Finaly 3.0 is somewhat out (and stuff)

[Botsareus]

Thx Num , but I think I'll post my own "starter-bot" to

This is the robot is started with:
http://s9.invisionfree.com/DarwinBots_Foru...p?showtopic=155

This robot does "better" then the robot above:
http://s9.invisionfree.com/DarwinBots_Foru...p?showtopic=156

Its the evolved virsion. But it has a glitch: If it is fighting its not mutating self , the not mutating self wins 99% of the time.
 :help:

  :evil: :bigginangel:  :evil:

[Botsareus]

Here I posted my bots, help if you can;

I also wanted to add that I have 256mb Ram and 1.50Ghz, also: My system is not cluged up by usless programs when I run DB.

Endy, the problem is when there is a change in the max number of vegs. in the invirment the problem above happens automaticaly, I dont even have to wait until generations evolve to cause the problem. 300 vegs on high resolution means 75 vegs on low resolution , so that change is addapteble: its only from 75 --> 25 vegs. A robot should be able to evolve to survive under new settings, why it does not is what I am trying to figure out.

[Numsgil]

Bots,

what kind of frame rate are you getting for your sims?  How many robots do you run in your usual simulations?

[Botsareus]

Ok this is what I rimember from the 300 vegs experiment. In about 45 minutes it ran 100000 cycles. 300 vegs feed about 50 FirstBots.

Num, my speed is fine, dont get off topic please. Just beleave me I run the simulations long enugh and fast enugh. At higher resolutions just clickying the "turn display off" dramaticaly inproves performance.

Num, Shvartz, Do you understand the way I mutated this bot4g? I did not just leave a simulation running. As Num sayed: Watching a simulation for 3 days does not produce "better" bots.
Shvartz Because then you will understand how I figured out that mutation rates do not work.

[shvarz]

Bots, when you say 4g is "better", what do you mean by that?  How is it better if it is losing to the original bot?

[Botsareus]

You must of done some really funny settings shvartz...

[Botsareus]

Shvartz try doing F1 conditions , and disable all mutations.

BETTER QUSTION: What settings did you use shvartz?

[Botsareus]

Num , Shvartz left but you still here , what to do with all this info,. what do you think? mutation rate bugs , the problem with bot4g , etc.

[Numsgil]

It may be that the unmutated version is better against other bots, while the mutated version is used to living in a situation where there are no competitors.

Also, 4 generations is tiny tiny.  100 generations would be the minimum before I'd expect to see mutations racking up to actual evolution.  On the scale you're talking about you're just seeing random change, without an opportunity to sort out the good from the bad.

[shvarz]

OK Bots, here is what I think happened (I have to guess, cause you don't say it):

You started some sim (not F1 settings) with First bot.  let it run for a long time and it evolved into 4g.  Then you put them into F1 mode and First bot wins.  What did you prove?  That your 4g bot is adapted to the settings that you used to evolve it, not to F1 settings.  Exactly what you would expect.

[Botsareus]

The problem is that the mutated virsion does better agenst other bots and agenst the unmutated virsion.

"On the scale you're talking about you're just seeing random change, without an opportunity to sort out the good from the bad. " I am sorting the good exactly form the bad because my robot is supposed to evolve agenst [you]itself [/you]with mutations turned off. not agenst the robot I started with or the [you]unmutated[/you] robot. (one copy of the robot evolves the other one does not, when the copy that does not evolve dies out I save the new produced robot as a new generation, I already tryed to explain this in earlyer posts.)

non-mutating is not unmutated in any way.

[Botsareus]

"Then you put them into F1 mode and First bot wins."  FirstBot does not win shvartz, how are you getting this results? what are your settings?

"You started some sim (not F1 settings) with First bot." If I did that then first bot would not win anyway. There is some kind of bug/law that says even: non mutating clone does better then mutating clone in different settings.
It says also: If mutating clone has defult mutation rates, non mutating clone does worse then mutating clone in same settings but  If mutating clone has  mutation rates used by itself from the lest run, non mutating clone does better then mutating clone in same settings.
and: non mutating clone does worse then mutating clone in same settings for only 4 runs , after that: non mutating clone does better then mutating clone in same settings (clone from run 4 in run 5)

[Numsgil]

Okay, I finally understand what you're doing.

You're loading up two copies of the same species.  One has mutations enabled in the mutations panel (that is, they're set to non zero values).  The other copy has mutations disabled by pressing the disable mutations button in the mutations panel.

Right?

That either species is dying out doesn't make much sense, especially after so few cycles.  The answer to why the mutating version can't beat the non mutating version in F1 mode is simple: the mutated version must not be the thing killing off the non mutating version.

So the question then becomes what is killing it off, and why can the mutating version survive it?

[Botsareus]

"let it run for a long time and it evolved into 4g." no shvartz , I toled everyone to read post carfully it did not just evolve into 4g in one simple run. Its a complex system I run mutating clones agenst not mutating clones of the same robot. Then take the resulting robot and repeat the prosses . Above I listed bugs/laws that prevent me from going further in different settings or with better mutation rates or after about 4 runs.

'What did you prove? That your 4g bot is adapted to the settings that you used to evolve it, not to F1 settings. Exactly what you would expect. " , Well if its playing agenst itself weather it is addapted or not it should do better or the same. Thats why I want you to give me the settings shvartz were firstbot (witch is supposebly 4 runs worse then bot4g) beats bot4g.  Post the numbers for settings shvartz

[Botsareus]

The answer to why the mutating version can't beat the non mutating version in F1 mode is simple: the mutated version must not be the thing killing off the non mutating version.

In true F1 mode mutations are disabled , In my mode the mutating virsion has mutation rates: 2mofm the rest are 200m

what I mean by mofm is the mutation rates of mutation on top in the form.

P.S.
 :blink: Now I am bigging to realize how hard it is for people to undertand me , dam , $#% what to do? go back to elementry school $#% ? , ...