Formic Functions - 2018-04-22

« first day (273 days earlier) ← previous day next day → last day (464 days later) »

12:18 AM

Time for data analytics. Now, what data should I be gathering?

Was that instance of The Formation getting 10.something just a fluke?

Or 8.something? Either way, it seems I have some catching up to do with my previous versions

I have a slight suspicion that reducing the spawnrate further might be prudent, given that the formation can pick up food closer to the queen and then extend with the new spawnchanges, but...

7 hours later…

7:09 AM

@Draco18s Yeah I really should get around to excluding them automatically based on the comments on the answers. There are too many to keep checking by hand...

Steamroller Ants:

The latest edit of this player was disqualified: "A laden worker cannot move onto food." Input: [{"color":1,"food":0,"ant":{"food":0,"type":2,"friend":true}‌},{"color":1,"food":‌0,"ant":null},{"colo‌r":1,"food":0,"ant":‌null},{"color":1,"fo‌od":0,"ant":{"food":‌28,"type":5,"friend"‌:true}},{"color":1,"‌food":0,"ant":{"food‌":1,"type":1,"friend‌":true}},{"color":1,‌"food":0,"ant":null}‌,{"color":1,"food":1‌,"ant":null},{"color‌":1,"food":0,"ant":{‌"food":0,"type":1,"f‌riend":false}},{"col‌or":1,"food":0,"ant"‌:{"food":25,"type":5‌,"friend":false}}] R‌esponse: {"cell":6} — trichoplax Aug 3 '17 at 11:33

HalfThere:

Input: [{"color":3,"food":0,"ant":{"food":1,"type":4,"friend":false‌}},{"color":3,"food"‌:0,"ant":{"food":1,"‌type":4,"friend":fal‌se}},{"color":3,"foo‌d":0,"ant":{"food":1‌,"type":4,"friend":f‌alse}},{"color":7,"f‌ood":0,"ant":{"food"‌:1,"type":3,"friend"‌:false}},{"color":3,‌"food":0,"ant":{"foo‌d":24,"type":5,"frie‌nd":true}},{"color":‌3,"food":0,"ant":{"f‌ood":1,"type":4,"fri‌end":false}},{"color‌":3,"food":0,"ant":{‌"food":1,"type":4,"f‌riend":false}},{"col‌or":3,"food":0,"ant"‌:{"food":1,"type":4,‌"friend":false}},{"c‌olor":3,"food":0,"an‌t":{"food":1,"type":‌2,"friend":false}}] Response: undefined — trichoplax Mar 3 at 21:52

Steamroller has workers that move onto food when laden, and HalfThere can't cope with being surrounded on 8 sides.

2 hours later…

9:29 AM

The bounty is posted, and will be awarded to whoever is top of the leaderboard 7 days from ... now!

@trichoplax O crap, Martin remembered. Time to get back at full speed!

Well, I may have felt the time had come to give a reminder...

(As tempting as it was to leave the bounty-fueled update frenzy going until you all collapse from exhaustion)

@trichoplax Meh... I, for one, wasn't really in a frenzy. I just had the bounty at the back of my mind, not really doing much...

Oops...

Now, do I want to handicap myself by posting the Wildfirified Highway as a separate entry, or do I want to replace Highway with it? Hmm...

9:42 AM

Do you know how the score varies each way?

@trichoplax Well, let's start off by saying that I don't have the Wildfiriway even near being finished, so absolutely no data right now. I feel like the patterns will be similar enough to confuse each other, so I might just go ahead and replace it. If it won't affect the score much in a 100-gamer, I will want to post it as a separate entry.

Let me know if you'd like me to run a tournament each way once it's ready, to compare results

@Alion Score of Wildfiriway, for clarification. I don't care if Highway falls to the lower half of the leaderboard :P

First off, I need to finally finish remaking Lightspeed. Will just copy-paste and update logic to equivalent in framework v6, so no prettification, just manual cross-compilation. I don't have time for that! :P

Also, I might actually not have much of a chance considering how sophisticated SlM has become over the past couple of weeks. It's obliterating the test version of Wildfiriway, that's for sure.

Hm. I have no clue what past me was thinking. It's pretty much impossible to do all of the things I need to do to make Wildfiriway work on the hill in the span of a couple of days (5 at best).

Have fun :P

Now I've got to pray that I actually remember what v5 was doing so that I can painlessly cross-compile it...

10:26 AM

Good grief, this is one hell of a function: function logic(partnerTest, partnerOrthC, partnerDiagC, frontC, sideC, backC, backCells, moveCells, adjacentC)

protip: pass an options object

then it'll be function logic({partnerTest, partnerOrthC, partnerDiagC, frontC, sideC, backC, backCells, moveCells, adjacentC}) :P

@ASCII-only Does that actually help me if I have to declare all of the variables every time this function is called, though?

not really

@Alion You don't have to remember the order? :p

Oh, yeah. That. Though it'd be more verbose. Meh, I don't care enough.

10:32 AM

Allows default values without restricting which order you can miss them out

10:46 AM

A'ight. Porting successful. Replacing Lightspeed in a sec. Functionally identical so no new tournament needs to be run.

Famous last words

@trichoplax What do you mean ;-;

> Functionally identical

no need to test then...

I'll start a new tournament for each edit regardless, so it will get tested here :)

@trichoplax Oh, my bad, the solo-queen phase was revamped to match the Single Queen entry (because it seems to be the most efficient despite being up there in terms of simplicity).

Sec, I'm retarded. Forgot that I actually have to new a NOP in v6. Fixing...

@trichoplax Should be good now. Don't test version 2.5, it's broken. Test 2.5.0.1 if you want to.

I'll just run the very latest then

10:59 AM

(although it would probably never come up in real testing - the error would only happen if either of the Lightspeed partner had no legal moves left)

Hm.. Forgot to switch DEBUG_MODE to false. Shouldn't affect anything. shrugs

Will it affect the speed?

@trichoplax Unless the JS engine can optimize away some function declarations because of it - no, because the functions affected by this switch aren't placed anywhere in the code. Even if it optimizes it, though, it's gonna save at best ~0.003 ms (and since Lightspeed caches fantastically it won't even matter in the long run).

Great - no console.logs to worry about then?

@trichoplax Only console.log would happen if the error() function was called already, and at that point it makes no sense for the entry to continue as something has gone terribly wrong and I want it get DQ'd anyway.

Great. I'll set another tournament running

11:09 AM

Both error calls in my code are only related to things that make no sense if I coded this thing properly. I would get really confused if either of them got triggered...

I wonder how much more attention the bounty will attract.

Not sure how much difference it will make, since the challenge has spent a lot of time on the front page already recently with all the posts and edits

3 hours later…

2:04 PM

@Alion Something happened - Lightspeed was disqualified after 1 game

@trichoplax Fascinating, frankly. Do you have the view that DQ'd it?

I've commented on the post with the details

Hmm. Perhaps I should've thrown it at least one random game for testing... Found a DQ myself.

3 hours ago, by trichoplax

Famous last words

:P

Give me a break...

2:15 PM

I never expected this game to be taken so seriously people would be running tests on submissions before posting. I feel like it's outgrown me and I'm far out of my depth

So the partner managed to get lost. How the hell?!?

The worst part about this? None of the views I can get will actually give me anything. They're just post mortem. If the partner gets lost, then something's gone wrong already...

@Alion modified the maps size to 25x10 to encourage enemy collisions & saw the queen spawn a helper with one already in view

@dzaima .....

Not only does that not make sense, it also contradicts when I've been researching (unless it's a different issue!)

;_;

2:32 PM

okay why cant I see food

.find() instead of .some()

if (partner.find(DIRECTIONS)) would fail if it found the partner on cell 0.

oh fonts again

@Alion yep, the partner was in a corner

Yup, you sniped the bug. Thanks a bunch.

@trichoplax Should be all good now.

Great :)

heh highway does not like having too frequent food

2:40 PM

Surprisingly, it had nothing to do with other ants. It was able to simply implode by itself.

I have no clue how that didn't occur in 100 games.

@Alion you sure you had pasted the right thing? :p

@dzaima Uhh... maybe? Lemme check.

Oops, forgot a log() in there.

(it didn't happen to me in the ~8 games I let it alone run for so that's why I thought it was other ants)

I'll just disable DEBUG_MODE while I'm at it.

Okay, that should finally be everything. Good grief.

Oh, and by the way, in the game where the DQ happened to me there were other ants, but it just imploded on food alone, so that's why my conclusion was that it can implode by itself. It did have colors nearby, though it shouldn't take them into account in the LS phase...

(hopefully it doesn't ;_;)

I don't know anything anymore. Help.

Running a new tournament now. Will let you know if anything dies

3:13 PM

Hmm... In LS's routine, the partner walks clockwise to the queen and the queen walks counterclockwise. I wonder if there are some benefits (or disadvantages) to reversing this order...

4:04 PM

@trichoplax ah, that one's fairly recent. Steamroller I may have known about before, given the age. If you do update your controller to check that, let me know, and I'll pull it into my fork. Reminds me that I have a few other tweaks I was working on for it too... (namely running a tournament on a fixed seed, and making that seed update after every game, so that if something comes up in a "random" game, that single game can be repeated)

@Alion I'll see your logic func and raise you this

4:50 PM

Automatic formation ejection, coming up soon TM!

5:12 PM

Note to self: I should probably make the error message on throw customizable instead of "A critical error has occurred!". This doesn't give me much info when it happens during a tournament...

5:28 PM

A critical error has occured! ...and it turns out you assigned a function rather than calling it. Goddamn difficult to track down in code.

6:16 PM

I've no idea what I'm doing at this point. Because of the hiatus I've forgotten how to Formic :P :c

« first day (273 days earlier) ← previous day next day → last day (464 days later) »

Transcript for

Apr '1822

Formic Functions

A place to discuss the Ant QotH: codegolf.stackexchange.com/qu...

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile