Author Topic: Successes...and Failures?  (Read 6904 times)

Offline asterixx

  • Bot Neophyte
  • *
  • Posts: 48
    • View Profile
Successes...and Failures?
« on: April 27, 2008, 07:08:37 PM »
This is quite possibly a repeat topic, sorry for taking up space. (just refer me to the actual topic and delete this one if that's the case   )

My question is about the "Find the most successful robot" feature. Is this functioning properly? If so, what are the criteria for highlighting the bot in question? How is 'success' being defined here? Could there potentially be a "Find the least successful robot" feature in the future?

Lastly, and I think I've read this somewhere on these forums already but Im really not too sure; Are the DNA scripts functional or are they also going to be implemented in future versions? I ask this because my Zerobot sim is in full swing now and it would be wonderfully helpful to have someway of searching through this mess of bots.

"Tell a man there are 300 billion stars in the universe and he'll believe you.  Tell him a bench has wet paint on it and he'll have to touch it to be sure."

Offline shvarz

  • Bot God
  • *****
  • Posts: 1341
    • View Profile
Successes...and Failures?
« Reply #1 on: April 28, 2008, 01:14:09 AM »
Currently the best bot is defined as the bot with the most offspring. Not a perfect definition of "best", but anything more complicated would be ... more complicated and subjective.
"Never underestimate the power of stupid things in big numbers" - Serious Sam

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Successes...and Failures?
« Reply #2 on: April 28, 2008, 01:25:57 AM »
Right.  The Best Bot code also excludes veggies...

A "find worst" bot option would mean what exactly?  Define what this is and it could of corse be implimented.  Oldetst bot without offspring perhaps?  

I've never played with scripts.  Give it a try.  I'll fix any bugs you find...



Many beers....

Offline asterixx

  • Bot Neophyte
  • *
  • Posts: 48
    • View Profile
Successes...and Failures?
« Reply #3 on: April 28, 2008, 06:35:43 PM »
Firstly, I noticed that it selected a most successful bot that had 0 offspring over one that had 4 offspring...That's why I was confused. In any case, it doesnt really bother me all that much.

Quote
Oldest bot without offspring perhaps?

Yes, something like that. It could potentially allow for some interesting comparisons, and might help the user to understand what exactly went "wrong", so to speak.

"Tell a man there are 300 billion stars in the universe and he'll believe you.  Tell him a bench has wet paint on it and he'll have to touch it to be sure."

Offline bacillus

  • Bot Overlord
  • ****
  • Posts: 907
    • View Profile
Successes...and Failures?
« Reply #4 on: April 28, 2008, 06:46:46 PM »
I think a bot's total nrg (counting offspring, nrg, body, slime, shell, poison, venom) should define the best bot function.
"They laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown."
- Carl Sagan

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7742
    • View Profile
Successes...and Failures?
« Reply #5 on: April 29, 2008, 01:58:25 AM »
Quote from: bacillus
I think a bot's total nrg (counting offspring, nrg, body, slime, shell, poison, venom) should define the best bot function.


 This has always been my thought as well.  Was this ever implemented, or is it still the bot with the most kids?

Offline Moonfisher

  • Bot Overlord
  • ****
  • Posts: 592
    • View Profile
Successes...and Failures?
« Reply #6 on: April 29, 2008, 05:41:26 AM »
Hmmm sounds like there would still be loopholes... I agree that most kids and kills will end up selecting the worst bot in the sim (The one who makes small offspring and eats them)...
But if you just add up all that stuff then bots that make shell/poison/venom with no limit will seem realy good... but they're actualy very wastefull...
Can't you check for living offspring somehow ? I think that's the most important thing a bot can have... live children.... live grandchildren are even better...
In the end when it comes to evolution successfull reproduction is the key... a living grandchild is a strong sign that this a bot is doinf well... where a lot of shell usualy means it broke it's shell gene and soon it will be too heavy to move and die from starvation...

Offline Gambit

  • Bot Builder
  • **
  • Posts: 52
    • View Profile
Successes...and Failures?
« Reply #7 on: April 29, 2008, 11:18:51 AM »
lol most kills.

how about the total energy harvested since birth - total energy expended. when it eats kids that wouldnt double. it'd just be getting back the energy it put into them.

I think that works out...

Offline Moonfisher

  • Bot Overlord
  • ****
  • Posts: 592
    • View Profile
Successes...and Failures?
« Reply #8 on: April 29, 2008, 12:01:21 PM »
With energy gain you still have a problem with bots that don't reproduce... they always end up big with a lot of energy.
Maybe if you found the bot with most energy/body/shell/venom/asf, then picked a non mutated offspring... problem is most offspring will be mutated if you have high mutations...
It just seems like measuring most values would end up finding a bot with a broken repro gene, and looking at repro will often find a bot who's very good at eating it's own offspring before they get away... just seems like there needs to be a way to make sure the selected bot can atleast reproduce successfully or it won't matter how strong it is...

Offline asterixx

  • Bot Neophyte
  • *
  • Posts: 48
    • View Profile
Successes...and Failures?
« Reply #9 on: April 29, 2008, 05:46:14 PM »
I agree Moonfisher. A very successful bot might also be able to limit the amount of energy it absorbs from the environment (T_Preservans is a good example). Doesnt measuring values that are outside the scope of reproduction become a problem of subjectivity?

Quote from: EricL
I've never played with scripts.  Give it a try.  I'll fix any bugs you find...

As for this, I think the main issue I've been having is that when I set "DNA Scripts Enabled" checkbox, and run the simulator, then go back and check the settings, the box is unchecked.
"Tell a man there are 300 billion stars in the universe and he'll believe you.  Tell him a bench has wet paint on it and he'll have to touch it to be sure."

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Successes...and Failures?
« Reply #10 on: April 29, 2008, 06:01:10 PM »
Quote from: Numsgil
Quote from: bacillus
I think a bot's total nrg (counting offspring, nrg, body, slime, shell, poison, venom) should define the best bot function.

This has always been my thought as well.  Was this ever implemented, or is it still the bot with the most kids?
It's the bot with the greatest number of extant descendents, be they direct or indirect.
Many beers....

Offline Shasta

  • Administrator
  • Bot Destroyer
  • *****
  • Posts: 231
    • View Profile
Successes...and Failures?
« Reply #11 on: April 30, 2008, 12:12:36 AM »
Would it be possible to let us define our own "best" bot algorithm?

 By this I mean something like the program parsing a text file that we create using some key words (mostKills, mostOffspring, mostNrg... ect), operators and numbers. Then defining if we want the lowest or highest result when run for the bots.

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Successes...and Failures?
« Reply #12 on: April 30, 2008, 12:14:40 AM »
Quote from: Shasta
Would it be possible to let us define our own "best" bot algorithm?

 By this I mean something like the program parsing a text file that we create using some key words (mostKills, mostOffspring, mostNrg... ect), operators and numbers. Then defining if we want the lowest or highest result when run for the bots.
Anything is possible.  It's all just priorities and time.  Create a new post in the Suggestions forum and I'll add it to the list.
Many beers....

Offline gymsum

  • Bot Destroyer
  • ***
  • Posts: 215
    • View Profile
Successes...and Failures?
« Reply #13 on: April 30, 2008, 10:25:49 AM »
Quote from: EricL
Quote from: Shasta
Would it be possible to let us define our own "best" bot algorithm?

 By this I mean something like the program parsing a text file that we create using some key words (mostKills, mostOffspring, mostNrg... ect), operators and numbers. Then defining if we want the lowest or highest result when run for the bots.
Anything is possible.  It's all just priorities and time.  Create a new post in the Suggestions forum and I'll add it to the list.


Im not sure if thats wise, your work load has you coding until Janurary 20010, this will only push you back further

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Successes...and Failures?
« Reply #14 on: April 30, 2008, 11:07:16 AM »
Quote from: gymsum
Im not sure if thats wise, your work load has you coding until Janurary 20010, this will only push you back further
Same as it's ever been.  I only said it would go on the list, not that I would necessarily get to it but one never knows.  Soem things go quick and maybe one day I'll get some help...  
Many beers....