Bots and Simulations > Single store

Singularis(SS)(Commander Keen)

(1/1)

Commander Keen:
Making SS bots is harder than I thought!


--- Code: ---'Singularis by Commander Keen
'A single store tie feeder that isn't terribly efficient
'I commented out the reproduction gene, because it ends up killing itself
'I have considered commenting out the slime generation, maybe shell would be more useful, or neither
start
1 *.robage 10 mod sgn sub 20 *.velsx sub mult
1 *.robage 3 mod sgn sub 99 mult add
.robage 2 mod sgn add
1 *.robage 11 mod sgn sub 50 mult add
'1 *.robage 13 mod sgn sub 50 mult *.nrg 1000 sub 0 floor sgn mult add ' They just kill each other
1 *.robage 7 mod sgn sub 500 *.slime sub mult add ' never makes right amount, could comment out
1 *.robage 3 mod sgn sub .sharenrg mult
1 *.robage 10 mod sgn sub .sx mult add
*.robage 2 mod sgn .tie mult add
'1 *.robage 13 mod sgn sub .repro mult *.nrg 1000 sub 0 floor sgn mult add 'They just kill each other
1 *.robage 7 mod sgn sub .mkslime mult add 'Never makes the right amount, could comment out
1 *.robage 11 mod sgn sub .aimdx mult add
store
stop
--- End code ---

Moonfisher:
Yeah, making SS bots is damn hard.
I finaly finished Brevis, took ages, I thought to myself, I'll just try and SS'ify Excalibur real quick...
SS'ifying Excalibur is not something you just do "real quick", hard to rethink a lot of stuff and coordinate and prioriize....
In the end the condition blocks filled out more than a page each... it was getting impossible to debug.
So... thought of a way to hunt bugs more effectively, and it's a good way to be certain that no actions are overlapping before you start to convert the code.
SG'fy the entire code before I sart to put it together, and gotta test while assempling it... all the time...

I wrote the code I used for debugging Brevis in the same post, it's very usefull for making a SS bot.

bacillus:
I would never be mad enough to try make one of these, but by the looks of it, the slime/shell feature is one best left ommited, because with SS bots you not only have to think in % of nrg, but also % of time used up on each command. The time could be better put to use making more ties or feeding or reproducing. Maybe a cascading priority system is the bast way to go.

Oooh, this sounds like fun. I imagine no bot in the SS league would be particularly strong, just efficent beyond all other leagues.

Numsgil:
It does have a certain alluring madness to it

Like staring in to the sun

Moonfisher:
Well, it's not like you need to beat anything to get in the league...

cond
start
0 0 store
stop

Theoreticaly this is a SS bot that would be able to get a league spot
Basicaly saying you can submit your evolved zero bot and it'll get a spot in the league even if it doesn't survive for long
So the bar is still very low for the SS league, if you can make a bot able to survive with 1 store you'll already have beaten nano... actualy a zero bot beats nano....
What I'm saying is... start small... and if you're having problems debugging, just get the debug code from Brevis (Posted as a reply to the topic) I would never have gotten that thing to work propperly without that code (Or it would atleast have taken 10 times longer).

Navigation

[0] Message Index