototype weapon that is boosted by a complex damage modifier.
  • Prototype weapons have a pretty bad price:performance increase ratio compared to using an advanced frame that provides an additional high slot for a complex damage mod.
  • Skill points give a bigger improvement to damage over equipment.
Now there's some balance issues that weapons have right now, but generally the difference between tiers of weapons is reasonable.

But while this formula applies to most weapons in the game, there are a few notable exceptions.

The Swarm Launcher

These are every Duster's first anti vehicle weapon. Every player is given a militia variant of these on one of their starter fits. They also have both the best and the worst damage scaling in the game. With a z multiplier of 1.25, they get a massive boost of damage between ranks. This would be great, if it wasn't scaled from the top down.

What does that mean?

Typically when balancing competitive games, you want to make sure the game is balanced at it's highest levels of play. This is due to the fact that you want your game to either be as good or better as players improve at the game. It's also harder to balance lower levels of player skill, as it becomes harder to determine if there's an imbalance within the game or between the player's level of skill (although developers should try). In Dust, this balance also exists between tiers of gear, with standard vs standard, advanced vs advanced and prototype vs prototype supposedly being balanced against each other. This is all further complicated when you throw vehicles into the mix.

Because of this balance issue, the need for prototype to perform well against prototype means that scaling has to be fairly equal between the different tiers. In the case of Swarms, their huge scaling multiplier makes them weak at lower tiers. When there's nothing stopping someone from using a prototype fitted tank against a player who only has access to a militia swarm launcher, you end up with a situation where the new player's anti vehicle weapon deals negligible amounts of damage. Considering the gap between a basic swarm launcher and an advanced, it wouldn't surprise me if they weren't performing well enough to kill tanks at their own level.

In summary, the difference between most weapon tiers is a complex damage modifier. The difference between swarm launcher tiers is 2 complex damage modifiers and 5 ranks of specialization. This is really bad for new players and forces people to use prototype weapons to be competitive.

The Forge Gun

Unlike the swarm launcher, the forge gun benefits from having a more normal scaling system. With a z multiplier of 1.1, it tends to be balanced at different tiers. Forge guns were bugged when they entered 1.8, so it's hard to determine how balanced they are.

From a personal perspective, they seem to perform well at killing vehicles and at killing infantry. So maybe 1.1 is the correct z multiplier for an anti vehicle weapon? I'd be interested in hearing opinions on this.

If so, it puts the plasma cannon in a poor place. As the Gallente anti vehicle weapon, it gets stuck with a 1.05 z multiplier. The regular plasma cannon does a reasonable job against standard tanks, but considering it's level of scaling compared to other anti vehicle options, maybe it could do with being on the same z multiplier as forge guns.

Nova Knives

These have a pretty crazy z multiplier at 1.66. If you're looking into using knives, you should definitely be using the advanced ones at least.

From a damage over time perspective, they also lose charge time between tiers. The base skill also increases this too, so they become amazing at the prototype level compared to the standard ones. Although it could be argued that the advanced ones already do enough damage to make the jump to prototype not worthwhile.

What's somewhat disturbing is that a Minmatar scout with maxed out suit and knife skills does less damage with standard knives than someone without bonuses using advanced.


I'll finish with grenades as it's getting late and I'll look at vehicle turrets another time. But these seem to be on a separate scale when it comes to damage.

The AV grenade has a z multiplier of 1.2, thus making it closer to the swarm launcher than other weapons. AV grenades fell out of fashion after being nerfed in Uprising 1.7, but can still do serious damage to vehicles when used as a group.

The other 2 grenades get a 1.25 z multiplier, so perhaps it's the role of a grenade that makes them unique for this. Flux grenades bring great utility beyond just their damage and said utility increases at a 1.1 multiplier due to their radius increasing. Locus grenades I'm not so sure about, since their biggest impact on me is to force me out of cover. They do this well regardless of their damage, since I'm trying (and get the opportunity) to avoid that completely.

Grenades have a fairly low cost at all tiers, so you could say they give the biggest bang for your buck when you upgrade to higher tiers.

So What Does All This Mean?

Well if you're looking for better performance at killing vehicles, I'd suggest that you work towards the better tiers of swarm launchers as the lower ones are under-performing right now. If you're looking for a lower level means of anti vehicle, both the forge gun and the plasma cannon have better standard level performance.

Nova knives scale like crazy and you should be using the advanced ones as they truly shine over the standard. Skills are less important compared to ISK investment and prototype level knives are mostly only worthwhile if you're looking to chew through health stacked sentinels.

Grenades are an extremely cost effective means of bumping up anti-vehicle damage, although they are more skill point intensive. If you're looking for other long term investments of skills, you should pick up weapon specializations and commando skills (for light weapons).

When it comes to other weapons, skills are king here too. Specialization  You're also better off using standard or advanced dropsuits with complex modules, rather than upgrading your weapon from standard to advanced. Upgrading a weapon to prototype gives the lowest benefit for cost, so really the last upgrade to a suit you should make.

Friday, 18 April 2014

Design 514: The New Player Experience

"We're Killing Dust"

So this topic has been brought up again with the recent "WE ARE KILLING THIS GAME" videos by CEOPyrex. During this, the point was brought up that certain players have been entering public contracts in squads of full of proto gear and higher tier vehicles. This isn't a new problem by any means, as newer players have been getting the sharp end of the stick on this for something like 15 months now. The problem actually pre-dates the game entering open beta. I remember week 2 of the game going live on Tranquillity, struggling with my exile assault rifle against GEKs and people who were already using tanks.

But how long this has been a problem isn't exactly relevant to fixing it. Looking at the solutions that CCP have tried and failed with during that time does. Lets look at a few.

Offering Better Alternatives

I'm sure there's a carrot/stick analogy here, but whatever. Offering a better place for veterans to really go full throttle against each other is a pretty standard affair. The challenge is greater, but so long as the rewards are too, many high level players will be willing to try. Every attempt Dust has made towards providing this has failed on a few basic levels.

  • The original corporation battles had a poor reward:risk ratio.
  • Faction warfare rewards don't provide a sustainable way to play. Game mode doesn't provide a reliable way to play with a full team. Could potentially work with full team queuing and a method of trading assets between players.
  • Planetary conquest has a huge barrier of entry before you even get into the match. There's too many other issues to fully list them all. CCP are fully replacing it and scrapping it's current form in the future.
  • Tournaments lack the tools for third parties to run them effectively. CCP lack man power to run regular events themselves.
There's also the final issue that none of these provide as many skill points per hour as running public contract ambush matches. So are inferior options to people simply wanting to acquire SP.

The Current New Player Experience

So with the carrot not being...carroty enough for the veterans, CCP decided to give new players a stick free zone in the form of the academy. This provides separate matches for players under a certain amount of lifetime warpoints. Much like it's carrot based cousins however, this stick free zone has it's issues.

For starters, there's nothing stopping veterans from remaking alts to play in it. Some people enjoy being vindictive and likely won't stop just because they're asked. There's the issue of people saving up skill points and then coming back to play with a large advantage over there peers. New players have no means to play with more experienced friends in this safe haven of sorts. Finally, the academy does little to prepare those who go through it for what awaits them in public contracts.

Designing A Better New Player Experience

Until today, I had no real suggestion that would meaningfully address this problem. That was until I read this article by game designer David Sirlin regarding the same topic on one of his own games. In it he talks about how when teaching people to play his games, he would provide a simpler ruleset to get them used to the basics. Having used this method with one of his earlier games, it's how I teach new players before introducing them to the more advanced parts of the game.

Looking around, other games have this too. Counterstrike has some lighter modes to allow people to get used to the movement and weapons. Smash Brothers has a variety of modes and stages that allow for people of all skill levels to enjoy. Early levels in platformers start with easier jumps and weaker enemies. World of Warcraft starts your character with a couple of basic skills and progressively gives you more.

Dust 514 throws you straight into everything with a few boxes of text between you and the enemy. Tanks don't care if the pen is mightier than the sword, because this isn't the 19th century. Dust is a complicated game and gives a serious advantage to those who are knowledgeable. The closest thing we have to easing the transition for new players is slight limitation of tanks in ambush. Not to mention the gap between the basic starter fits and 

So the solution seems fairly obvious now. We add a way to play that boils Dust down to it's simplest form, while keeping the current modes intact. What I would propose are "Training Matches" that only allow a select amount of free, premade fittings to be used. This would include basic gear with a small variety of fits to allow players to get used to using different weapons and equipment in a number of different roles. It would also limit vehicles to basic LAVs only in order to get people used to vehicles in a simpler form. Also removing things like the warbarge strike and possibly turret installations.

The downside to this mode would be limited rewards. You wouldn't be earning nearly as much ISK per match. The SP rewards could be smaller post-cap as well to keep public contracts more competitive. There's the possibility to add a further bridge between this and the more competitive modes with a full team version. Entering the full team queue would provide a mode with slightly more advanced mechanics, such as higher tier outfits and more vehicle types. The ISK rewards would still be lower than public contracts, but it would provide a way for people to get used to playing in larger groups.

There's obviously more to something like this than a couple of simple paragraphs can cover. As a basic concept though, doing something that's based on what many other games have done with a large amount of success seems like a good place to start. Adapting it to fit Dust is the harder part.

Sunday, 6 April 2014

Design 514: Balancing Cloaks

As a quick note, I have only 4 or so hours of laptop battery to last the next 3 days while I'm visiting family. Boots on the Ground was meant to happen before I left, but ended up not due to DST shenanigans and me needing to be up early the next day. I'll be writing a couple of blogs while I'm gone, including this one.

So anyway, Uprising 1.8 was released recently and while not perfect from a balance standpoint, it seems much better when looking at the overall design. Classes now feel like they have actual roles. Sentinels are kings of point defence, Logistics are no longer just better assaults with more equipment, Commandos actually have a role as long range damage dealers and Scouts are fast, sneaky little buggers. There's still balancing to be done (Assaults are close to being good), but picking a dropsuit gives you a purpose on the battlefield.

One of the biggest additions with this update was the cloaking device. This was seen as controversial from the moment it was announced, but has actually turned out pretty well. Moving around while cloaked makes you visible to the keen eye, allowing good players to kill bad cloaked players. Cloaking in and of itself has been designed well and has allowed for new play styles to arise.

What has become problematic however is not cloaking itself, but rather the other benefit of the cloaking device. Not only do these provide the cloaking effect, but also a (stacking penalty free) 25% reduction to dropsuit profile. This means that scouts can be invisible to most types of detection based purely on using a cloaking device. The player Haerr has provided a handy spreadsheet that shows how this all plays out. There's a few key points I want to touch on based on this info.

Gallente can get under all detection with skills, a cloak and an enhanced profile dampener. This means that I can (and do) run a basic suit that costs barely anything but cannot be found by people spending more on a single Gallente bonused scanner or a Caldari Scout with a full stack of precision mods. The fact that my cheap suit that isn't even dedicated to avoiding it's counters can do this is a balance problem.

Outside of Gallente, other suits can get under most detection with a single complex profile dampener and under the rest with a single enhanced. So not only are they part of the same problem, but have made the benefit of being Gallente a moot point. Why specialize into Gallente when you can pick another race and get more out of it? The only advantage left to Gallente would be if you didn't care that much about avoiding the specialized counters. Unless you don't care about being a Scout.

Which brings me to the armour stacking. At the current time of posting this, the Gallente Scout that simply stacks armour is a much better than the Assault that does the same. Sure you might lose when purely comparing health, but the increased speed, the reduced dropsuit profile, the extra equipment slot and the easily fit cloak means you're just better where it matters. Throw on an enhanced dampener instead of the 4th plate and you're basically an invisible Assault. The other races are having similar issues, but the Gallente is really where the problem is most apparent.

The final problem is that there's no reason not to use the cloak when it's available. I can't think of a single instance where being cloaked isn't advantageous. Cloaking just seems so mindless when compared to other actions in the game. When I fire my weapon, I sacrifice ammo and risk the enemy knowing my position in exchange for damaging my opponent. When I drop a nanohive, I exchange an equipment slot for my team gaining additional ammo and the risk of losing said gains to the enemy destroying it. When I cloak, my only risk is when I can't find a position within 30 seconds to decloak, at which point I get my cloak back after a short period of waiting. There is no risk when it comes to cloaking. It's essentially an anti-risk device that has no drawbacks other than short, manageable periods of time where you don't use it.

While I can see what CCP were going for when they gave the cloaking device it's profile dampening, it just dumbs down the gameplay of those who use it. It allows players to be great at avoiding detection without the need to invest into anti-detection modules. It makes racial parity somewhat pointless and gives Scouts the capability to do other roles better than those who should do them best. It enables a 1 dimensional play style that has no drawbacks.

So what's the solution to this? Give it a drawback.

When CCP announced the cloaking device, my initial thoughts were that the scanner would be it's hard counter. If they reversed the dropsuit profile bonus on the cloaks, that could very well be a reality. Switching it from a 25% reduction to a 20% increase would force people to think more about when they should use their cloak. It would also force people who want reduced detection to use modules for it. It would keep Scouts from being able to do both their own role and the Assault's role at the same time. Running some quick numbers, new detection would work out like this.

Moderately dampened uncloaked Gallente Scout
Heavily dampened uncloaked non-Gallente Scout
Heavily dampened cloaked Gallente Scout
Gallente Logi focused scanner/Heavily precise Caldari Scout
Moderately dampened cloaked Gallente Scout
Lightly dampened uncloaked Gallente Scout
Moderately precise Caldari Scout
Moderately dampened uncloaked non-Gallente Scout
Focused scanner
Heavily dampened cloaked non-Gallente Scout
Gallente Logi proto scanner/Lightly precise Caldari Scout/Heavily precise non-Caldari Scout
Lightly dampened cloaked Gallente Scout
Lightly dampened uncloaked non-Gallente Scout/Regular dampened uncloaked Gallente Scout
Regular Caldari Scout
Lightly dampened cloaked non-Gallente Scout
Regular uncloaked non-Gallente Scout/Regular cloaked Gallente Scout

There's more to that list of course, but in terms of balancing it's much better than it is now. There's a sense of progression for both sides of the detection war. The best detection can only be trumped by a large investment into dampening. Being cloaked opens you to being vulnerable against certain targets while being safer against everyone else.

Does this fix everything? No. There will need to be more numbers tweaked and fine tuning. But this provides a first step in a more balanced direction.