Remove spread a.k.a randomness

[raynor]

In the light of the complains about increased turn rates that prevent dancing, I would like to understand if we considered simply removing spread for ranged units. It seems to me that in that case dancing would not be an issue, and we could consequently go back to normal turn speed. In my opinion, it would also reward better micro management of units.

[Yekaterina]

Well, removing spread means even more op archers, slingers and skirmisher. I can make a little mod for you to try it out.

If all ranged units are buffed then we need to improve the melee.

 

[ChronA]

Do you mean removing the random spread element from projectile targeting, or are you talking about removing ballistic projectile simulation for ranged attacks entirely? (The latter would be like the system used in the StarCraft games, where ranged attacks launch a homing projectile that are guaranteed to hit .) It's an important distinction because zero-spread projectiles can still be dodged by dancing.

[raynor]

@Yekaterina I would be down for you to do a mod, and I think I will also start looking into modding. For sure in this case we would need to nerf ranged units damage, and/or buff melee damage/pierce armour

@ChronA honestly I am out of my depth, but yes I was thinking of a SC/SC2 system where we would remove the ballistic projectile simulation. I understand this might be controversial, but it seems to me that it could solve several issues at once (dancing and turn speed) at the coast of something less realistic for sure. Still, dancing is very toxic imo and turn speed has shown very problematic in A24

Edited May 12, 2021 by raynor

[Yekaterina]

3 minutes ago, raynor said:

@Yekaterina I would be down for you to do a mod, and I think I will also start looking into modding. For sure in this case we would need to nerf ranged units damage, and/or buff melee damage/pierce armour

@ChronA honestly I am out of my depth, but yes I was thinking of a SC/SC2 system where we would remove the ballistic projectile simulation. I understand this might be controversial, but it seems to me that it could solve several issues at once (dancing and turn speed) at the coast of something less realistic for sure. Still, dancing is very toxic imo and turn speed has shown very problematic in A24

I can remove spread and buff all melee. How do you suggest I buff them? But it would also mean that ranged units are killed very quickly in battles, because every shot hits them. Valihrant made a mod that solves turn speed. I can increase them to 18 radians per second if you want

[Yekaterina]

raynor.zip

This removed all spread. 

Haven't changed melee yet

[Dakara]

Hello,

i test the mod tonigh . Maybe we have to nerf a little unit big range like arrow and slingers for balance.

[Yekaterina]

27 minutes ago, Dakara said:

i test the mod tonigh . Maybe we have to nerf a little unit big range like arrow and slingers for balance.

The thing is, a well trained slinger in real life can throw stones up to 300 meters. Archers can shoot up to at least 100 metres. 

But sure, I can make the changes. Just tell me the values you want to use. 

[alre]

@Yekaterina does that only remove the spread or does it also assures that every shot hits?

I personally think that the best part of random spread is its rendering, wich looks better than 100% precise projectiles. But that's not the only thing, I also think that it's a good thing that ranged units deal more damage from closer, so if we remove spread, I'd also make damage dealt dependent on distance.

All in all I'm in favour of such a rework, I think it would make the game simpler and more consistent.

Just now, Yekaterina said:

The thing is, a well trained slinger in real life can throw stones up to 300 meters. Archers can shoot up to at least 100 metres. 

But sure, I can make the changes. Just tell me the values you want to use. 

Archers can shoot much further than that. We know it well enough because archers were used in good numbers in more recent times (english longbows). We have some conflicting sources on which weapon had longer range, it's generally speculated that slings were somehow superior in that regard, but most probably they achieved fairly similar results.

[Yekaterina]

Personally I prefer having some spread.

Consider projectile mass, drag coefficient, kinetic energy of projectile at launch point...

Arrows are lighter, so they fly faster but also experienced more air resistance. Stones have smaller kinetic energy but large mass + not really affected by air resistance.

[LetswaveaBook]

On the topic of bow range, if you shoot an arrow for a very long distance, I guess it will lose a considerable amount of its energy to air resistance. So that means its effective range might not be that high.

I like the spread feature as a gameplay balance. Spread means that archers can shoot far, but at large distance melee troops are fairly safe from them because archers are more likely to miss. I think it hurts gameplay if substantial damage can be done at long range and therefore I like inaccuracy at long range to be in the game.

[Grapjas]

In regards to the OP, i dont think having 100% accuracy will counter dancing if we had the old turn rates.

The actual problem is the amount of time it takes for the projectile to arrive (but this should not be higher imo). Players can instantly walk the other way when projectiles are still in the air.

The thing is, archers where primarly used as a weakening/demoralizing tactic and in great numbers, guaranteeing hits because of the great volley of arrows. The game does this pretty well imo. Missing the main target when firing with a group to a group will likely hit the person next to him instead.

Making friendly fire (against own units, not ally) a toggleable game option could be interesting as a balance to units though. But thats a different subject i guess.

[LetswaveaBook]

6 minutes ago, Grapjas said:

Making friendly fire (against own units, not ally) a toggleable game option could be interesting as a balance to units though. But thats a different subject i guess.

I like this suggestion for sake of realism. However I wouldn't like if your ranged units target an enemy automatically and then automatically decide to inflict more damage to your own troops than to the opposing troops.

[raynor]

I was under post limitations policy but @Stan` just bailed me out :D.

@Yekaterina Thanks for the mod, I forgot to say but I am away from home until end of next week most likely, so I won't be able to test it before that time, however I am fully trusting @Dakara 

 

I definitely understand that my proposal is less realistic but the dancing or current options to mitigate it are terrible gameplay downside in my opinion, and a SC2-like projectile mechanics would kill dancing without any other compromises. I understand now that my original idea is actually to remove the projectile simulation, not only removing the spread. I also like @Dakara idea of decreasing damage over range, but it seems to me that might be more work. Would it be possible to simply remove the simulation and make projectiles always land with a mod, or do we need to touch the cpp engine to do so?

[Grapjas]

4 minutes ago, LetswaveaBook said:

I like this suggestion for sake of realism. However I wouldn't like if your ranged units target an enemy automatically and then automatically decide to inflict more damage to your own troops than to the opposing troops.

Yes there should be an 'hold fire' / 'fire at will' button too.

Edited May 12, 2021 by Grapjas

[raynor]

@Grapjas @LetswaveaBook I think friendly fire for all units would be insane, and would force people into pure ranged vs ranged units which would be a drama imo. But for the sake of keeping things into scope, please let's open another thread for this topic

[Grapjas]

3 minutes ago, raynor said:

Would it be possible to simply remove the simulation and make projectiles always land with a mod

That's definitely doable with a mod and simple to do aswell. The only problem would be a visual one, it will look like the projectile missed, when it was actually registered as a hit.

[Freagarach]

17 minutes ago, raynor said:

or current options to mitigate it are terrible gameplay downside in my opinion

And please note that making projectiles always land (i.e. making them cruise missiles) can be considered terrible gameplay downsides as well. (_At least_ by me and other people I play with.)

[raynor]

Why do you think this would be bad in term of gameplay? From a competitive aspect of the game, it would make more sense imo. As I see things, the more randomness, the less competitive the game is.

Moreover, as we saw in a24 turn time has a significant part in making rush hard to execute, no turning time implies dancing etc so this randomness leads to other gameplay issues. But I also realise that I am really biased by my previous SC/SC2 gaming experience, so glad to hear other opinions like yours

[LetswaveaBook]

1 hour ago, raynor said:

As I see things, the more randomness, the less competitive the game is.

While individual arrows have random hit chances, it does not impact a game with many arrows. The law of large numbers says that if many arrows/slingshots are shot, the amount of hits is extremely likely to be close to the average.

Edited May 12, 2021 by LetswaveaBook
@Yekaterina, CLT is cooler as LLN does not meantion the square root of the sample size.

[Yekaterina]

Central limit theorem

[Stan`]

1 hour ago, raynor said:

Why do you think this would be bad in term of gameplay? From a competitive aspect of the game, it would make more sense imo. As I see things, the more randomness, the less competitive the game is.

Because @Freagarach is not into competitive gameplay, at all. And sadly what makes sense for competitive gameplay doesn't always make sense for the other players.

[ChronA]

(LetswaveaBook and Yekaterina with the ninja!)

2 hours ago, Grapjas said:

That's definitely doable with a mod and simple to do aswell.

Yeah, after some thought, I guess you are right. Just put a binary override onto the collision test of MissileHit() in DelayedDamage.js. Best practice would probably be to add an "always hit" tag to ranged weapons that would control this behavior. That way, if for instance someone wanted to make a Myth style total conversion using the game assets, they would have an easy time re-enabling projectile simulation.

2 hours ago, Grapjas said:

The only problem would be a visual one, it will look like the projectile missed, when it was actually registered as a hit.

I don't think updating the visualizations is really necessary. Individual projectiles are hard to see anyway, and clear misses should be pretty rare if projectile speeds are fast enough. It would be much weirder to see arrows bending to track their targets.

Edited May 12, 2021 by ChronA

[raynor]

Quote

While individual arrows have random hit chances, it does not impact a game with many arrows. The law of large numbers says that if many arrows/slingshots are shot, the amount of hits is extremely likely to be close to the average.

It is however not true for smaller group, and this also indirectly rewards turtling because smaller groups attack will most likely be less successful, since randomness will play a higher role in this case (basing my self on your analysis on archer spread). I am pushing it a bit, I have to admit ;). Removing simulation projection in favour of something like SC2 might be not the right solution then. But I still don't see any way to avoid either long turn rates or dancing with current approach.

 

Edited May 12, 2021 by raynor

[LetswaveaBook]

1 hour ago, raynor said:

I am pushing it a bit, I have to admit ;).

Indeed, you are pushing it a bit. Randomness can give both above and below average results. You view randomness as a disadvantage whereas it can also give results better than expected.

 

If you have 70% chance of hitting an enemy (which I estimated to be about the same chance that a camel will hit something from 60 when firing at a group) and your group fires 60 shots. We can calculate probabilities with the binomial distribution. Then you will hit 42 shots on average. There is only 3.6% chance that you hit 35 shots or less. You have 76% chance to land at least 40 shots. The chance to hit with 47 shots or more is about 10%. So the randomness is fairly mild.

Good execution plays a way more important role than randomness.