So after seeing the two threads addressing balance outcomes in Atlas, I wanted to make a personal thread mostly out of reflections on the thousands of games I have had since Beta. In the hope to tell the story of at least my opinion, what I liked and what has made the game unplayable for me for at least the past year. This is of course, a long long post (boo!), but I feel writing down my thoughts will help some players think about the game.

So this post is structured as follows, you can skip sections as you see fit, I'll try keep each section self contained, but rather there is a narrative across them, so I appreciate if you take the time to read through. It goes like this: my personal thoughts on the game, what I think makes the roles interesting, and some mechanics good or bad, then thoughts on current changes (since beta) and most recent ones. I feel that the further this game progresses the further we stray from the original things I liked about the game. So I will try to sum up what I really liked in the game. Then explain what this introspection and the recent buffs to frontliners have made me think about. Because I think it revealed that some other things need work.

Personal Gibberish (and Thoughts about the game)

Personally I hopped on the hype train of Atlas Reactor, as someone who has used the simultaneous turns in some Text-Based MMO I designed long ago (in 2007), it was something similar akin to having strategies and adding some conditions to fights, since fights can happen even when you weren't online. It was akin to writing small AIs to control your actions. To do so, I needed the fights to be split down into phases, and items/skills could be used differently in phases (A shield could defend in one phase, but bash in another). Something like that, so I immediately hopped to see how they managed to do that in a 3D setting and whatnot. And I have to say, I was pretty impressed, so I clocked hundred of hours in the first week or so. And well even after thousands of hours I still suck at the game so hooray!

What liked about AR was how subtle one can be. That is particularly how it rewarded good positioning, LoS kiting, cammmo pads were interesting. Also how it allowed for sprinting as a way to "lose" frontliners. These were the fundamental things that I found were solid in the game. Forcing opponents to think about the next turn. At its beginning AR had a stable EV (Expected Value) and a reasonable one for many things. Zuki for instance could sticky to ignore shields. Helio comps could be starved of health orbs, so it was okay to risk your frontline being in a messy spot if it could negate a health orb and what not. The reason I bring this up, is that you can still do all of that, but the EV has increased variance now.

Personally I always always supported rewarding good positionning, you can clearly see my early advocation of adding cover reduction to ults, and reducing ignore cover mods, which ironically could have contributed to the double support meta I loathe terribly. Another thing I liked about AR and I used to play a lot of Nix, is the abiliy to kite fights, and double move at key turns (so that EV was maximized, becuase you dont deal damage this turn, but you can deal more damage the next few ones, or not be in danger). Similarly due to positioning, and I reiterate this claim, I liked that supports had half momentum heals for a full momentum FL/FP action. That meant, that a support could at most negate the action of one enemy Firepower, and sometimes even half. However if your guy was in cover, you could always negate all of it, or multiple Firepowers. An aurora big heal would heal completely cover but around 70% a full hit.

Another aspect I enjoyed thoroughly about AR eventho I do not myself use it often, is Taunts and Overcon, because it gives this friendly banter to the game, which makes it worth watching, I think the devs have since Alpha never explored improving this system at all, which I think is unique to the game (maybe add counter-taunts! or automatic taunts when someone goofs). I think there is much to explore here, and it would increase the appeal of the game further. But again this is mostly cosmetic but important. It adds to the allure of the game being a full experience.

So now that I have detailed on the things I like about AR. I want to go over some mechanics, esp recent ones, that have kind of made the game a bit of a bore in certain cases. Especially in terms of rewarding bad plays.

Thoughts on the Ranked system (~OK)

I never really liked the way the ranks are computed, I do understand however that the model is being refined and that the player population is low. I do understand that perhaps this is the best we can have for now. So I am not going to suggest much. I just want to comment on the fact that, perhaps having some form of objectives would help alleviate from the deathmatch aspect of it to give players a stronger edge. I see that the new X-Ray buff brings a strong emphasis on grabbing and controlling areas, a direction which I like (sadly, I really hate the buff itself, and the vision changes but more on that later). I think the way to put it, is that if I have someone on my team that I know is going to keep dying no matter what I do, the question to ask is: "what can I do to punish my opponents when hes doing that", and in the current way the game is designed, its really "not much", I really hope we could have some other stuff to do besides just killing people, I think the orbs themselves are less important than expected (although I do understand they are super important for high level play) just less meaningful in scenarios of ranked or whatnot. Also the way they are spread out and well having celeste just breaks this whole thing. I think ranked will be more meaningful, if players had more options, more things to do to contribute to the win condition.

Other than that, I understand this is far-fetched, I do not expect much changes here, and it kind of serves its purpose. So kudos for at least having a working model, I personally have no idea how to improve it past just adding more things to do. And perhaps less grinding, but thats the direction the team is taking anyway so.

Thoughts on the Respawn Mechanic (BAD!)

I think the respawn mechanic has rarely been revisitted for a year now, and I have no idea what the developers intend to do with it, frankly I have not have had the time to keep up with the streams or the game often so far, but in my limited understanding, I dont think any plans to revisit it are in the works. I have generally found the respawn mechanic to be flawed in many ways. I am going to explain it here, because I think this could lead to some frustrating things in the game (and no not just because I play Nix alot).

Firstly, I would like to point out that I do understand that games like this want to somehow provide players with a come back mechanic to ease out the death. However I mean, its deathmatch... and theres no gold like MOBAs, so I don't think we have to go hard on the comeback mechanics, as such, I think we have gone a tad bit overboard.

Firstly down for 2 turns seems reasonable as is right now. I really like the idea. Problem with it is that beacon provides HUGE amount of vision (for free!), and the person coming out of the beacon has a sprint so really he can just do whatever. Person coming out of the beacon can even autofollow (what?), and most imporantly you can spawn within one distance of an orb to negate it. So if I have someone respawning on the enemy team, I lose my ability to control orbs all of a sudden, talk about rewarding bad plays, simply because he can literally spawn on the other side of the map. Maybe if he can spawn within 6 squares of where he died, it would make more sense in that regard.

Either way I find it quite rubbish, firstly you can use it to reveal bushes, which means someone else on your team can use that vision to autofollow from across the map (YAY FREE PROBE!). Also yeah I do understand that if you choose to autofollow, you might give up the orb because it happens after movement. But I mean think of the EV of that... hmm thats pretty twisted. I can just use it to provide vision, then sprint out! Like woa. Not healthy at all.

So what should we do about it, hmm its hard to say, I would firstly if I want to give it a vision mechanic make it so its a grey drone area of vision, so something like about 1/3 of the normal vision, you can still do it in bushes, and globally thats fine. But at least it shouldnt reveal that large a map spot. I think that is super important. Why do we allow autofollow? thats not something that should be allowable out of a respawn. Perhaps we can also make it so its like the first turn movement, maybe not even waypoints are allowed. Harsh I know, but I mean... really.. we cant just have it as it is, its too weird for EVs of a serious game.

Ok maybe I really really hate the autofollow mechanic, but ill detail that later, but then again autofollowing from across the map is my real issue, vision is also my bigger issue. I think people tend to grossly downplay the importance of vision in this game.

Thoughts on Autofollow (~Good , needs polish)

A recent thing I wanted to revisit, was autofollow, because I think thats a good mechanic in the game, but yet it is easily abusable with respect to expected outcomes for certain situations.

Firstly,I would like to point out that I understand why you do not want to autofollow. If you do not know why, or are a new player playing against a PuP, think hell dash you? Just move and ask the aurora to ion cloud where you will be headed. Not convinced that it can be bad, hmm well I made this video for ****s and giggles.

Autofollow is a double-edged sword it implies you are taking the following risks:
  • Opponent may lead you to a bad spot, possibly through traps, which is horrible in general
  • You may relinquish cover on the next turn. As you may autofollow in the open (high probability).
  • You may not negate good movement spots, orbs, or clash players for a good position (that is, autofollow happens after movement, so you are giving away the priority on the strengh of your current position)

But it also rewards you as follows:
  • You are guaranteed to be near the target (unless you force a major cooldown, stealth or fade), excluding you being knocked back of course
  • You can autofollow if someone dashes (line dashes) so, you do not have to do much guessing, if the opponent can dash in 4 different directions (woa thats some high EV).

This is what makes autofollow a healthy mechanic, in general. Because the EV fluctuate very nicely between risk and reward. However I feel for certain conditions, an autofollower always wins. I am going to list some cases, and suggest some polish that could help alleviate some of the problems.

Autofollow can sometimes negate heavy disengages, which are high cooldowns (and thus it yields tremendously high reward), consider a gremo dash, if the person autofollow, then they will end next to the gremo next turn. If they are hasted they can even primary, and thus have an incredibly high EV, as the gremo dash has a ridiculously high cooldown. Sure it is easy to negate that, say you can knockback the frontliner or something like that. But I mean in non-coordinated play that is very unlikely to happen, or happens in very small number of cases. You can almost always autofollow a target if you are expected a dash (rip LW), especially if that target has a line dash (yup thats gremo, BB, LW, Celeste unmodded, Khita, TolRen, Oz, Su-Ren [excluding FLs and point dashes in bushes]). So the dash for the firepower which is their expected only way to survive (outside supports) is completely negated (hmm maybe that contributes to double support in a way... who knows). In a typical PVP/Ranked setting, if I am a firepower and the FL decides to autofollow, its always beneficial for them to do so if they have even a 50% guess I will dash. I do not think that it is healthy. And I may be exagerating a bit. Again good players only use autofollow under certain conditions, but this is a condition where it always yields a positive outcome, because an FL can play this game given that the FP has a high CD on dash, and lower health pool.

Another situations where autofollow is ridiculous I would say, is mostly due to long-range autofollows. That is, I can decide as a frontliner to second wind and follow an aurora from halfway across the map. Which is cool, but the aurora can do little about it. And def she can't smart position (again in terms of uncoordinated play this is only made worse).

Here are some polishes I would make for autofollow, which I think would make things smoother, these are things that can be and/or, and would improve skill required to play FL and would reduce the fact that you can play dumb 3 FL comps currently and smash everything. I want FL to have high damages, but only if they are playing well.

Make autofollow only possible, when within a given range of a lancer. Preferably within 5-6 tiles. Autofollow from across the map should not be as rewarding as it currently is.
Add a penalty for autofollow, something like a slow, which means if you autofollow you may risk completely losing the enemy if he double moves. Also it alleviates the requirement of all FPs to take extra range on their dashes or shield/stealth. It gives FPs more room to kite FLs. Again I like to reward players for thinking about their positionning, and guess enemy movement right, with more than just cover.
Thoughts on Probe (~Good but Dangerous)
I like the fact that we needed to add a vision catalyst, I still think probe is too strong, mostly because it can reveal large amount of starting cammo in the game, it also allows for the long-range autofollow to happen, and amongst other things, point-click dashers (Kaigin, PuP) to have a reasonable advantage. my problem is that its range is too big, it has to be cut down. One probe can reveal a large chunk of the start positions, sure people will say but that it was its intent, to which I say cool, but now people can play even more passive and hide even further back if they really want to.

Also its hard to tell if someone picked probe, so you can play passive anyway on first 2 turns. I think probe's range should be slightly bigger than a grey drone, in terms of vision, in terms of placement range its pretty good. Ability to autofollow in probed area compounds this problem.

For example, take garrison and titus with a helio, and 2 probes. What are the odds that you can dodge the blackhole on a titus dash combo due purely to a good positioning and opener. I think this oversimplifies the game and makes it quite boring to be honest.

Thoughts on Endless Momentum (TM) on Supports (Interesting but Dangerous)

One of the major reasons I think that double support became meta, is that supports can do alot in terms of momentum. Either by negating it or gainging it heavily for their team. Supports have great tools in terms of momentum, and also have ignore covers on some of their attacks, if not ults (with the exception of perhaps Su-Ren). I feel that going with this direction encouraged double support meta, and frankly the fact they can heal and negate a ton of momentum with the heal means that we have an issue on our hands. I'll take for example the Meridian Heal, the Orion Heal as well, Fate Transfer and Helio shields as examples.

Meridian heal can negate multiple actions, due to its weaken, due to the fact that it heals both meridian and his target, the lower they are, thus negating past momentum of people actually being in danger to just shoot them, one can see Meridian heal which is on a 3 turn cooldown, as something like a double aurora flare. Helio for example has great momentum due to 2 charges but ability to negate 2 FP actions if target is in cover, or 1 if not, that is the helio shield has no downside of using it, esp if you have another support who can heal every 3 turns, or each turn (Aurora). A fate transfer is a huge momentum issue due to it shielding orion anyway, and the target, so orion at least can get a 20 shield in the worst case.

I do not want to say I have a problem with this per se, but I think it is the reason support metas have existed, these supports have ults that are reasonably okay to tilt games, ignore covers for the most (Finn, Aurora, Khita ults), helio ult is an insane momentum shift etc. I just find it dangerous, and its rewarding aggressive reckless play more often than it is not. I also think we should tune it down a notch, becuase its making FP/FLs not a worthy pick. An Orion can do similar primary than gremo (not exactly, but similarly effective), with still an ult that gives him ridiculous survivability compared to a gremo, while the gremo ult is ridiculously strong still.

I am just not particularly with the things supports can do for free, or in terms of EV. Esp for things that reward weird plays. And the combination of these together create odd momentum shifts in the game. I think supports need a smaller health pool, but thats just me, at least make it so that we can punish them if we are taking the risk to position to shoot them. Not just take 30-40 damage to deal 40 damage to something that is negated for free. I feel FPs are most struck by this. Esp that now when they fire they reveal their position.

Before people could not position recklessly because they can get erased by a good focus fire, I feel the game is providing a lot of contingencies for focus fire, but more than it should. Which is making it dangerous. 2FP versus 2Supp will most likely put the 2Supp in a great advantage against the 2FP.

Closing Remarks

I think pretty much the FL buffs were reasonable, albeit slightly more aggressive than I hoped (I would want -2 damage nerf in general, but its all good). I think supports still have weird health pools given what they offer, in terms of risk and reward. And I think the game would be much better if we revisit the spawn and autofollows to polish them just a tad bit.

(Reddit Link)