on "push" mechanic

[alre]

I had read that a new push mechanic was added to the game, and I wanted to try it out. This is a comment on that feature. Tag to @wraitii as he/she wrote it.

I love the idea, but I don't think it plays out too well. With this new feature, formations are totally worthless, because people move even too well. Since there are no chain bumps slowing down mobs, there is almost no friction when moving blobs of people, which makes retreat very easy. I don't think I like the idea: the damage taken retreating is a big element in game strategy. Also, if moved to point when not in formations, people can form very dense masses that are both OP and weird. I suggest to raise the distance at wich there is repulsion.

Another thing I noticed, is that when moving people and rams together, people can pass trough rams like they are made of air, but rams can't do the same, with the result that moving rams is possibly even more frustrating than it already was. Is it possible to add "mass" to rams and elephants, so they can push other people and maybe not be forced to make long roundabouts to avoid people and instead push it? Formations could do the same thing.

[wraitii]

Hey, thanks for checking it out  

On 23/4/2021 at 7:11 PM, alre said:

I love the idea, but I don't think it plays out too well. With this new feature, formations are totally worthless, because people move even too well. Since there are no chain bumps slowing down mobs, there is almost no friction when moving blobs of people, which makes retreat very easy. I don't think I like the idea: the damage taken retreating is a big element in game strategy.

I will counter that by saying that you actually need some strategy to hit your enemy on retreat, e.g. putting some units in their path, instead of just getting freebies because the pathfinding is bad.

I can see how you could dislike the change, but I think ultimately it's just different, not better or worse. And as you noted formations already allowed this smooth movement, so this is just straightening things out.

Quote

Also, if moved to point when not in formations, people can form very dense masses that are both OP and weird. I suggest to raise the distance at wich there is repulsion.

Yes, that is a bigger concern. I am not certain that it makes them OP, though.
There is a limit on the distance at which units push each other, as things start behaving really badly if the numbers get too high, but there's probably room for improvement.

Quote

Another thing I noticed, is that when moving people and rams together, people can pass trough rams like they are made of air, but rams can't do the same, with the result that moving rams is possibly even more frustrating than it already was. Is it possible to add "mass" to rams and elephants, so they can push other people and maybe not be forced to make long roundabouts to avoid people and instead push it? Formations could do the same thing.

The one constraint the system has right now is that static units cannot push moving units, and so no, it can't fix the rams. That is there because otherwise things start going awry quickly, and i've not found a great solution.
I am considering reducing the ram and formation "size" for A25 though, since there is a bit of overlap anyways, because I'm generally thinking better movement > some overlap.

----

I think I'll give another look at the numbers when we get closer to the release.

 

[ChronA]

6 hours ago, wraitii said:

I will counter that by saying that you actually need some strategy to hit your enemy on retreat, e.g. putting some units in their path, instead of just getting freebies because the pathfinding is bad.

Looking at other games, that can definitely work out well as a design choice, but only if it is supported by other aspects of the balance plan. And without huge stat revisions I don't think 0 AD is likely to qualify.

Units here are too immobile relative to the rate at which they kill each other to pull off massive force reorientations in the heat of battle, but also not immobile enough to be effectively locked into any compromising position they land in. If one player divides their forces to form a hammer and anvil, the opponent will just destroy the anvil with their superior force before the hammer can get into place. Moreover the game's economy & map size is such that continuous reinforcement is pretty common. That means the opponent's rear is almost never entirely unguarded, and you more likely to force a 90 degree pivot than actually encircle the enemy. This is certainly what we see in AoE2.

If you want to have encirclements work in 0 AD, there are a few ways to go about it. One would be to massively slow the time to kill by boosting armor and hp, or lowering attack damage. this would encourage reactive maneuvering. Or you could speed up the time to kill and harden counter-strength into the one-hit-kill territory for advantageous unit matchups. That would incentivize ambushes. Either way you'll probably want to give a big buff to cavalry movement speed, while slightly nerfing infantry movement speed. You would also need to slow down unit production so player will stop parade-marching reinforcements, and nerf static defense a bit. Otherwise no one would ever commit to a fight on open ground.

 

An alternative perspective: maybe this should be paired with some compensatory mechanical revisions that would buff chasing. Running charges and directional armor come to mind. 

[maroder]

Generally judging by the screenshots provided by @Langbart here: https://code.wildfiregames.com/D1490

It looks to me like a improvement to me. But I do see the dense masses that can be formed without the formation.

The solution to formations being useless would be the planned formation bonuses: https://trac.wildfiregames.com/ticket/3523

 

[alre]

In ancient battles, only a handful of men were usually killed during each battle, most kills were scored while pursuing the retreaters. I don't think we should make retreating easier with this new pathfinder, without balancing things out in some way*. Cutting the retreat with additional forces is helpful and always will, but shouldn't be necessary for inflicting reasonable losses to a retreating force. But maybe I'm exaggerating the role of the pathfinder.

*@ChronA already mentioned running charges and directional armor, other options that come to my mind are slowing people down when hit, or having soldiers moving at randomly different speeds when not in formation.

3 hours ago, maroder said:

The solution to formations being useless would be the planned formation bonuses: https://trac.wildfiregames.com/ticket/3523

That's very possible. Also, I don't see any other way to differentiate the many avaiable formations. But to be honest, I'd like it more if formations had benefits by themselves, rather than being a liability that needs a hard bonus to be useful.

Edited April 25, 2021 by alre

[azayrahmad]

5 hours ago, alre said:

I don't think we should make retreating easier

Have we implemented directional attack bonus yet?

[ChronA]

31 minutes ago, azayrahmad said:

Have we implemented directional attack bonus yet?

I had a version of directional armor modded in for testing in A23. Would not be hard to merge it into A24, if no one else has a better version of the same concept already at hand. (Mine's kind of bare bones: just simulation side, no GUI.) 

[maroder]

3 hours ago, ChronA said:

I had a version of directional armor modded in for testing in A23

Sounds very interesting. A few questions: Do you apply directional armor from all attacks? Is it the same for all units or can you specify it in the templates? What exactly means directional; front, side, back or the exact angle?

If it is the same for all units it can have very different results, depending on the unit speed.

[azayrahmad]

5 minutes ago, maroder said:

Sounds very interesting. A few questions: Do you apply directional armor from all attacks? Is it the same for all units or can you specify it in the templates? What exactly means directional; front, side, back or the exact angle?

If it is the same for all units it can have very different results, depending on the unit speed.

I'm not sure how @ChronA implemented it, but I remember from @Angen's old Survival mod that it was quite simple. It uses angle difference between two units rotation and decide a tolerance/threshold value on which is considered side and back. You can check in this part of code. I adapted a part of this in my Morale mod.

Not sure about speed because IIRC unit stopped moving first before attacking, so I haven't implement charging yet.

All in all, I think this mechanism could help discourage retreating and kiting.

[maroder]

10 minutes ago, azayrahmad said:

but I remember from @Angen's old Survival mod that it was quite simple

Totally forgot about that one. And there are so many nice ideas implemented:

Attack and Armour: Dealing of damage is directional | Shield size defines how much damage unit blocks from front

Mercenaries: need to pay them else they turn to gaia

Rams: need to be garrisoned to be able to attack | can be captured

 

But anyway, my concern with directional armor (that is equal for all units) and unit speed is that slower units will be much more affected by this change. Retreating with cav would mean you can get out fine, but retreating with ptol pikemen would kill them. This can make sense, but it can also get op very quick or would require other balancing changes.

[wraitii]

I have doubts about directional attacks so long as then engine doesn't really care about rotation for units (and it doesn't, for the most part). That being said, perhaps I'm a bit pessimistic on this.

20 hours ago, ChronA said:

[...] This is certainly what we see in AoE2.

I won't lie, I'll finish that sentence with "and it works very well for them".

----

Realistically, I don't think we'll get to directional damage in A25, and I am unconvinced (and have been for years) by talks of more advanced formation combat and such. 0 A.D. has a fundamental economic aspect and any fight system will suffer from the 'continuous reinforcement' effect, which is _completely_ unrealistic and breaks down any comparison with history or indeed games like Total War. And I personally don't particularly care to try and change it. I like Age of Empires a lot.

Making retreating harder will simply make the fights even more snow-bally than they already are, and I fear that's not a great direction to go in right now.

[ChronA]

52 minutes ago, wraitii said:

22 hours ago, ChronA said:

[...] This is certainly what we see in AoE2.

I won't lie, I'll finish that sentence with "and it works very well for them".

Oh absolutely! You'll get no argument from me about that!
I'm just pointing out that if the intention is to make retreating/kiting punishable, then there are options (some of which are already in the official plan). However it is also perfectly legitimate to just have very weak chasing. In fact I point to AOE2 and Dawn of War 2 (and even SC2 to a degree) as games where that approach not only works but is core to their identities.

1 hour ago, wraitii said:

Realistically, I don't think we'll get to directional damage in A25

What makes you say that? The technical implementation is trivial (and you have at least one contributor--me--offering to put in the legwork) so I assume you mean that it's not a good time in the game's balance evolution to try such a massive change. However, are you really expecting there to be a better time down the road? It sounds more to me like they simply don't fit the vision anymore. If that's the case, then fair enough, but then what's the plan B?

[ChronA]

8 hours ago, azayrahmad said:

I'm not sure how @ChronA implemented it, but I remember from @Angen's old Survival mod that it was quite simple. It uses angle difference between two units rotation and decide a tolerance/threshold value on which is considered side and back.

Yes mine works much the same way: calculate the angle between the target unit's heading and the vector of the attack, and if that angle is greater than the threshold then calculate damage using "flank" armor values instead of the main armor values. I also made it so the flanking angle threshold and flank armor values can be customized for each unit in their templates.  

[wraitii]

1 hour ago, ChronA said:

What makes you say that? The technical implementation is trivial (and you have at least one contributor--me--offering to put in the legwork) so I assume you mean that it's not a good time in the game's balance evolution to try such a massive change. However, are you really expecting there to be a better time down the road? It sounds more to me like they simply don't fit the vision anymore. If that's the case, then fair enough, but then what's the plan B?

I don't really have an answer to any of these questions. All I know is that an RTS like 0 A.D. is a complex system, and this looks like it could have a lot of subtle impacts, and I'm not convinced it would actually make the game more _fun_, or even more interesting to play.

I'm just not very involved in gameplay at the moment (well, really, since forever), because I'm currently more interested by engine changes (including e.g. pushing, though I recognise it has gameplay impacts). I think the 0 A.D. gameplay is kind of coincidental, and we should mostly focus on improving what we have rather than changing it, because we don't have a very good plan on where to go if we were to do that, but we do have players that enjoy the gameplay right now.

[Thorfinn the Shallow Minded]

2 hours ago, wraitii said:

I don't really have an answer to any of these questions. All I know is that an RTS like 0 A.D. is a complex system, and this looks like it could have a lot of subtle impacts, and I'm not convinced it would actually make the game more _fun_, or even more interesting to play.

I think that this feature in moderation could add some nuance to the game.  In general ranged units probably could have little directionality to their armour by virtue of them rarely wearing much.  Instead it could be focussed on melee infantry, making things such as flanking them a relevant way of countering them.  A way I could see this particularly working well would be that a simple kiting manoeuvre would be suboptimal since the ranged unit would basically just strike in the direction with the most resistance. Instead it would be a matter of surrounding the force to hit their vulnerable sides.

[alre]

3 hours ago, wraitii said:

I don't really have an answer to any of these questions. All I know is that an RTS like 0 A.D. is a complex system, and this looks like it could have a lot of subtle impacts, and I'm not convinced it would actually make the game more _fun_, or even more interesting to play.

I'm just not very involved in gameplay at the moment (well, really, since forever), because I'm currently more interested by engine changes (including e.g. pushing, though I recognise it has gameplay impacts). I think the 0 A.D. gameplay is kind of coincidental, and we should mostly focus on improving what we have rather than changing it, because we don't have a very good plan on where to go if we were to do that, but we do have players that enjoy the gameplay right now.

I'd be fine if retreating would be as harmful as it already is (particularly when in a bad position), my point is that I fear the new pathfinder changes that, and thus I'm asksing for a counterbalance.

46 minutes ago, Thorfinn the Shallow Minded said:

I think that this feature in moderation could add some nuance to the game.  In general ranged units probably could have little directionality to their armour by virtue of them rarely wearing much.  Instead it could be focussed on melee infantry, making things such as flanking them a relevant way of countering them.  A way I could see this particularly working well would be that a simple kiting manoeuvre would be suboptimal since the ranged unit would basically just strike in the direction with the most resistance. Instead it would be a matter of surrounding the force to hit their vulnerable sides.

I think such a mod would be very interesting. Actually, among all proposals, this is the one that has both the most likely side effects, and the most interesting implications. Have we any testing program or schedule?

[wraitii]

13 hours ago, alre said:

I'd be fine if retreating would be as harmful as it already is (particularly when in a bad position), my point is that I fear the new pathfinder changes that, and thus I'm asksing for a counterbalance.

I'm not sure how much of a difference it makes, since it's also easier to chase units (and efficient kiting would already use formations).

15 hours ago, Thorfinn the Shallow Minded said:

I think that this feature in moderation could add some nuance to the game.  In general ranged units probably could have little directionality to their armour by virtue of them rarely wearing much.  Instead it could be focussed on melee infantry, making things such as flanking them a relevant way of countering them.  A way I could see this particularly working well would be that a simple kiting manoeuvre would be suboptimal since the ranged unit would basically just strike in the direction with the most resistance. Instead it would be a matter of surrounding the force to hit their vulnerable sides.

That could work, but it could lead to issues: javelineers, with their lower range, will have trouble flanking melee units. If we increase their damage to compensate, they'll become OP from the back. And/or it relegates them to being anti-ranged units. Flanking is mostly a cavalry thing. And how much would it change DPS in combat situation? Archers are already close to being OP, and this could make the more powerful (but it could also be a large nerf).

That being said, the anti-kiting potential is interesting. I'm not certain directionality is the way to go, but maybe it'd be a useful addition for this particular setup.
I'm not _opposed_ to the idea, as I do see how it could add tactical depth, I'm just quite wary of unintended side effects.

[Thorfinn the Shallow Minded]

2 hours ago, wraitii said:

That could work, but it could lead to issues: javelineers, with their lower range, will have trouble flanking melee units. If we increase their damage to compensate, they'll become OP from the back. And/or it relegates them to being anti-ranged units. Flanking is mostly a cavalry thing. And how much would it change DPS in combat situation? Archers are already close to being OP, and this could make the more powerful (but it could also be a large nerf).

That being said, the anti-kiting potential is interesting. I'm not certain directionality is the way to go, but maybe it'd be a useful addition for this particular setup.
I'm not _opposed_ to the idea, as I do see how it could add tactical depth, I'm just quite wary of unintended side effects.

Fair points.  As I would see it, javelinists would be the best ranged units equipped to face infantry head on since effectively fighting melee unit with them already is fairly involved.  Personally I would like seeing archer and slinger missiles doing pretty minimal damage head on, making their role one of softening melee units unless there are some units in place pinning them down, which would allow for them to surround the force.  Naturally that's just one approach, and undoubtedly it would have side-effects.  Those problems in my mind would be navigable.

[ChronA]

I haven't been able to try out this feature until now... but DEAR GOD!!! Did no one test this thing?! Units have effectively no collision now, and that's saying something because they never had very much to begin with.

This is what 200 units looks like. And they can even move like this. I assume you are familiar with the term death ball, well this is practically a death black hole.

Edited June 4, 2021 by ChronA

[wraitii]

49 minutes ago, ChronA said:

I haven't been able to try out his feature until now... but DEAR GOD!!! Did no one test this thing?! Units have effectively no collision now, and that's saying something because they never had very much to begin with.

This is mentioned in the very first post in this thread. I don't really like it, but I also don't think it's a very realistic situation outside of toy examples. I also don't expect it to be a particular balancing problem, but maybe I'm incorrect on that account.

The 'pushing distance' can easily be modified by changing pathfinder.xml if you want to test out other settings.

[ChronA]

 To clarify, I do not disagree with the feature itself. The much smoother pathing it enables is VERY nice, but obviously I have concerns about the gameplay implications. Admittedly it doesn't seem that much worse when compared to the deathball situation in a24 and a23... except, like I say the situation was already extremely bad IMO.

I will definitely play around with the pathfinding parameters a bit, and try to get a feel for what's possible.

[wraitii]

10 minutes ago, ChronA said:

Admittedly it doesn't seem that much worse when compared to the deathball situation in a24 and a23... except, like I say the situation was already extremely bad IMO.

Right. I don't see the A24 situation as that problematic either (or at least I don't really mind), so the change isn't horrible to me, merely unfortunate.

Note that it's still hardcoded in C++ that idle units are allowed to bunch up together more than walking units, I could make these settings different so we can tweak the parameters individually.

[maroder]

I actually don't mind it that much either. Sure, visually they are a bit clumped, but gameplay wise it seems fine to me.

[Lion.Kanzen]

2 hours ago, ChronA said:

I haven't been able to try out this feature until now... but DEAR GOD!!! Did no one test this thing?! Units have effectively no collision now, and that's saying something because they never had very much to begin with.

This is what 200 units looks like. And they can even move like this. I assume you are familiar with the term death ball, well this is practically a death black hole.

I was realizing that yesterday that I couldn't move a citizen.

[ChronA]

1 hour ago, wraitii said:

Note that it's still hardcoded in C++ that idle units are allowed to bunch up together more than walking units, I could make these settings different so we can tweak the parameters individually.

I would absolutely appreciate more parameters to customize pathing behavior. Pathing is one off the most powerful factors in establishing the unique feeling of an RTS. It is the most basic way you interact with the game world. So even relatively small changes in behavior can have a huge impact on the flavor of a product (for good or ill).

Given that the 0 AD project aspires to be not just a single game, but a platform for making ancient warfare games, I think embracing gameplay-differentiability is desirable. I would be much more interested in Delenda Est and Hyrule Conquest as stand-alone experiences if they didn't appear to play so much like straight reskins/expansions of EA. Tweaked pathfinding settings would help them to distinguish themselves as their own things.

1 hour ago, maroder said:

I actually don't mind it that much either. Sure, visually they are a bit clumped, but gameplay wise it seems fine to me.

Well I can't hold habituation against you. But ultra dense unit packing does cause problems for gameplay and balance:

1. Choke points have less gameplay impact. This limits the ability of map designers to finesse away certain balance issues, and reduces gameplay diversity. (See this situation for an example of choke points not working: https://www.youtube.com/watch?v=XPuNEQEomGQ&t=2229s).
2. Same story with attack surface-area and concavity.
3. It also reduces combat readability and control, which makes it harder to used mixed unit compositions effectively and reduces opportunities for players to distinguish themselves with individual unit micro.