The Nineteenth Byte

Doorknob

12:17 AM

Actually, the spec said Only wolves that are submitted as an answer will be included. It never said "All answers must be represented by 'W'" or anything.

undergroundmonorail

1:32 AM

@Doorknob And "wolf" is defined as "animal represented by a W" so your point continues to not actually work

Doorknob

@undergroundmonorail The point is that it says "The only wolves will be submissions," but it never says "The only submissions will be wolves." But this argument doesn't really have a point anymore at this point :P

Geobits

Completely pointless, so let me add... To completely clear things up, I suggest the following: "Only wolves (submitted as an answer) will be included."

I think that covers both readings fairly well :p

ProgrammerDan

Roses are red and violets are blue?

non-sequitors are fun!

Geobits

So are sequiturs!

Quincunx

3:15 AM

@Rusher Lion-follower is now finished. :-). In my tests, AlphaWolf now always gets 10+ lower than LazyWolf.

Rusher

I just ran Lion-Follower all by itself and it ended with 7 living....

@Quincunx Is the sole purpose to destroy Lazy?

Rusher

4:02 AM

I wish SE had private messages...

Quincunx

4:39 AM

@Rusher Well, other wolfs use Lion-swapping techniques. So the purpose was to destroy them :-). And it destroys Alpha much more than Lazy.

@Rusher I didn't answer because I was offline, in-real-world.

@Rusher Yup. They don't care how many survive, as long as they get to follow lions :-)

And it's not called Lion-Follower. It's called WhenIGrowUp.

@Rusher What? Only 7 living? I ran it ~10 times and I average 30 living.

Or did you mean with no other animals?

When that happens, I do get 7 living

~~~~~~~~

Something I noticed is that after somewhere around 10-50 generations, the speed significantly increases (without the Thread.sleep(100); ofc). Is this just the JIT Compiler kicking in?

It fits the bill.

zenflux

No, it's just the population rapidly thinning, suddenly most of the less-smart wolves are dead when they run themselves into others

I believe most deaths occur cycle < 20 or so

Quincunx

5:01 AM

@zenflux False. Try running with only Lions.

This is a clear case of the JIT Compiler optimizing.

zenflux

@Quincunx Sorry, I'm not set up at the moment to test it, but try running many full runs in a loop, if it's the JIT then the speedup will only happen on the first go around, subsequent runs will be constant speed

Quincunx

@zenflux This is the JIT. It takes a while for it to kick in though. I can think of no other thing that would increase the speed that much. It is running in a loop, for 1000 times. After the first ~20 generations, the speed speeds to a fast constant speed.

zenflux

@Quincunx Yes, I know, but then does the speedup occur again on the next first 20? If so it's not the JIT.

run1000() //tests if speedup at g = 20

Quincunx

????

zenflux

for(int i=0;i<10;i++) {

run1000() // does that speedup occur every time?

}

Quincunx

5:09 AM

Oh, let's see

zenflux

Some days I'm more literate in code than english :P

And today is one of those days...

Quincunx

Lol

The speedup is always at ~100 (I decided to count this time)

And it is still fast when running tests again.

After running through once and resetting the board, it still keeps the fast speed.

zenflux

@Quincunx Hmm, that's decent evidence I guess, there should really be a VM flag or something to tell it to log JIT activity

Would be nifty as hell for benchmarks

Quincunx

Would also be nice if I could just tell it to immediately use the JIT from netbeans. Would be useful for things like this.

zenflux

Or ya know, just static compilation... to bad it would be slower than the JIT when it's warmed up.

...and unportable, and less cool-factor, etc etc

Quincunx

5:21 AM

Static compilation. I'd love it. Guess I need to press forward in learning C++

zenflux

Yeah I've been keeping C++ in "maybe later" pile, and it will remain there for the forseeable future.

Quincunx

That's the same problem I have.

except I've already written a few programs

zenflux

I wrote a distortion VST cause I needed it, but that's as far as I've ventured

reddit.com/r/programming/comments/21uyk5/darkest_corners_of_c/…

Woot! My automatic test suite just returned that my wolf it's tweaking is winning!

Quincunx

????

zenflux

Anyone else play around with automated wolf AI optimization?

DoorknobChatbot

5:31 AM

New post! My daughter's alphabet

zenflux

It runs tests (average of 100 runs of 1000 iterations) while tweaking some constants, looking for the best numbers for the AI to use

I'm breeding a superior wolf! Mwuahauauha

ProgramFOX

6:32 AM

@Rusher: When you was in this chat room, I already left. So I will ask my question now: is it allowed to create several wolves that do a different task, but work together?

plannapus

7:00 AM

Am I the only one being slightly frustrated at the fact that people prefer posting their wolves along with lengthy results of test runs rather than explanations of what their wolves do and why they differ from the previous ones? I know most of them are new to this site and need to be shown how we do things here but i feel like a nitpicking jerk doing so :)

ProgramFOX

@plannapus If you have a wolf with a lot of code, it's not very easy to post a detailed explanation.

plannapus

It is not harder than posting results of test runs (which are kind of worthless anyhow considering: 1) they don t include all the wolves and 2) Rusher is going to update the leaderboard anyway(

But even without a detailed explanation, at least a sentence or two explaining the concept would be welcome

ProgramFOX

@plannapus That's true.

Saying what your wolf actually does is a minimum, IMO.

A very detailed explanation is good, but not very necessary.

Quincunx

7:24 AM

Wolf Koth tests: 100 lions added per 100 animals added:

Lion: 126
HerjanWolf: 76
OmegaWolf: 70
ProAlpha: 48
AlphaWolf: 40
ShadowWolf: 36
LazyWolf: 32
HybridWolf: 28
StoneEatingWolf: 17
GatheringWolf: 15
Wion: 10
BlindWolf: 7
CamperWolf: 7
Sheep: 7
HonorWolf: 5
WhenIGrowUp: 4
ForrestWolf: 2
MimicWolf: 1
MOSHPITFRENZYWolf: 1
LionHunterWolf: 1
WolfRunningWithScissors: 0
Stone: 0
WolfWithoutFear: 0
GamblerWolf: 0
Bear: 0
EmoWolf: 0
MigratingWolf: 0

plannapus

@Quincunx ahah ok this time there is a clear difference

Though the leaderboard doesn t change that much

Quincunx

Yup.

plannapus

Wion performs better though

Quincunx

Now I'm increasing the size of the map accoding to the correct amounts.

Lion: 216
HerjanWolf: 89
OmegaWolf: 84
ProAlpha: 69
ShadowWolf: 53
AlphaWolf: 53
LazyWolf: 40
HybridWolf: 36
StoneEatingWolf: 31
GatheringWolf: 19
Sheep: 19
Wion: 14
CamperWolf: 8
BlindWolf: 7
WhenIGrowUp: 6
HonorWolf: 5
ForrestWolf: 5
MimicWolf: 2
MOSHPITFRENZYWolf: 1
GamblerWolf: 1
Stone: 0
Bear: 0
EmoWolf: 0
WolfRunningWithScissors: 0
WolfWithoutFear: 0
MigratingWolf: 0
LionHunterWolf: 0

Interesting

Kind of funny that my wolf is doing as well as it is.

@plannapus maybe I shouldve said that after the 100 times test, I decided that the difference is so small that I'm only going to run it 10 times.

plannapus

ok, that probably doesn t change that must though

your wolf is probably getting protected by the many wolves on the field

Quincunx

7:29 AM

@ProgramFOX Of course it's allowed!

ProgramFOX

Nice!

Quincunx

I was thinking of creating a MinionWolf and a HiveQueenWolf, but I decided it would be too hard.

MinionWolf would fight any non-HiveQueenWolf

The problem is that it is impossible to tell what kind of wolf you're facing.

plannapus

how would they recognize each other?

ah

Quincunx

I tried using reflection, but it doesn't work. :-(

ProgramFOX

@Quincunx Yeah, that's why it would be very hard to create two wolves that work together.

Quincunx

7:32 AM

Wolf Test with 200 Lions added per animals added and size adjusted accordingly:

Lion: 1871
HerjanWolf: 89
OmegaWolf: 62
Wion: 37
ProAlpha: 34
ShadowWolf: 32
WhenIGrowUp: 25
LazyWolf: 24
AlphaWolf: 23
GatheringWolf: 11
ForrestWolf: 4
StoneEatingWolf: 1
HybridWolf: 1
Sheep: 1
Bear: 0
MimicWolf: 0
GamblerWolf: 0
CamperWolf: 0
EmoWolf: 0
BlindWolf: 0
HonorWolf: 0
WolfRunningWithScissors: 0
Stone: 0
MigratingWolf: 0
MOSHPITFRENZYWolf: 0
LionHunterWolf: 0
WolfWithoutFear: 0

plannapus

i ve been trying to develop a wolf pack that tries to find each other (after mapping the field and using fixed object, i e stones and wolves that don t move, to compare maps) and then hunt together but it is far beyond my capacities at the moment :)

Quincunx

HerjanWolf has the best anti-lion strategy.

ProgramFOX

My wolf really has the worst. It is really too addicted and only wants to attack...

Quincunx

HerjanWolf Thought I'd post a link so we can more easily upvote it.

plannapus

i would really like him to explain what his wolf does before upvoting to be honest.

(not that i think i won t upvote whatever his explanation will be, really just to enforce good behavior)

Quincunx

7:40 AM

I'm trying to run the test for 1000 lions added per 100 animals added (with extra space). It's going slowly. Been ~5-10 minutes so far.

plannapus

so that s what? 24 000 lions? :)

Quincunx

:-) Some huge number.

I give up

Re-running with a lower number.

Manu

The problem with AlphaWolf and swapping with the lions is that the lion may have moved already... then it moves right into the lion -.-

i need to store the surroundings for one round...

Quincunx

Now adding 10000 lions (not per 100 animals, alone). What will happen?

Manu

pretty sure the lions will win ;)

Quincunx

7:46 AM

@Manu Nope. I think not. I think my impatience will win.

Manu

:D
Can you try it with 1000 wolves per type? to see which wolf has the best wolf-avoiding strategy?

Quincunx

Test for 10000 extra lions added, size-adjusted:

Lion: 5854
HerjanWolf: 89
Wion: 56
WhenIGrowUp: 50
ShadowWolf: 48
OmegaWolf: 26
GatheringWolf: 16
ForrestWolf: 7
LazyWolf: 7
ProAlpha: 6
AlphaWolf: 3
HonorWolf: 0
Sheep: 0
CamperWolf: 0
LionHunterWolf: 0
WolfWithoutFear: 0
Bear: 0
WolfRunningWithScissors: 0
Stone: 0
BlindWolf: 0
StoneEatingWolf: 0
GamblerWolf: 0
MOSHPITFRENZYWolf: 0
EmoWolf: 0
HybridWolf: 0
MimicWolf: 0
MigratingWolf: 0

@Manu sure

@Manu would you like the board to be adjusted according to the increased number of wolves?

Wait, no

That's too many wolves.

That would take too long.

I'll run at 500, size adjusted

Manu

@Quincunx Cool! :)

plannapus

The normal run is already a good test of wolf-avoiding strategy since most of the animals on the field are wolves (100 lions, 100 bears, 100 stones and ~30 x 100 wolves).

Quincunx

@Manu Um, was I supposed to not increase the other things? because I did.

I'll add normal wolves too?

Manu

7:51 AM

@Quincunx I don't care, just some more wolves :D

Quincunx

I give up on the 500 run. Too long.

2* number of wolves, size adjusted, no normal animals.

OmegaWolf: 164
HybridWolf: 162
HerjanWolf: 161
ProAlpha: 154
HonorWolf: 154
StoneEatingWolf: 138
AlphaWolf: 137
ShadowWolf: 136
CamperWolf: 100
LazyWolf: 98
BlindWolf: 88
GatheringWolf: 72
Sheep: 62
EmoWolf: 53
LionHunterWolf: 49
GamblerWolf: 34
WolfWithoutFear: 25
MigratingWolf: 22
ForrestWolf: 13
MOSHPITFRENZYWolf: 12
MimicWolf: 1
WolfRunningWithScissors: 0
Wion: 0
WhenIGrowUp: 0

Manu

@Quincunx Impressive... StoneEatingWolf so far up? haha

Quincunx

Now for no increase on player wolves, but same number as default wolves as other wolves (as in default wolves are half of all animals)

Also, no normal animals:

OmegaWolf: 164
HybridWolf: 162
HerjanWolf: 161
HonorWolf: 154
ProAlpha: 154
StoneEatingWolf: 138
AlphaWolf: 137
ShadowWolf: 136
CamperWolf: 100
LazyWolf: 98
BlindWolf: 88
GatheringWolf: 72
Sheep: 62
EmoWolf: 53
LionHunterWolf: 49
GamblerWolf: 34
WolfWithoutFear: 25
MigratingWolf: 22
ForrestWolf: 13
MOSHPITFRENZYWolf: 12
MimicWolf: 1
WolfRunningWithScissors: 0
WhenIGrowUp: 0
Wion: 0

It's amazing how similar those two runs were.

Well, g'night.

plannapus

Actually there are so identical that i suspect a copy/paste mistake on your part :)

ProgramFOX

@plannapus It can't be a copy/paste mistake, because in the first result, WhenIGrowUp is last and in the second, Wion is last.

plannapus

8:00 AM

ah true i haven t seen that. Yet the numbers are exactly the same and the "default wolves" don t appear in the last one.

Quincunx

Yeah the default wolves don't appear because I did not add them to the class list. I was too lazy.

plannapus

:) ok this explains that then

Quincunx

At your post, I also thought I made an error. Perhaps I'll rerun it.

I should also say that don't trust these tests down to the +-1, because I use double and rounding.

This looks more like what it output the 1st time I did the test.

Last test re-done due to copy-error:

Wolf: 238
HonorWolf: 79
BlindWolf: 75
CamperWolf: 75
LazyWolf: 72
OmegaWolf: 62
ShadowWolf: 58
EmoWolf: 55
StoneEatingWolf: 50
HybridWolf: 49
ProAlpha: 39
GatheringWolf: 37
Sheep: 37
LionHunterWolf: 35
HerjanWolf: 33
GamblerWolf: 23
AlphaWolf: 21
ForrestWolf: 8
WhenIGrowUp: 0
WolfRunningWithScissors: 0
MigratingWolf: 0
WolfWithoutFear: 0
Wion: 0
MOSHPITFRENZYWolf: 0
MimicWolf: 0

g'night

Rusher

1:05 PM

@ProgramFOX If your wolves are obviously collaborating, then I guess I still have to include it, but I'll encourage everyone to downvote the hell out of it.

Haha wow. You guys did some serious testing. Last night I got every single updated, legit Wolf (including a few non-java submissions) but I passed out at midnight. Will be able to update the real scoreboard tonight. :)

plannapus

@ProgramFOX instead of writing several packs of wolves collaborating with one another, why not creating one single pack of wolves that is split into different subpacks collaborating? Sure that means less wolves but that s conceptually the same.

Rusher

I want to retract what I said about downvoting collaborating wolves. I actually don't think you can do it with a vision range of 1. You'll just step on each other.

You can't read from or write to files written by other wolves either, so the only way to communicate is with a variable that is public. If you make a public variable, I will destroy it with my own submission lol

But I don't know much about reflection so... if you can trick me then kudos to you.

What happened to @Chris-JesterYoung? He must have a real life.

Manu

1:34 PM

@Rusher The variable would have to be public static, so every wolf could change it easily...

Rusher

For anyone that is worried about GamblerWolf, it doesn't actually work. I instantiated his Wolf and then printed myRand.nextDouble() and got a random value.

The same goes for Mirror. His Wolf tries to modify the Attack enums. I printed them out after instantiating his class, and got Rock, Paper, etc.

Manu

@Rusher I also tried to modify static fields of other wolves, but it didn't work :(

ProgramFOX

Yeah, I get your points @Rusher. But there is still one option to communicate: using Sockets, but it will probably take some weeks before that will be ready...

Other wolves could also connect, but I could create a Socket that only allows 200 wolves.

Doorknob

1:57 PM

@Rusher A real life? What's that? :-P

Rusher

@Doorknob You wouldn't know, because you are a Doorknob. The only real social experiences you get are with people entering Wonderland.

I've pretty much forfeited my life this week. 8:00-5:00 work. 5:00-9:00 play Magic: The Gathering. 9:00-11:00 cry because I can't keep up with all the wolf submissions. Sleep and repeat.

Do any of you live in Texas?

how to escape < and > on stack in a post?

ProgramFOX

@Rusher Write < or >

< for < and > for >

Rusher

Thanks!

ProgramFOX

That's called a HTML Entity:

List of XML and HTML character entity references

In SGML, HTML and XML documents, the logical constructs known as character data and attribute values consist of sequences of characters, in which each character can manifest directly (representing itself), or can be represented by a series of characters called a character reference, of which there are two types: a numeric character reference and a character entity reference. This article lists the character entity references that are valid in HTML and XML documents. Although in popular usage character references are often called "entity references" or even "entities", this usage is wrong. ...

Rusher

Haha I love the "this usage is wrong.[citation needed]"

Haha and the HTML folks even made Wikipedia issue this warning: This article contains special characters. Without proper rendering support, you may see question marks, boxes, or other symbols.

Rusher

2:32 PM

If the owner of the challenge has edit without review privilege, could someone submit a code-golf answer that takes over the mouse momentarily, edits all other submissions to be something larger than their submission, and then exits?

ProgrammerDan

Nefarious!

Rusher

It's pretty nefarious for me to be honest. I never see loopholes like that

Averroes

4:25 PM

@Manu Hey! Stone Eating Wolves are a evolving specie!

:P

Rusher

Holy crap. This is freaking impressive codegolf.stackexchange.com/a/25676/18487

ProgrammerDan

whoa

good thing I didn't start a Wolf yet, no point! :P

ProgramFOX

@Rusher Wow! That wolf is great!

Manu

5:15 PM

@Rusher impressive!

Herjan

7:37 PM

Yeah I thought the same when I saw the averages, but when I added that wolf to my list it really wasn't all that great, when I tested single rounds I saw a lot of instability, sometimes it had 6x wolves left, my (10 rounds 1000 iterations) test:
class animals.Bear - 0.0
class animals.Lion - 0.0
class animals.Stone - 2.2
class animals.AlphaWolf - 75.8
class animals.HerjanWolf - 85.0
class animals.GatheringWolf - 37.5
class animals.OmegaWolf - 86.0
class animals.ShadowWolf - 70.4
class animals.MOSHPITFRENZYWolf - 10.1

zenflux

Yeah, a little while ago I ran a 100x run avg, DeepWolf had mid 70's I believe. Doing a 500x now, but it takes so long when Deep is taking 75% of the total sim time (!)

Oh wait, it's because Hybrid is making a pet instance of everything for each of it's instances

Herjan

That's a lot, can't wait to see such reliable results, and Hybrid doesn't make a instance of every class, only from HonorWolf, ProAlpha,OmegaWolf and AlphaWolf I believe

zenflux

well it updated to include a few more, including deep.

and yes, the 'pets' are on each instance

I made them a static list for testing

Herjan

Ok lets update my HybridWolf as well then

He has added my wolf as well, I feel honored ;)

ProgrammerDan

So many fascinating subversion techniques. I looked through DeepWolf's code but couldn't determine at a glance how mapping was being handled, but fun stuff all the same.

zenflux

7:49 PM

@Herjan Yours seems to be the best in my tests, you should also feel proud ;)

@ProgrammerDan And I have one that nobody is using currently as well, I plan to post it today, stay tuned

ProgrammerDan

@zenflux I can't wait! This is a great contest, where the complexity of the posted answers increases nearly exponentially with the amount of time since the contest was posted.

Herjan

@zenflux I feel great! :)

zenflux

@Herjan You seem to have struck a rare intersection between a local minimum in code complexity and a local maximum in effectiveness. Respect.

Quick profile of the sim: i.imgur.com/kMUVa8A.png

can't seem to get that to inline the image

But as you can see, DeepWolf is insane. Game.moveAll is usually top of the list

Herjan

Is that with all wolves then?

zenflux

Yes, or nearly so.

Geobits

8:02 PM

Once the wolves die down...

0

A: Proposed Question Sandbox - Mark XII

The Hunger Games king-of-the-hill Let's face it, The Hunger Games isn't strictly about hunger. This is. If you don't eat, you die. Overview There is an island populated with a herd of prey animals. You control a pack of five predators. Your object is to not starve your pack. Do this by eating ...

zenflux

@Herjan Says I'm running 28 classes.

ProgrammerDan

@Geobits SHWEW. Finished reading through the specs. I'll read again later tonight and see if I can poke any holes, but it looks respectable at a glance.

Geobits

Thanks! Hopefully there's nothing terribly wrong with it. Minor things I don't mind, lol

Rusher

@Geobits I'd suggest using a title that isn't copyrighted

Geobits

Honestly, if they want to sue me over it, let them. I'll change it at that point.

ProgrammerDan

8:16 PM

The Hunger Island?

zenflux

The Hunger Island Games?

aha, see, it's totally fresh!

Geobits

If I come up with something better, I may change it. If not, I don't really see much of a problem. I'm pretty sure that trademarks can be reused anyway, so long as the works are in entirely different fields. Just don't hope for a novelization of the results ^^

Rusher

@Geobits I'd be less worried about being sued and more worried about someone filing an infringement notice, at which point your account will be deleted without them even bothering to read most of your post.

Geobits

If they file an infringement notice over this and SE doesn't laugh it away, I will have truly lost faith in the system.

Seriously.

Rusher

Who do you think is responding to infringement notices on Youtube? Nobody. It's accepted unless challenged.

If challenged, THEN the video is actually looked at.

Geobits

8:21 PM

Right, but they receive heaps of thousands of notices every day. From what I've seen on meta.so, they actually read them here, because there aren't so many.

Rusher

If you get multiple notices, then it is taken down regardless of whether it was actually copyrighted or not.

Ah, well you have more information than I do then. I hadn't seen that on meta despite my search two minutes ago.

zenflux

See, now you need to post a golf question asking for scripts to generate names involving "Animal" "Hunger" "Death" and "Games"

Geobits

The only action I've seen was a moderator manually removing the post. Not the account, etc. Any rebuttal can be brought up at that point.

Rusher

I like the Rock Band name generation. My favorite was "Unlimited Death Tomato"

zenflux

Hey, @Rusher, you updating the wolves leaderboard today/tonight?

Herjan

8:30 PM

@Rusher I heard zenflux made a 500 rounds 1000 iterations test with 28 classes, maybe you may use his results ^^

zenflux

I just did a quick 10 round:

HybridWolf 85.7
MyWolf 84.0 *
HonorWolf 83.7
DeepWolf 82.7
HerjanWolf 82.2
OmegaWolf 80.9
ProAlpha 80.1
ShadowWolf 73.7
AlphaWolf 73.6
StoneEatingWolf 60.7
LazyWolf 50.0
BlindWolf 44.1
Sheep 40.6
GatheringWolf 27.2
CamperWolf 26.7
LionHunterWolf 26.0
EmoWolf 21.3

MyWolf, is, obviously, mine. It's in development.

I have a 100x and the 500x still cooking, likely going to be a while.

ProgrammerDan

When you reset your simulation, @zenflux, how are you handling the fact that the wrapper code doesn't "reset"

Rusher

@programmerDan he didn't include the non-java entries

zenflux

@ProgrammerDan By not using the wrapper :P

ProgrammerDan

Ask a silly question ... :P

zenflux

8:36 PM

I really should invest some time into optimizing the Game code if I'm going to be doing this much testing...

Herjan

@zenflux Its just now that I see you're BurntPizza from java-gaming lol

zenflux

Yep, my account names are messed up right now

Rusher

@zenflux I was up until midnight last night adding submissions to the project, checking for changes to the ones I already had, and making sure everyone's submissions "worked" by running them in small batches. I also streamlined the output so that it writes to a .csv file that I can just copy and paste into Google spreadsheet. If all goes well, results will be posted tonight.

Herjan

And how did your wolf do in the 500 round? Im way more curious about that one

ProgrammerDan

Nice, @Rusher. That's dedication.

zenflux

8:38 PM

@Herjan It's still running

@Rusher Man, I've lost quite a bit of sleep this week due to your dang contest

Rusher

@ProgrammerDan It would be easier if I didn't have a MTG tournament coming up this weekend.

ProgrammerDan

@Rusher wow, yeah. Candle, burning at both ends much?

zenflux

furious coding Oh, it's 3AM, when did this happen? Oh, I have to wake up at 6 you say?

That's my life this week

Rusher

zen, did you come from StackOverflow to here?

And @Herjan, where did you come from?

zenflux

@Rusher I more or less just started on both

I came from java-gaming.org

as did Herjan

Herjan

8:40 PM

I just saw this contest on java-gaming.org, zenflux made a thread about this contest and it looked cool so I bumped in

Rusher

Hah. That explains a lot...

zenflux

Actually, now that I think of it, I saw this on reddit, I'll try to find the link

reddit.com/r/programming/comments/22h2py/…

ProgrammerDan

Nice, I wonder if the poster of the idea behind DeepWolf on Reddit is the same as the creator of DeepWolf here?

zenflux

8:57 PM

This guy? >How about a wolf that loads all other known competing wolves and tests how they would behave and uses that to create a move tree and reacts?

That's basically Hybrid

Herjan

[–]alias_vim_eq_reddit 3 points 2 days ago

How about a wolf that loads all other known competing wolves and tests how they would behave and uses that to create a move tree and reacts?

I just wanted to post that as a comment for Hybrid lol

but its really genious

zenflux

I just think it's interesting how well it works, it's definitely better than it's parts

ProgrammerDan

Actually, yeah, my bad, that's what I meant.

zenflux

But damn if it's not the slowest thing

@Geobits, good on ya for specifying a 200ms timeout on Hunger Games, these unbounded wolf AI are getting ridiculous

surprised there isn't a RabiesAI yet

or a WolfOfWallStreet

Geobits

@zenflux I needed the timeout, really. If a submission just hung, it would just stall me otherwise. I can't sit there waiting forever.

The 200ms is just an arbitrary number I came up with. Computing one turn should take that long. I mean, you have a max of five moves to figure out.

Rusher

9:13 PM

@Geobits When I update the scoreboard tonight, I'm also going to add an expiration date of 1 week from today. That way, you can align the introduction of your challenge IF you choose. You obviously don't have to. Either way, I can't keep losing sleep forever.

Geobits

11:29 PM

@Rusher You don't need to hard-expire it or anything. If you want to keep it open, just add that you will only be running scoring rounds every X days/weeks, something along those lines. If you're just tired of it, though, that's a different story.

Transcript for