Loading...
  OR  Zero-K Name:    Password:   

Physics simulation vs balance

30 posts, 905 views
Post comment
Filter:    Player:  
Page of 2 (30 records)
sort
3 years ago
Related to Brackman's overshooter widget, and ignoring the short term discussion, how much do players and devs care about/like "Physically simulated units and projectiles" (copy paste from landing page, and perceived by me as a major selling point of Zero-K)?

More specifically for most ratio's of unit and projectile speed, there will be a "retreat range bonus", because that is just physics (this is mostly due to the fact that projectile's have the speed comparable to the speed of units, not hundreds of times faster - I doubt for example that phantom has noticeable retreat range bonus).

I see two extremes to the opinion range: either "let's forget about physics simulation and focus on balance" or "let's embrace physics simulation and adjust accordingly" (ex: projectile speed can be increased as compared to unit speed which will make the effect much less noticeable, but might require different visual effect - like smoke behind projectile rather than projectile).

I understand the tactical decisions to just rule some things unfair/block/police them and considering the few development resources focus on what could be easily improved, but still think it is relevant to at least talk about ideal long term goals.
+0 / -0
3 years ago
quote:
More specifically for most ratio's of unit and projectile speed, there will be a "retreat range bonus", because that is just physics

Lol as if.

+5 / -0


3 years ago
A more appropriate title would be "Physics simulation vs fun". Balance is pretty flexible. We had a pretty well balanced game with retreat range bonus and we have decent balance now. What matters is whether the balance and physics combine to create something enjoyable.

My approach to physics for ZK is to take what seems fun and discard the rest. The fun stuff is exaggerated to make it chunky and salient, and is often what makes discarding the rest necessary. Physics is a tool, or maybe a foundation, for creating nuanced unit interactions underwritten by things we're pretty familiar with, ie the behaviour of stuff flying around in the world. Visuals and mechanics align to give units more weight and to make interactions more intuitive. What you see is what you get. It isn't that Zero-K aims to have physics or to simulate physics particularly accurately. Rather, physics is used to achieve other goals.

If Zero-K, or similar games for that matter, had full physics then I expect either:
  • The physics would have to be neutered to the point of being boring, at which point why even have physics?
  • The game would be janky to the point of unplayability, at least for the types of quick and/or casual games you can play in ZK.

Take Ravager for example. It moves at 40% the speed of its projectile. As unknownrankTinySpider points out, Zero-K does not have relativity. Full physics implies relativity. So a full physics Ravager would have significantly more range in its direction of travel. I haven't figured out the maths, but my intuition says the range boost would be significantly greater than 40% (for (1/2)mv^2 reasons). So a full physics game is a game of constantly morphing range ellipses as units move and rotate. A game of Ravagers jittering forwards and backwards to toss their projectiles at distant Stingers.

What is the solution to such a janky physics Ravager? Neuter the physics. Make Ravager shoot at more 'realistic' velocities so that the range bonus is negligible. But few things can reasonably dodge such a projectile, so why even have the physics in the first place? This is the paradox of game physics, at least unless the player has direct control of every manifestation of their will in the world (like in a FPS).

An even harsher restriction on the current Ravager is the imposition of range rings. Ballistic weapon have a range limit but many of them could in theory shoot further. However, without range rings there would be a fixed relationship between projectile speed and range, which constrains unit diversity. Similarly, homing missiles tend to be able to home for quite a bit further than they can shoot. This is a matter of unit AI. Imagine how frustrating it would be for units to keep wasting shots at units ducking in and out of the edge of your units physical missile ranges. As a general rule, more complicated physics generates more ways for unit AI to screw up in frustrating ways.

In terms of ideal long term goals, it all depends on what is fun and below the too janky threshold. In theory, stuff like Reaver overshoot could be solved by giving it a maximum fuse and reworking the unit AI in the engine so that it aims and fires with the goal of dealing some minimum fraction of its AoE damage to the target. The fuse could even depend on the angle of fire (for some weird reason) to replicate the current range behaviours. But I don't see anyone doing that, and something should be done now.

On the retreat range bonus, it still exists on all eligible weapons except for Glaive, Bandit, Scorcher, Dagger and Pyro. Try a Ravager duel. The retreat range bonus was removed by giving those units no leading, so that they fire at everything in range (like a Lotus), and then adjusting the projectile once it is fired to add the leading back in. This works because projectiles have a little bit of overshoot, they don't disappear at exactly the edge of their range. Removing retreat range bonus for Ravager is probably infeasible and would at least look ridiculous.

This reminds me of feedback conversations with supcom/PA people about the retreat range bonus. They said it was something weird that was added onto the game. I asked them why it didn't happen in the games they were familiar with and they didn't know. I'd be interested to know the answer because it must be there unless it was explicitly removed. If other games have anything as extreme as Ravager then its existence, or the method of removal, should be clear after some investigation.

Retreat range bonus was removed from raiders because it was a trap, confused people coming in, and was a big unit AI failure. Losing to equal forces with 0% damage dealt based on quick micro tricks feels pretty bad. The fun of mastering it and destroying incoming players did not seems sustainable. So it wasn't that physics gave way to balance, more that it game way to overall fun. And it wasn't even "real physics" because in real physics Glaive vs. Glaive wouldn't have a retreat range bonus due to relativity.
+5 / -0
3 years ago
The video strengthen the point. Imagine two trucks going one after the other, (truck A "retreating" from truck B) and both releasing a ball ("shooting", with speed 0). Now, do you claim that truck B will NOT hit truck A ball?

The "not natural" thing is that the bullet in ZK just "disappears" after a given range and as far as I know has constant speed (in real life it would decrease speed and fall on the ground). If you would want better physics modelling you would model the decreasing bullet speed (hence lower damage at higher distance), but that would not affect at all the retreat "bonus" (even make it larger as the impact speed of the bullet with the following vehicle will be bullet speed + vehicle speed, while with the retreating vehicle the impact speed will be just bullet speed)
+0 / -0
3 years ago
Cross-posted with GoogleFrog, so answering that.

I think that many times physics in ZK allowed unexpected strategies that were not planned (ex: ramps, space jacks, terraforming to protect against some weapons, needing line of sight for lances, etc.). I still play ZK after all these years because it feels dynamic and always new. I am afraid that the idea of restricting physics for fun could reduce dynamics on long term. Maybe improving physics simulation would give more strategies that can work in some situations (versus declare "A is fun, B is not fun"). Ex: for the retreat advantage, you get an advantage just if units chase you, the chasing units can go after a mobile target (ex: mex, building, etc.).

quote:
Full physics implies relativity.

Don't think ZK models friction (or does it?). I think the effect of friction comes before relativity. Friction at different air speeds will be very different (larger friction higher the speed). So I think the "vehicle following" will "hit" harder the bullet.

quote:
A game of Ravagers jittering forwards and backwards to toss their projectiles at distant Stingers.
Which is a different strategy which makes sense for slow moving projectiles. Not sure if it is good or bad, but for sure I never saw in a strategy game. Should we not have line move because nobody else does it? (now it is painful for me to play anything else because of that :-p).

quote:
So a full physics game is a game of constantly morphing range ellipses as units move and rotate.
Statement sounds incomplete to me. You mean for weapons where the vehicle speed is close to projectile speed, and both high enough to matter. For example that already is the case with lance, no? Do you argue that lance ray should be "stationary" (if the vehicle speed would be ignored)? And for phantom you would not care (for example).

quote:
Ballistic weapon have a range limit but many of them could in theory shoot further.
Not sure I get it. How is crabe/emissary spire a thing then?...


quote:
However, without range rings there would be a fixed relationship between projectile speed and range, which constrains unit diversity.
Not sure I get it. Which units are currently more "diverse" due to the effect you mention? (maybe I get better what you say)

quote:
But I don't see anyone doing that, and something should be done now.
Sure, I perfectly agree with you, this is more about objectives, "it would be nice if", rather than next 6 months. I would love to have the water flowing thing that I worked on included somehow, but I waited on engine changes, other stuff came up and that's life... But I think it is good to think of ideas sometimes.

quote:
The retreat range bonus was removed by giving those units no leading, so that they fire at everything in range (like a Lotus), and then adjusting the projectile once it is fired to add the leading back in.
Would increasing projectile speed (by 100x) and making another visual effect be another solution? (again not saying to implement it, just curious)

quote:
because in real physics Glaive vs. Glaive wouldn't have a retreat range bonus due to relativity
Depends how you see it. The impact speed will be different for the 2 Glaives. (one is vehicle speed - bullet speed, for the other is vehicle speed + bullet speed). If you get close to the speed of light I guess those two values will get closer. But for newtonian speed it matters. If glaives go with 20km/h and bullet has 25km/h, do you think an impact at 5km/h is different from one at 45km/h?...
+0 / -0
Significant air resistance could add some retreat bonus to a full physics game because a bullet fired at a lower speed relative to the air will lose less speed relative to the moving units by the time it hits the chasing unit. But this still all depends on having exaggerated weapon properties.
+1 / -0


3 years ago
quote:
This reminds me of feedback conversations with supcom/PA people about the retreat range bonus. They said it was something weird that was added onto the game. I asked them why it didn't happen in the games they were familiar with and they didn't know. I'd be interested to know the answer because it must be there unless it was explicitly removed. If other games have anything as extreme as Ravager then its existence, or the method of removal, should be clear after some investigation.

My understanding of it, having played zero seconds of PA and Supcom, but having watched a few videos and having discussed it with people under influence of alcohol is that these games do go for high projectile velocity to unit velocity ratio, which makes any retreat advantages insignificant. Supcom also seems to have more "ReAlIsTiC" ranges in comparison to ZK, so things like accuracy and leading can still matter - at the cost of making the game devolve into icon wars much faster.
+0 / -0
3 years ago
quote:
The video strengthen the point.

I am ever impressed by the ability of people to just solidify their beliefs in face of contradictory evidence.
+0 / -1
unknownrankTinySpider

quote:
I am ever impressed by the ability of people to just solidify their beliefs in face of contradictory evidence.


Can you explain how that example is contradicting the claim because I don't see it as a slam dunk.

The only thing the video demonstrates is that a cannon firing a projectile at the same speed as a vehicle is moving in the opposite direction will make that projectile immobile horizontally. If I drop a bowling ball from an overpass and you drive a car under it at freeway speed, do you think it won't hurt you just because the bowling ball has no lateral velocity?

The claim is that shooting backwards gives you more range than moving forward at the same speed and shooting in the same direction. If they're firing at each other the fleeing vehicle is moving away from the projectile aiming at it, but the chasing vehicle is moving towards the projectile.

Under normal circumstances it wouldn't matter because relative to the speed of the projectile, the vehicles might as well be immobile. But in ZK bullets are VERY slow, at least compared to how you would expect realistic military weapons to function. It's just not realistic to take one step backward and suddently a bullet doesn't reach you anymore.

I actually don't know, but it sounds like fleeing would be an advantage here.
+0 / -0
3 years ago
CArankGalamesh
quote:
Can you explain how that example is contradicting the claim because I don't see it as a slam dunk.

quote:
there will be a "retreat range bonus", because that is just physics

quote:
<video>

quote:
The video strengthen the point.
+0 / -1

3 years ago
Yes, that is also what I read. Your response does not address my question.
+0 / -0
3 years ago
You mean your tangential question that in no way relates to a range bonus?
+0 / -0

3 years ago
I think both are related. I always assume that there's something I don't know, which is why I asked.

I was thinking about this and the only thing I see that could change the outcome is friction. The vehicles being still or moving wouldn't change anything if there is no atmosphere. Unless I'm missing something, for the fleeing vehicle to have an adventage, it would have to be moving fast enough that the chasing vehicle's projectile would slow down enough due to friction to fall short.

It could happen, but maybe it's not likely?
+0 / -0

3 years ago
quote:
The video strengthen the point. Imagine two trucks going one after the other, (truck A "retreating" from truck B) and both releasing a ball ("shooting", with speed 0). Now, do you claim that truck B will NOT hit truck A ball?

Obviously truck B will at some point hit the ball via driving over it but this has more to do with how the ball is big and does not expire rather than any range effects via movement speed.
+0 / -0
quote:
The vehicles being still or moving wouldn't change anything if there is no atmosphere.

What are you talking about? Do you think the physics demonstrated in the video are due to atmosphere?
+0 / -0
quote:
Obviously truck B will at some point hit the ball via driving over it but this has more to do with how the ball is big and does not expire rather than any range effects via movement speed.
I am just commenting on the example given. So, based on the example the "follower" truck gets "hit".

I think the example is bad, as probably what it would like to say is (but not sure): in vacuum it does not matter if vehicles move or not (they will both hit each other). I commented thinking implicitly about the more daily life physics (where the drag - force opposing movement - is not linear but squared for high speeds, so the retreating vehicle will get some advantage). I imagine ZK having an atmosphere, but maybe that is just me. It is a good point and worth discussing (do we have or not an atmosphere? how much we want to approximate? what are the parameters? etc.), but it's kind of hard to do that if you start just with "lol as if" and let the others wonder what you meant (at least someone else did not get the idea: CArankGalamesh)...

(I think) the root cause is that speed projectile in spring is constant and relative to the ground, and not constant and relative to the vehicle it fires. The not intuitive problems arises much more due to the small ratio between projectile speed and vehicle speed. In real life, it does not matter if you fire a bullet while moving fast forward. The bullet has muzzle velocity of 5000km/h, so your added speed will make a small difference (plus the drag will probably make it even smaller). But in ZK the speeds are so close that firing from an accelerating vehicle will give an elliptic "range" (that would be a fun thing to try).
+0 / -0
3 years ago
Can one of you explain to me why you keep mentioning air friction and drag? It's ok to admit you're wrong instead of writing walls of text that try to obfuscate the fact.
+0 / -2
quote:
quote:
The vehicles being still or moving wouldn't change anything if there is no atmosphere.

What are you talking about? Do you think the physics demonstrated in the video are due to atmosphere?


I think he means that as long as there is no effect of air friction, in physically simulated case, for two vehicles shooting each other, there would be no difference between the situation where both vehicles are stationary and situation where both are moving in same direction and speed.

1. In stationaty case:
Vehicle A shoots ball at speed V. Ball hits stationary vehicle B at relative speed V.

2. In moving case.
Vehicle A escapes from vehicle B, in other words vehicle B is chasing vehicle A.
Both vehicles move at speed V.

Vehicle A shoots ball at relative speed V. But since vehicle A itself moves at speed V in the other direction, the absolute velocity of ball becomes V-V = 0, as demonstrated in the video.
But since vehicle B moves at speed V toward vehicle A, it will hit the ball which is now in its way. The relative speed of the collision will again be V as in previous case.

3. Now lets say the chasing vehicle B shoots ball as escaping vehicle A.
It shoots ball at speed V relative to itself. Since it shoots in the same direction as it is moving the absolute speed of the ball with be V+V=2V.
The ball will hit the escaping vehicle A at absolute speed 2V, but since vehicle A moves in same direction the relative speed will 2V-V=V, so exactly the same as in previous two cases.

Thus in physically simulated case (but without air friction) there is no difference and no retreat advantage.

quote:
Can one of you explain to me why you keep mentioning air friction and drag? It's ok to admit you're wrong instead of writing walls of text that try to obfuscate the fact.


Because in physically simulated world air friction would provide retreat advantage.
Lets look again at cases 2 and 3, but with air friction having effect.
In case 2 the ball has zero absolute speed so it won't be affected by air drag and will hit the chasing vehicle with same relative speed V.
But in case 3 the ball has absolute speed 2V so it is affected by air friction and would lose speed as it moves, eventually hitting vehicle A with absolute speed less than 2V, so with relative speed less than V, doing less damage.
Thus with air friction the chasing vehicle can't hit escaping vehicle as hard as escaping vehicle can hit the chasing vehicle.

+4 / -1
https://zero-k.info/Forum/Post/243736#243736
quote:

Take Ravager for example. It moves at 40% the speed of its projectile. As TinySpider points out, Zero-K does not have relativity. Full physics implies relativity. So a full physics Ravager would have significantly more range in its direction of travel. I haven't figured out the maths, but my intuition says the range boost would be significantly greater than 40% (for (1/2)mv^2 reasons). So a full physics game is a game of constantly morphing range ellipses as units move and rotate. A game of Ravagers jittering forwards and backwards to toss their projectiles at distant Stingers.

This is incorrect. One easy way to think about it: imagine the Ravager is stationary and target is moving at the Ravager's maximum velocity.
On flat terrain, the maximum range will still describe a circle, however it won't be centered around the unit, if it is moving.
+2 / -1
The equivalence of the stationary and the moving case as PLrankRafalpluk explained is the most important thing.

By considering gravity, projectiles effectively get a finite life time due to their independent vertical movement. Their range is a rotational paraboloid that is shifted by (their velocity minus the target velocity)*projectile life time which depends on height. In the vertical cross-section, the range is still a circle (not elliptic) shifted forwards for stationary targets, but not at all for targets moving at the same velocity. So still no retreat bonus.

In physics, a small retreat bonus comes from air resistance (not only to damage, but also range). In ZK it came from deviations to physics: The best candidate is that velocities might not add up but are constant relative to ground as FRrankmalric said. So, ranges are not shifted for stationary targets, but are shifted backwards for moving targets. How has the retreat bonus actually been removed? By adding movement speeds to projectile speeds or by shifting ranges forward artificially?
+0 / -1
Page of 2 (30 records)