Follow up react to: Could lower collision size fix pathing?
Vložit
- čas přidán 26. 06. 2024
- Follow up to this video: • Could lower collision ...
Live Stream also on @ / theviper
Main Channel: / theviperaoe
MoreTheViper: / @moretheviper
TheViper Merch ► teespring.com/stores/theviper
Donate ► theviper.live/
Discord ► / discord
Twitter ► / theviperaoe
Instagram ► / theviperaoe
Facebook ► / theviper
Business Inquiries: theviperbusiness@gmail.com
TheViper | Professional Age of Empires 2 Definitive Edition Player for GamerLegion | Sponsored by Re-Bo - Hry
Reducing collision size may change the dynamics of vil fight in dark age. Maybe as we age up to castle age we can have reduced collision size as it will help with boom.
just a short question, and i might embarrass myself with this idea since i'm fairly new to the topic. but why are we even talking about "collision size"?
when villagers don't continue moving after being blocked due to collisions, that isn't a collision size issue. it's merely a pathing issue, and a "memory" issue (not RAM, but villager memory of where they were tasked to go).
meanwhile, the collisions on camps could be fixed by just making them "react": so when someone bumps into another, does the other then usually just stay there? talking about realism, the other person should consider moving a little bit as well. of course, when 5 people would have to move to fix this issue, this would mabybe even create bigger pathing problems and deadlocks, but implemented right, this would fix ANY kind of collision problems, and wouldn't introduce many more other problems.
and this would also mean that villagers could NOT easier escape when attacked by MAA for example, as their collision size wouldn't change, and the enemy forces wouldn't "react" to them getting closer or colliding with them.
also, what i can see from the clip at 2:15, this isn't a collision problem either. the problem seems to be that after a collision, the villager reevaluates their path and then moves accordingly, not taking ANY other villagers in their close proximity into account, besides the one that they've just collided with. they seem to try and "not hit the same guy again", however meaning that both villagers then take VERY different paths than optimal. from my understanding just from this video, the villagers seem to not have a "smooth line" to arrive at their final destination. when you send them around a corner, they're likely to have like 3-4 straight lines and then just move to specific target points directly.
those multiple target points that they have when you send them around a corner are calculated during pathing, and keeping that number low makes it easy and fast to process, even when the amount of units is huge.
and that's the problem already! if that number was bigger, and the villagers would have 30-40 of those points where they'd move in a straight line, they would move around each other much smoother.
however, this would increase load, most possibly on the servers and on the clients respectively. and that is what the devs don't seem to want to do. and maybe they have good reasons for that as well.
so sure, making the collision boxes smaller decreases this problem as well, but it's a cheap fix which won't fix all the issues either. increasing the number of target points in a path however would mean that the villagers would need to "think" more to get to their destination, and this would increase load.
in that case it would be important to have efficient algorithms, which are smart enough to differentiate between a huge amount of villagers in a small space, where they collide often (and where they'd need that additional destination points, and so CPU load), and other parts, where not so many villagers pile up, and so the amount of those points can be kept fairly low.
and from what i can judge, such a thing is not in place right here. be free to correct me here, but i suppose we could enhance the pathfinding AI in that way, and possibly even lower the CPU load in many cases as well through good optimizations.
Rewriting the AI spaghetti code from 25 years ago is a mess, there's a reason they haven't done it. One small change and it all falls apart.
Slightly altering collision sizes is a simple fix to a very complex problem. It's not going to be perfect, but it doesn't need to be, it just needs to be better without being worse in some other way.
It would be awesome if there were different % for "traveling" vs "working" - so that villagers don't stack up on the same spot when mining (looks wrong) but have no problems navigating around obstacles/other units.
Radius reduction upon age-up! New ages means more vills, usually. So 10% reduction each advance in age should work fine
Reducing vills colision box would change everything bc it would be more vills able to gather resources, more vils building i.e. wonders, more vils surrounding a enemy unit, more vills repairing or destroying buildings etc.
A change like this would disrupt the entire game economy of actions and resources.
Couldn’t we just increase the collision size of militia and men at arms if that’s a concern? I like the sound of reduced collision size for vills. It won’t fix everything, obviously, but it would be a small quality of life change.
Not that im against it, but increasing efficiency will lead to quicker uptimes, new build orders, new meta and EGAALOS HUUVERING THE EKONOMY AT MINUTE 12
You can have different collision sizes for different unit interactions!! Do 20% reduction for villager-villager collisions, and the rest remains utterly unchanged
Bumping this topic for the devs.
Do I just have my rosy tinted glasses on, or was userpatch just... good? Maybe they need to hire Scripter haha
Scripter gave us the holy scripture that was userpatch 1.5, and we have only tarnished his legacy
Just increase the base attack speed of man at arms and maybe +1 base attack, they should run these tests and let us experiment it. If get to overpower, just reset in the next update patch, keep thinking in something else and life goes on!
You have to do it "outside of main stuff" IMO. There is redbull grind and you dont want to mess with crazy stuff right before it ends
Am I crazy here? I feel like collisions were never an issue several years ago (AoK, AoC, HD, and first few years of DE). I took a break from AOE for a couple years and came back about a year ago to find the collisions VERY noticeable. Did the devs change something to make them worse? I feel like we're proposing solutions to a problem that never originally existed.
Yea, maybe im tripping, but I feel like when DE released there was no problems with stuck vils, at least not close to this frequency.
Also i cant remember this when I was a child, so I guess you are right. Overall I feel like pathing just became worse and worse with every patch.