I'm seriously considering making a spiritual successor.

[PBarnum]

Hey guys, sorry for the long hiatus, I've been putting in 40 hours a week since I started my new internship.

Great job on it so far Xeropace! I took one look at the GUI windows/void color and instantly thought Unity, haha. I love working with Unity and could code in it all day long so I can help if you need it AND if I have the time. Level design in Unity is practically non-existant because it is like Thothie said, model based. There is a nice piece of software for creating environments, but it does cost money. Modeling for Unity is easier because the source work flow is half modeling, half conversions (lol). I did just find out however that I cannot use my recent models anyway due to using the student version of Maya 2014. If I ever get a real copy (with or without money), then I'll keep going with Maya but I might need to just bite the bullet and switch over to blender.

@Orochi, sorry to see you go. You should still stick around for consultation!

Keep up the good work!

 

[Xeropace]

I may have purchased a WiiU over the weekend and eaten some of my dev time.

I've been working on integrating Unity's Particle systems into a way that can be utilised to produce similar effects to how Half-Life did it.

Player PoV:


Other players PoV:


The blood effect itself needs some work as it looks pretty crap but at least it means attacking/damage is sync'd over the network (though you may notice the final doesn't cause a blood splat: the cube dies on the server an instance before the other players recognise that the attacker has 'hit').

I really need a pair of animated FPS arms for more feedback. I tried over the weekend and totally failed. My rigging was awful, I think I aligned the fingers on the world axis, instead of their own local axis. Animating the fingers produced very strange results

 

[Thothie]

Might wanna see if Barnum can help ya with the viewmodel there. Mind that you might wanna think ahead a bit here - for instance, about how you're going to handle two arms, weapon attachments, and different armors.

Hope the blood splat sprites are client side - no need to track them server side, save for initialization. Also keeping them client side lets them move more smoothly, as you don't depend on net updates (plus, if they are server side, a lot of blood splat = a lot of lag). As for aesthetics, I assume you can adjust their alpha, that'd help... Not sure how you go about getting beyond the Goldsrc era particle system though - maybe dynamic scaling would help some (not that you can't do that in Goldsrc).

 

[Stoned]

It's premature to ask this, but how will the game world be made? Will there be individual maps like we have now, or will the players be able to explore Daragoth without being impeded by loading screens?Is Unity capable of rendering an extremely large world, like one in an MMO, without severely affecting performance?

 

[Xeropace]

Might wanna see if Barnum can help ya with the viewmodel there. Mind that you might wanna think ahead a bit here - for instance, about how you're going to handle two arms, weapon attachments, and different armors.

Good points. At the moment two hands was just going to be a single model with various poses and animations for all the weapon styles, two handed, single handed, shield, magic, etc. Which is easily achievable thanks to Unity's animation system (it's like a State Tree where animations follow different paths based on variables). It's easy to change the skin on the fly for the hands, which would be 'good enough' at changing armours, if different models were used instead (to give some actual depth/detail to the armour), as long as the bone configuration stays the same, the animations could be shared across them.

The main issue with re-targetting animations is that the mesh might not deform as nice as the original model.

Hope the blood splat sprites are client side - no need to track them server side, save for initialization. Also keeping them client side lets them move more smoothly, as you don't depend on net updates (plus, if they are server side, a lot of blood splat = a lot of lag). As for aesthetics, I assume you can adjust their alpha, that'd help... Not sure how you go about getting beyond the Goldsrc era particle system though - maybe dynamic scaling would help some (not that you can't do that in Goldsrc).

Blood splats are client side. All the server sends to the other players is, "This player just clicked attack", the animation, the raycast and the blood splat are all worked out, spawned and played clientside. This is why the final blood splat fails to play. Not entirely sure what to do here, I should probably be sending "player ___ hit ____" messages instead (there's really no need for them to do raycasts for anyone but themselves).

Now another interesting point is, currently all combat calculations are done server side. The player only raycasts to draw the blood effect, they do not tell the server, "I hit this", they just tell it, "I attacked in this direciton". Problem is, if I'm going to have hit rolls like MS:C does... I need to somehow sync a random number generator over the network, so that the clients next roll matches that of the server. Otherwise, the client might draw blood when the server missed or the player might not draw blood when the server hit.

Things like size and colour over time are possible. I just haven't played with it much. I only spent a couple hours on particles and that was mainly coding a class to handle it all. The time put into the effects was minimal. I did try Alpha over time but it didn't seem to work, though that may be an issue with the shader I was using.

They do look a different to how they do in the gif as the gif only captures at 8FPS, which seems to look more like snapshots.

It's premature to ask this, but how will the game world be made? Will there be individual maps like we have now, or will the players be able to explore Daragoth without being impeded by loading screens?Is Unity capable of rendering an extremely large world, like one in an MMO, without severely affecting performance?

Unity can do larger worlds than Hammer, I'm sure. It depends on how the creator sets it up though. You could have a tiny complex map with lots of lighting, monsters and physics based stuff and it would be slow, or you could have a 1km x 1km terrain with nothing on it and get 1000 FPS. Plus with networking involved, the more dynamic objects (players, monsters, items, etc) the more data that has to be sent. So at some point I do need to mess with the networking stuff to see how stable it is and how much lag it will produce under stress.

I was still aiming for maps like MS:C though, mainly as I think it's more fun. If people wish to grind/farm, they can run through the map, clear it and run it again. It gives each map a small, instanced, more purposeful meaning. The only thing I could see changing would be merging maps that are currently split in to a single map. As far as I know, maps were only ever split in MS:C due to limitations and/or to avoid crashing.

 

[Jelly]

Crazy thought here, how about making the firstperson camera the thirdperson camera and not rendering the head?

 

[Thothie]

Ah, the Mount & Blade approach is always an option. This can look a bit awkward, from time to time, but it certainly has its advantages. Mount & Blade's models are fairly low poly, as it has to render a hundred or so of them at a go. Wouldn't need to be the case for MSU, and you could just have player models with considerably higher poly than those of NPCs.

I need to somehow sync a random number generator over the network, so that the clients next roll matches that of the server.

I'd really love to do away with the "hit chance", but even if you wanted to keep it, you wouldn't need to do that. In MSC, the client tells the server it attacked, and then the server decides whether or not the attack did damage, and reports back with the result of whether or not it rolled a hit. There's other reasons an attack maybe mitigated, so the final hit/damage resolution needs to be handled server side - though other things may crop up that could require such a seed sync.

It would, however, be really nice to keep as many things client side as possible - makes for a better feel all around, in addition to less load on the server. My dream was to work up a way to make a client side mob, that just periodically synced across the network, but I couldn't work up a way to do it with the limited client information, and prevent the mob from teleporting about during syncs. Closest I got was hunting particle spell effects - the sort of mob that works, so long as you can't hit it.

Unity can do larger worlds than Hammer, I'm sure.

I think what Stoned really wants is the sort of open world effect you get in most MMO's. Even in such massive games, the maps are streamed in and out under a chunk system, and not loaded in all at once. Chunks with no active mobs or player adjacent to them are unloaded, thus it makes it viable (sometimes, in such games, different chunks sets are handled by different servers). It leads to oddities, in some of the more poorly written games... But I'm not surprised Unity doesn't have a rig for it. I've never found an open source engine that does, and only one retail engine that is easily available (and it doesn't look very impressive otherwise).

Not entirely related but, believe it or not, I did actually fiddle, a little, with procedurally generated dungeons, in MSC. (Found this little VB proggie that made random maps.)

 

[Xeropace]

Having the external 3D model visible from the players PoV has it's ups and downs. The main reason it doesn't work for the current idea is that when you look up and down, the spine matches your rotation on a 2 to 1 angle, so 90 degrees up, the body only pivots 45 degrees (as any further, it starts to look very painful). Also with a pivoting body, the position the camera should be at will change, so the camera will no longer line up with the body. Most FPS games have a separate 3D model that only has the torso/legs when you look down, the arms/guns are still separate first person models.

I'd really love to do away with the "hit chance", but even if you wanted to keep it, you wouldn't need to do that. In MSC, the client tells the server it attacked, and then the server decides whether or not the attack did damage, and reports back with the result of whether or not it rolled a hit. There's other reasons an attack maybe mitigated, so the final hit/damage resolution needs to be handled server side - though other things may crop up that could require such a seed sync.

No hit chance? What's your reasoning? Funny, I thought the same thing but Skilla said to keep it At the moment there is no hit chance. To me, hit chance feels weird in such a game as you physically get in to position to strike, if you miss, you miss because your aimed incorrectly, not because a die roll said otherwise. And yes, you can seed random number generators so they produce the same order of numbers, but ensuring the random rolls stay in sync would be a mini-challenge though I already have a solution in my head. I wanted to sync it, rather than poll the server for "did I hit?" so that the hit sounds, particle effects, etc could all be played when the player expects, not anywhere between 10ms-300ms later when the server responds. I did think maybe to poll the server for the roll result in advance (e.g. at point of clicking attack) so that it had time to tell the client what the outcome was when the animation hits.

It would, however, be really nice to keep as many things client side as possible - makes for a better feel all around, in addition to less load on the server.

I still think most of the important stuff like that should be done on the server, mainly to avoid hackers. Currently, if they got access to the memory of say, a mobs health and they set it to 0, nothing would happen, the server is in control and will basically tell that player, "no the health is 50". If suddenly the mobs are client side, there's nothing ensuring that values aren't modified.

As for procedural/open world, I was hoping to avoid that. Most games feel so flat and empty that it made me realise how awesome MS:C was with its hand crafted maps. I'm sure I could add some procedural things, on lets say, a maze map but really, I'd like to keep the maps focused/instanced. The current server plan is the same as MS:Cs, lots of smaller servers running whatever map they're on. It allows players to easily pick what they want to do, it keeps server load down, it's easier to code for An MMO style world/server is a totally separate project. I don't think the network engine I'm using would support such a system.

 

[Xeropace]

Sounds are in, first person animations are in. It now feels somewhat 'playable' as the screen actually responds to what you do.

The model is a free asset from: http://www.blendswap.com/blends/view/26032

Rigging and animation is by myself, which is why it looks so awkward

http://www.gfycat.com/NervousNeatCub

 

[Thothie]

Most FPS games have a separate 3D model that only has the torso/legs when you look down, the arms/guns are still separate first person models.

Most do... Mount & Blade is one of the exceptions, and it works well enough there. On the other hand, the only other one I can think of is a game some efriends of mine have been working on called Starforge - where it is done, horribly, awfully wrong. If you turn quickly, you have to wait for the model to shuffle its feat and look around, and the uneven terrain and physics have the model constantly slipping, causing your view to bob randomly, so - yeah - sadly, I can't play their game more than 10 minutes without getting ready to puke, thus I've not been able to help them with it (insomuch as I'd be any use).

No hit chance? What's your reasoning?

Like ya said, since you have to get into position, and deal with a mob, who might even be physically dodging, it maketh no sense. I probably woulda axed it ages ago, if not for the fact that it'd mean rebalancing every weapon and be one of those fundamental "core" changes that I've been trying to avoid. More importantly, I'd hope MSC's successor would lean more towards an action skill oriented game (and perhaps strategic character/equipment builds), rather than purely time invest, so that players of disparaging levels can work together more easily (with maybe a 200-400% difference between min and max level, including equipment, rather than an exponential one).

I still think most of the important stuff like that should be done on the server, mainly to avoid hackers.

Two things 1) it dun make a sodding bit of difference, if you have listen or user hostable servers, and 2) you are looking at a slippery slope that leads to a dark chasm that I know, from painful experience, you don't wanna go down. Half of MSC's performance issues and mystery bugs are due to efforts to deal with hackers, and it's just a never ending war. I think you're much better off combining the Borderlands "f*ck it" approach, with a central server and a volunteer police force and/or "/report" function, to deal with cheaters who go *way* over the top and start messing with everyone. You certainly do not want to be favoring cheat protection over performance at this primordial stage.

Though, as far as client side mobs are concerned, at least, it's hard to make that work at all, nevermind the cheating potential.

http://www.gfycat.com/NervousNeatCub

Those poor, helpless red cubes... How does your hit boxing work, at the moment?

 

[Xeropace]

More importantly, I'd hope MSC's successor would lean more towards an action skill oriented game (and perhaps strategic character/equipment builds), rather than purely time invest, so that players of disparaging levels can work together more easily (with maybe a 200-400% difference between min and max level, including equipment, rather than an exponential one).

Action Skill orientated? I'm not sure what you mean with that part. Do you mean people wouldn't have to grind their stats and instead would be more focused on a players skill (as in, the player behind the keyboard is 'good' at the game). One thing I liked most about MS:C was the fact most weapons were very unique, with their own set of skills and usages. It meant you weren't just throwing away all your old weapons in favour of the newest. Things like Bone Blade combined with Urdual Shield let me play with the big boys as I could just out heal the damage Sure my damage wasn't great but I was helping Me and Skilla spoke about it and we wanted to avoid boring number based upgrade on gear, otherwise you end up in a WoW situation where you spend all your time and money getting an item, for it to be obsolete a couple of days later.

You certainly do not want to be favoring cheat protection over performance at this primordial stage.

Well, at this stage it's hard to tell what may potentially be performance hits The game runs at like 1000 FPS+ unlocked as it's not doing much rendering or processing at this stage. I'm one to always profile as I go along to hopefully spot bottlenecks as early as possible.

Those poor, helpless red cubes... How does your hit boxing work, at the moment?

Currently, every uh, hittable creature has at least two 'colliders'. One (a cylindrical collider) is used for positioning them in the world, where they stand, if they're grounded, blocking people from running through them, etc. Then the other(s) is/are used for the hit detection in combat, they can be cube, spherical, cylindrical and even mesh colliders (though I will probably stick to primitives). The red cube only has one hit box, a cube that is just slightly larger than his model. Each hitbox has a parent "entity" that it belongs to, when a hitbox is hit, the server will update that entity's health based on who/what hit it. The idea behind multiple hitboxes was so I could do things like various levels of armour/damage scaling, so something could have an armoured front with a weak back, or maybe headshots on humanoids do more damage, a boss with an evil eye that have you hit, etc.

 

[PBarnum]

Unity's Network abilities or Photon Network? I can't stand working with Photon and haven't tried Unity's.

 

[FER]

Action Skill orientated? I'm not sure what you mean with that part. Do you mean people wouldn't have to grind their stats and instead would be more focused on a players skill (as in, the player behind the keyboard is 'good' at the game). One thing I liked most about MS:C was the fact most weapons were very unique, with their own set of skills and usages. It meant you weren't just throwing away all your old weapons in favour of the newest. Things like Bone Blade combined with Urdual Shield let me play with the big boys as I could just out heal the damage Sure my damage wasn't great but I was helping Me and Skilla spoke about it and we wanted to avoid boring number based upgrade on gear, otherwise you end up in a WoW situation where you spend all your time and money getting an item, for it to be obsolete a couple of days later.

I do agree on what tothie says, because thats the only way for a game with rpg element where you have actual control over your character.
I dont know how to explain it but a way to understand it is to look at games like monster hunter (or dark souls maybe): characters dont level up but you get stronger weapons and armors focus more on special skills given to the player than its defensive capabilities.

 

[Xeropace]

Unity's Network abilities or Photon Network? I can't stand working with Photon and haven't tried Unity's.

http://www.boltengine.com/

Much less fiddly than Unity's Networking yet much more developer friendly that Photon. I tried Photon too and was forced through so many hoops that it put me off before I even got started.

Bolt basically handles all the boring/hard stuff and gives a high level code approach to networking, so I still have nearly fully control on how things work.

FER, Monster Hunter is one of my favourite games but not entirely sure how to adapt it for something like MS:C. Are people saying they don't want things like Strength, Sword Skill, etc?

I was still planning to use them as well, in something like MS:C, it stops people just borrowing/being given rare items and suddenly being up to scratch... I feel like some time should be invested.

 

[ceriux]

yeah i feel like stats need to stay in some form or another.

 

[Thothie]

Action Skill orientated? [...] Do you mean people wouldn't have to grind their stats and instead would be more focused on a players skill (as in, the player behind the keyboard is 'good' at the game).

Well, not *exactly* - there has to be some progression, otherwise it's not really in the same genre as the original. It'd just be nice if there was more of an emphasis on actual player performance and build decisions, rather than a focus purely on the time sink aspect. This also helps players of varying levels of experience work together more smoothly, meaning new players can join in and have fun more easily, in addition to casuals and basement dwellers working together with little to no issue.

Ya still need stats and such, but the difference they make should be more marginal - just enough to create the illusion of progression and provide development goals.

One thing I liked most about MS:C was the fact most weapons were very unique, with their own set of skills and usages. It meant you weren't just throwing away all your old weapons in favour of the newest.

...Darn... That's a bit where we may differ, as I was hoping to abandon that quirky and randomly stair stepped equipment based progression, in favor of a wide range of item generation (Diablo II/Torch/Borderlands style), which, yes, probably would have you throwing away items left and right - though I was thinking it'd be nice to combine it with a weapon mastery bonus (ala Defiance, though maybe more drastic), in addition to weapon customization mods (gem socket/blade modification/upgrade/appearance etc) in order to make you more hesitant to do so.

The idea behind multiple hitboxes was so I could do things like various levels of armour/damage scaling, so something could have an armoured front with a weak back, or maybe headshots on humanoids do more damage, a boss with an evil eye that have you hit, etc.

Excellent [/mrburns]... Half-Life1 actually has this functionality integrated into the model structure, but the way Dogg setup hitboxes in MSC seems to have killed it (with the exception of players, everything's really just a box). Tis a shame... Closest I can do is actually have an entirely separate script as the hitbox and attach said entity to a model bone, as I did with Undamael's head, which is clumsy as hell.

 

[Xeropace]

...Darn... That's a bit where we may differ, as I was hoping to abandon that quirky and randomly stair stepped equipment based progression, in favor of a wide range of item generation (Diablo II/Torch/Borderlands style), which, yes, probably would have you throwing away items left and right - though I was thinking it'd be nice to combine it with a weapon mastery bonus (ala Defiance, though maybe more drastic), in addition to weapon customization mods (gem socket/blade modification/upgrade/appearance etc) in order to make you more hesitant to do so.
.

Incoming a bit of a rant, apologies in advanced

My main gripe with random rolls on loot is that it's just as much as artificial time sink as grinding character stats, only the progression isn't guaranteed. For example, when Diablo 3 originally launched, most people hated it. Every single piece of gear was about rolling the best possible combination of stats (dex/str/int, vit, all resist, DPS, etc), every item was a plain old yellow with -nothing- special about it, in terms of cosmetics and abilities. They ruined Legendaries, one of the key features to why people like Diablo 2 so much. Uniques/sets/rune words in D2 made the builds SO much more interesting. E.g. the very rare/sought after Enigma which enabled other classes beside the Sorc to use Blink, which gave birth to the Hammerdin build for Paladin. Granted most of the 'good' items in D2 gave +x to all skills, which in turn let characters improve their customised builds more. Diablo 3 also did away with the builds... Making every character essentially identical (you either used the cookie cutter build/gear or you couldn't deal enough damage, it was awful).

Both D2/D3 games had little special about the different classes of weapons, didn't matter if a Warrior was using a sword, axe, mace, etc, the gameplay was identical. Even Borderlands feels like that to me, you only ever keep a hold of the top 1-3 guns you've found, the unique effects are either boring or useless, so you end up using a random rolled item yet again.

MS:C though, every weapon was different, with different skills unlocking at various levels (mace got stun, dagger got the speed boost, etc), then on top of that, you also got access to a wider array of weapons as you became proficient to equip them. Each weapon potentially had their own abilities, increasing your possible options, without ever forcing you down a single path. Not to mention, every item was uniquely identifiable by it's world model, which made collecting/showing off that much more fun. Okay sure there were cookie cutter items (Dark sword, hoarfrost, blood drinker when I played) too but you still found yourself swapping weapons lots of times in a single run, which I personally think was a lot of the fun. Strangely, I've only noticed this looking back. I think MS:C was ahead of its time.

Every bit of exp you got in MS:C, was a step forward. You could play for a whole day, not get a single item, yet still come out stronger (or at least a step closer). Diablo 3, I used to play for a week and get -nothing- of value and my character would be no better/different to the start of the week, whilst those who got a head start had furthered the gap between us (better gear = faster runs and higher item drop chance) and had started to exploit the market by charging millions of gold (which people paid because no one wanted to farm random rolls). Every weapon you find in MS:C, is an extra tool to your arsenal (okay, some exceptions, like the earlier weapons that are just damage upgrades), none of the time spent getting the other weapons is wasted. Every item is named, every item has a possible personal attachment, every person has their own favourite. Even if the weapon isn't designed for combat (Look at Raven Mace for example, I know for one Skill and I love that thing) it still potentially has a place. In Diablo 3, your items mean nothing, it's only the numbers you care for. You throw away items left, right and centre, potentially items you spent weeks farming/saving for. The problem with things like MS:Cs exp system, is yes, how do you balance it in terms of time invested to stop those who can play it all the time, getting too far ahead. Games like Planetside 2 have horizontal progression, as in, you don't level up and do more damage, you just get access to different tools/weapons, that way everyone can contribute but you still go, "Oh man, I want that weapon". So maybe... Weapon skills affect what you can equip but don't scale your HP/damage? I honestly have no idea

Interestingly enough, Diablo 3 recently buffed Legendaries and have tried to bring back the unique and useful feeling for named items, they also made them bind on pickup to discourage people abusing the market. It was too little too late for me though I can't name you a single item from Diablo 3 or the Borderlands series. I can name you countless items from MS:C and I even still remember loads of Diablo 2, which I haven't played in years and years. Hell, I even remember Phantasy Star Online's weapons and it's all because they are unique in terms of model and effect.

 

[Skillasaur]

See also: My signature.

 

[FER]

IM not too fond on weaponstat rolls either, tough prefixed weapons with sockets to put some stuff is okay.

However im with the idea that health shouldnt be able to raise as you level up, instead be affected by skills or race class. Same with defense being dependant only by the armor.

 

[Monika's_BFFEx0256]

It may be a little too early to be bringing up complex mechanics, but seeing as how MS:C has a somewhat dynamic battle system; Perhaps a combat system similar to Zelda (Or Demon/DarkSouls) where you can lock on, and dodge your opponents' attacks. This is already in MS:Ckinda and it'd be amazing to see it in a better adaptation!

 

[Xeropace]

They're aliiiive!

http://www.gfycat.com/ImaginaryFloweryAsianpiedstarling

 

[Regorty]

Soon, they will stack upon each other and create pillars of red cubes so there will be no hiding in high places from them!

 

[Red Cell]

Hahaha reminds me of imprinting with ducks.

 

[Thothie]

Soon, they will stack upon each other and create pillars of red cubes so there will be no hiding in high places from them!

Our mobs already do that! And it's totally deliberate, and not at all the result of crappy step size adjustment scripting on my part.

[Wall of text rant worthy of Thothie]

Gawds, where to start... While I agree with most of that in principle (especially about Diablo 3 being crap, from what little I've seen of it), it is, as mentioned in the WoT, very difficult to balance items that way, and having every item be unique, means it's difficult to keep players, well, playing, once they've nabbed everything.

It's probably too early to really be discussing this anyways, but I'd suggest maybe combining the two strategies. Provide level generated "unique" items with unique abilities at predictable locales, as well as the randomly generated items, and item modification parts (blades, hilts, socket gems, enchantments, etc). Thus, you retain unique items, provide motivation to go and get them again (after X# of levels makes modifying them to match up with the higher level variant too resource intensive), and give players something to do between expansion patches, even if they have already nabbed all the uniques and hit the level wall. Added advantage being that even your unique items are varied, depending on customization.

But it'd be very nice to have a predictable leveling system, where you can say level X item is going to be doing X dps with X +/- variation based on X level modifications. One biggest meta problems that makes developing for MSC difficult is a lack of predictability in levels and development. That's further exasperated by the stair stepping all the unique item/spell powers involved, but you can at least mitigate that somewhat by assuming characters of a certain experience are going to have access to certain equipment - though that conversely means players without said are going to get unexpectedly stomped.

As for procedural generation, it's nice to have small randomized areas, *maybe* some randomized instanced dungeons for the truly bored, but yeah, you don't wanna go all Minecraft and randomly generate your whole world. I've never seen that done well... Seems it'd be *possible*, but you'd need a complex and intricate system beyond anything I've yet been witness to.

It's your show though, just making random suggestions based on previous experience.

Hahaha reminds me of imprinting with ducks.

Strawberry Jello ducks...

 

[Xeropace]

Yeah it is too early, considering I don't even have an inventory/loot system yet It's still good to talk/figure it out early though otherwise one may start developing down a certain path without realising.

I hadn't thought about what happens if a player doesn't have the desired items... You either have to make it so every map/boss/whatever can be beaten with stock gear and then any extra weapons are just bonuses... Saying that though, most people only even attempted certain things in MS:C thanks to the security of Armour of Bravery.

Rather than farming to strengthen their weapons with modules... They should farm... Hats and cosmetics! Then we can add keys you have to buy with real money to unlock chests!