Poll

What do you like doing most in RA2/RA3/robot combat games?

Building robots
Local single player battles with manually-controlled robots
Local multiplayer battles  with manually-controlled robots (PvP with controllers or a shared keyboard)
AI-only tournament battles
Other?  Please comment below.

Author Topic: Robot Rumble 2.0 - Robot Combat Simulator - Under Development  (Read 187485 times)

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1120 on: December 19, 2019, 11:15:54 AM »
oh yea, the update did something to self collision so this is fun i guess

Darn it.  I thought I fixed this.

I guess I need to look at it again.  Would you mind sending the .RR2Bot file?
Sending atm

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1121 on: December 19, 2019, 11:17:49 AM »
Ok so some feedback:
1. Flame pit instantly breaks off parts

Found and fixed the problem.  Thank you guys for pointing it out.  I feel like I should have caught it. *sheepish*

2. Im not a fan of the damage. Even at 50% parts fall off too fast and it makes not difference than on 200%. You dont even need to hit a bot with the weapon to break his part off

Good to know.  I might retune the slider to give a 25% option.

3. for the damage. Maybe you could make an option to couple up few smaller parts to turn it into a bigger part. this way you could avoid losing a fair bit of parts because one small part that was holding the weapon together reached 0hp

This might be tough to do.  I will have to think about it quite a bit and ask @tashic to see what he thinks.  It would require additional stuff in the .RR2Bot file.  It would also require new UI. 

It also fundamentally affects the design of the game.  Would it always be optimal to just make the entire robot one group so that nothing ever falls off?

Following that train of logic, what if component destructibility isn't possible at all.  Is the game better if the entire robot is just one big bag of hit points?  Everything works perfectly until the robot is dead?

Offline KupaTec

  • Antweight
  • Posts: 63
  • Rep: 6
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1122 on: December 19, 2019, 11:20:11 AM »
    I'm not too sure the damage slider is working when set to 0% as parts still seem to fall off but again that could be down to the flame pit mentioned earlier
Okay, after a bit of testing I can say the damage slider is working apart from the fire pit issue, however it seems that even on 200% that bot chassis are now immune to damage completely. House robots still take damage correctly but custom bots cannot be killed now, is this expected behaviour?

EDIT:
Following that train of logic, what if component destructibility isn't possible at all.  Is the game better if the entire robot is just one big bag of hit points?  Everything works perfectly until the robot is dead?
I think component destructibility is a major feature that needs to stay in the game, however the current handling could use tweaks. One thought I had is to have each tree of components form a bag of HP and once that is depleted then the entire tree falls off? Somewhat similar to the previous updates handling, but using the total tree health instead of just the weakest parts.

Would it always be optimal to just make the entire robot one group so that nothing ever falls off?
Unfortunately there are always going to be people that optimise the fun out of any system, but I don't personally think that should be a reason not to do something.

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1123 on: December 19, 2019, 11:27:22 AM »
Good to know.  I might retune the slider to give a 25% option.
The main thing is that you can hit an opponent with a wheel, and the part may fall off. Everything that is not a weapon prolly shouldnt deal damage.

It also fundamentally affects the design of the game.  Would it always be optimal to just make the entire robot one group so that nothing ever falls off?

Following that train of logic, what if component destructibility isn't possible at all.  Is the game better if the entire robot is just one big bag of hit points?  Everything works perfectly until the robot is dead?
Ok this one is kinda hard i guess, ofc i think with the way damage works people will need to relearn how to build. The problem is that the smaller the part is the less durable it is, which is okay, the issue however is that it seems like the scaling is off as small parts even on 50 thickness steel fall too off easily, Maybe make only weapons groupable?

As for the damage multiplier, i think you should also add 5% and 10% because 25 might still be too high

Offline KupaTec

  • Antweight
  • Posts: 63
  • Rep: 6
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1124 on: December 19, 2019, 11:53:13 AM »
The main thing is that you can hit an opponent with a wheel, and the part may fall off. Everything that is not a weapon prolly shouldnt deal damage.
[snip]
As for the damage multiplier, i think you should also add 5% and 10% because 25 might still be too high

I have built a "Fire Pit Fix" for the current version and now at 50% and even 100% damage slider it feels right for my bots at least. Something about the way the fire pit was dealing damage (mostly a lot of it every frame) was causing issues with durability. I think that is what was causing the problems with the lighter parts health being seemingly out of scale also.

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1125 on: December 19, 2019, 12:07:38 PM »
Another issue is that the custom wheels fall off by themselves overtime. This is a shame really, as lot of my bots use custom made wheels


i think that the main culprit and issue is self damage, which breaks stuf. I was hitting an opponents chassis with my axe, and the axe fell off, he did not have a weapon at all

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1126 on: December 19, 2019, 12:38:17 PM »
Another issue is that the custom wheels fall off by themselves overtime. This is a shame really, as lot of my bots use custom made wheels


i think that the main culprit and issue is self damage, which breaks stuf. I was hitting an opponents chassis with my axe, and the axe fell off, he did not have a weapon at all

I'm really struggling with this conceptually, and could use some help with this:

1. Objects should be damaged because they absorb kinetic energy (or impulse, or whatever).  This could be from a spinner hit, but also from just naturally bumping into stuff.  Case in point: a flipper launches a robot 2 meters in the air.  The robot lands.  It breaks into 2 pieces.

2. The advantage of saying that "only weapons can cause damage" is that is a very clearly understood idea for players.  The disadvantage is that it means that you can give literally anything you create a "weapon" tag, resulting in everything damaging everything.  This puts us right back where we started.

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1127 on: December 19, 2019, 12:51:26 PM »
1. Objects should be damaged because they absorb kinetic energy (or impulse, or whatever).  This could be from a spinner hit, but also from just naturally bumping into stuff.  Case in point: a flipper launches a robot 2 meters in the air.  The robot lands.  It breaks into 2 pieces.
The problem is, i was just driving the bot, i did not bounce into everything nor i was hit by a spinner/flipper/axe. Bot bouncing may have also been a culprit

2. The advantage of saying that "only weapons can cause damage" is that is a very clearly understood idea for players.  The disadvantage is that it means that you can give literally anything you create a "weapon" tag, resulting in everything damaging everything.  This puts us right back where we started.
Maybe a weapon could only be parts that are not rubber and they are only on motors/pistons/servos.
Maybe also a function where of you pressed f12, you could see what is a weapon (parts would be highlighted green), and what is not. This would be helpful for hosts i guess

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1128 on: December 19, 2019, 12:58:48 PM »
1. Objects should be damaged because they absorb kinetic energy (or impulse, or whatever).  This could be from a spinner hit, but also from just naturally bumping into stuff.  Case in point: a flipper launches a robot 2 meters in the air.  The robot lands.  It breaks into 2 pieces.
The problem is, i was just driving the bot, i did not bounce into everything nor i was hit by a spinner/flipper/axe. Bot bouncing may have also been a culprit

2. The advantage of saying that "only weapons can cause damage" is that is a very clearly understood idea for players.  The disadvantage is that it means that you can give literally anything you create a "weapon" tag, resulting in everything damaging everything.  This puts us right back where we started.


Maybe a weapon could only be parts that are not rubber and they are only on motors/pistons/servos.
Maybe also a function where of you pressed f12, you could see what is a weapon (parts would be highlighted green), and what is not. This would be helpful for hosts i guess

1. Maybe the answer to this is that a minimum amount of kinetic energy needs to be dissipated in order for damage to occur.  This should prevent random vibrations and bumping around from doing anything.

2. Giving things a "weapon" tag would definitely help with player understanding.  I like the idea of pressing the f12 key (or something similar on the mac keyboard) to highlight things.  It doesn't address the damage caused by being launched into the air though.  Should a big fall cause damage?  What about getting launched sideways into the arena wall at 10 meters/second?

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1129 on: December 19, 2019, 01:03:44 PM »
1. Maybe the answer to this is that a minimum amount of kinetic energy needs to be dissipated in order for damage to occur.  This should prevent random vibrations and bumping around from doing anything.
I like the idea, this way, the vibrations/bouncing wont hurt thr bot. Also maybe at 0hp, you'd need to reach a certain force of KJ to rip the part out too!

2. Giving things a "weapon" tag would definitely help with player understanding.  I like the idea of pressing the f12 key (or something similar on the mac keyboard) to highlight things.  It doesn't address the damage caused by being launched into the air though.  Should a big fall cause damage?  What about getting launched sideways into the arena wall at 10 meters/second?
Atm, Flipper damage should not be touched till the damage itself is sorted. As for the launching stuff, i guess internal damage would be the best, however if you would hit the side/floor at a high speed with a spinning weapon, then i guess external damage would work there

Offline powerrave

  • Giga Heavyweight
  • Posts: 5372
  • Rep: 2
    • View Profile
    • Awards
  • See profile for gamer tags: Yes
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1130 on: December 19, 2019, 03:02:11 PM »
I've downloaded and started up the new version for a short bit. May be trying to build stuff with mirroring now in place.

What I did notice is an issue with the forward heading. Some of my bots (namely Stopping Power and Miasma) have it at 180 (or -180, doesn't matter), but every time the botlab is entered with one of them, the forward heading gets reset to 0.
"Always be yourself, unless you're a loser"


Offline Badnik96

  • tired of your shit
  • *
  • Posts: 17537
  • Rep: 3
    • Badnik96GTM
  • Awards BOTM Winner
    • View Profile
    • BattleBots Wiki
    • Awards
  • See profile for gamer tags: Yes
  • Skype: Badnik96
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1131 on: December 19, 2019, 04:21:43 PM »
i feel like what needs to be done (if possible) is to find a certain energy threshold where damage will begin to occur. two robots bumping into each other at low speeds typically doesn't cause damage, and neither do many low-power attacks with weapons. if there's a way to make it so all damage calculated below this threshold is zero, i think that should be the way to go.

weapons shouldn't always cause damage because then a heavy bar dragging slowly across a chassis will do quite a lot more damage than in real life, and no one likes that.

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1132 on: December 19, 2019, 05:51:12 PM »
I've downloaded and started up the new version for a short bit. May be trying to build stuff with mirroring now in place.

What I did notice is an issue with the forward heading. Some of my bots (namely Stopping Power and Miasma) have it at 180 (or -180, doesn't matter), but every time the botlab is entered with one of them, the forward heading gets reset to 0.
Alternatively, you can rotate the bot. Select the chassis and do a 180. Heading to 0 and voila!

Also chassis scaling is broken. Cyar will tell whats wrong

Offline CyarSkirata

  • Antweight
  • Posts: 114
  • Rep: 0
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1133 on: December 19, 2019, 06:07:17 PM »
My Panic Attack has been functionally destroyed by this update. :(
The chassis turns into a janky mess when the bot is being simulated, and the weapons put a ridiculous amount of torque on the whole thing. It's so bad it can't even self-right.

If you want to compare it how it is now to how it's supposed to work, its fight with Flipster in the first Parsec Rumble 1.0 video is a good example.
(Just opening that bot file in an earlier version would kinda work, but I de-wheelfixed the weapons in the hope that it would at least fix the torque issue. It didn't.)
Stuff I Done Competed In:

Offline Badnik96

  • tired of your shit
  • *
  • Posts: 17537
  • Rep: 3
    • Badnik96GTM
  • Awards BOTM Winner
    • View Profile
    • BattleBots Wiki
    • Awards
  • See profile for gamer tags: Yes
  • Skype: Badnik96
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1134 on: December 19, 2019, 07:21:06 PM »
also while i'm here, can i request the ability to run a weapon off of multiple motors? this would help tremendously with trying to make more compact robots.

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1135 on: December 19, 2019, 08:52:33 PM »
also while i'm here, can i request the ability to run a weapon off of multiple motors? this would help tremendously with trying to make more compact robots.

Were you thinking two motors, both turning the same belt?  You can request it, but it really complicates the component assembly tree.  I.e. it isn't really a tree structure any more.  The .RR2bot file and component assembly is currently built around a tree structure.  Each component has exactly one parent component.  With the possibility of two parents, things become much more complicated.

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1136 on: December 19, 2019, 09:28:32 PM »
My Panic Attack has been functionally destroyed by this update. :(
The chassis turns into a janky mess when the bot is being simulated, and the weapons put a ridiculous amount of torque on the whole thing. It's so bad it can't even self-right.  [ Quoting of attachment images from other messages is not allowed ]  

If you want to compare it how it is now to how it's supposed to work, its fight with Flipster in the first Parsec Rumble 1.0 video is a good example.
(Just opening that bot file in an earlier version would kinda work, but I de-wheelfixed the weapons in the hope that it would at least fix the torque issue. It didn't.)

I have an old version of Panic Attack from October 21st, as well as the new version you just sent me.  The old version seemed to work okay.  For some reason the flipper's limit angles weren't set properly in the new version.  Once I set the limit angles, things seemed to work okay.

My computer is really struggling with frame rate at the moment, but once I can get that under control I will try to get a video showing both versions of Panic Attack going head-to-head.

Edit:  Here's a gif showing how it is working in for me.  Please excuse the low resolution.  I had to turn things way down because my laptop is chugging:


Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1137 on: December 20, 2019, 12:33:46 AM »
I had to turn things way down because my laptop is chugging:
You wanna know chugging? 4 highly detailed bots in one arena, lagging out a 6 core pc, although it was the gpu that was struggling

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1138 on: December 20, 2019, 05:39:04 AM »
I had to turn things way down because my laptop is chugging:
You wanna know chugging? 4 highly detailed bots in one arena, lagging out a 6 core pc, although it was the gpu that was struggling

I'm noticing that too.  All of the robots I build are really simple.  No decoration, just basic blocky parts to see if the basic ideas are working correctly.  When I pulled up the second version of Panic Attack my memory usage went through the roof and the game slowed to a crawl.

I have a few ideas that might help:

1. Put each robot on its own collision layer during combat and disable all self-collisions within that layer.  I have a hunch that this should fix Crabsolutely Clawfull, Beetle Crab, Son of Ziggy, and any other robots whose motor armatures have things attached with overlapping colliders. (EASY FIX - can be done today)

2. Get the painter working.  Having an entire separate component dedicated to every piece of paint on the robot is just crazy: awesome looking, but computationally expensive.  (HARD FIX - will take weeks/months of building out the painting system)

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1139 on: December 20, 2019, 05:44:39 AM »
I had to turn things way down because my laptop is chugging:
You wanna know chugging? 4 highly detailed bots in one arena, lagging out a 6 core pc, although it was the gpu that was struggling

I'm noticing that too.  All of the robots I build are really simple.  No decoration, just basic blocky parts to see if the basic ideas are working correctly.  When I pulled up the second version of Panic Attack my memory usage went through the roof and the game slowed to a crawl.

I have a few ideas that might help:

1. Put each robot on its own collision layer during combat and disable all self-collisions within that layer.  I have a hunch that this should fix Crabsolutely Clawfull, Beetle Crab, Son of Ziggy, and any other robots whose motor armatures have things attached with overlapping colliders. (EASY FIX - can be done today)

2. Get the painter working.  Having an entire separate component dedicated to every piece of paint on the robot is just crazy: awesome looking, but computationally expensive.  (HARD FIX - will take weeks/months of building out the painting system)

I like idea 1 as temp idea
I kinda want to fix the damage stuff. Badnik has really suggested something good