weakest selection in draws

HDeffo
Neural Network
Neural Network
Posts: 212

weakest selection in draws

Post#1 » 15 Dec 2016, 05:07

Pier4r has been bugging the forums for a while now asking how gladiabots chooses target selections in case of a draw. After 3 days of testing on every possible map in every possible enemy positioning I could come up with.. in every possible scenario I believe I have finally found the solution for him and everyone else! :D (You are welcome). Select nearest target that fits the selection filter. If two or more enemies are of equal distance and match the selection filter however, the one with the lowest degrees of global rotation from the bot will be selected. It will choose an equal distance bot at global 156 degrees before a bot at 178 degrees. Global rotation means 0 is always north not forward.

pier4r
Skynet
Skynet
Posts: 3390

Re: weakest selection in draws

Post#2 » 15 Dec 2016, 07:32

Thanks, I put it right away in the FAQs.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

User avatar
NullPointer
Autonomous Entity
Autonomous Entity
Posts: 539

Re: weakest selection in draws

Post#3 » 15 Dec 2016, 09:36

Are you sure about it? Sounds very unlikely that this was implemented this way. My guess is that the first bot in the list of candidates would be picked up, which means the instantiation order would dictate who gets picked up first.

HDeffo
Neural Network
Neural Network
Posts: 212

Re: weakest selection in draws

Post#4 » 15 Dec 2016, 09:41

NullPointer wrote:Are you sure about it? Sounds very unlikely that this was implemented this way. My guess is that the first bot in the list of candidates would be picked up, which means the instantiation order would dictate who gets picked up first.


I thought this at first too and spent a day compiling a list of "orders" before i realized none of the orders made sense then i tried having bots switch positions which caused my bot to target a different one

User avatar
GFX47
Dev
Dev
Posts: 2914

Re: weakest selection in draws

Post#5 » 15 Dec 2016, 09:51

The entity selection process actually sorts the targeted entities using the following methods and keeps the first one:

  • weakest/strongest X:
    • lowest/highest X value
    • in case of equality lowest/highest X ratio
    • in case of equality closest (see below)
  • weakest/strongest X ratio:
    • lowest/highest X ratio
    • in case of equality lowest/highest X value
    • in case of equality closest (see below)
  • closest/furthest:
    • closest/furthest
    • in case of equality instantiation order (consider it as random)
* X = health or shield

I agree the last resort sorting is not satisfactory but don't have a better solution yet.

HDeffo
Neural Network
Neural Network
Posts: 212

Re: weakest selection in draws

Post#6 » 15 Dec 2016, 10:15

Are you sure there isn't a bug then? Because if it was instantiation order a lot of my tests involving switching which AI targets the enemy arent making any sense....the easiest map to test this is on meeting point. Different bots will choose different targets even when they're all at the same distances

User avatar
NullPointer
Autonomous Entity
Autonomous Entity
Posts: 539

Re: weakest selection in draws

Post#7 » 15 Dec 2016, 11:25

I raised this concern specifically about Meeting Point, I suspect that the instantiation order of the bots is not 100% symmetric between sides. I've seen a similar behaviour to the one you described where on the right side of the screen, the bot attacked is the one closer to the resources and on the left side the one further from the resources, even though the AIs were mirrored perfectly.

It's not that easy to reproduce this specific scenario, but it's not difficult to see the same AIs not drawing in this map.

Return to “Strategies”

Who is online

Users browsing this forum: No registered users and 1 guest