The Nineteenth Byte

TimmyD

17:03

TIL: About forty percent of all mammal species are rodents.

Nathan Merrill

TIL: Timmy learned that about forty percent of all mammal species are rodents.

Dennis

TIL: Nathan learned that Timmy learned that about forty percent of all mammal species are rodents.

trichoplax

TIL: 40% of recent TNB posts are about rodents

Geobits

TIART: People in here have too much time on their hands.

3

trichoplax

I have enough time spare to ask what that stands for.

Dennis

17:08

Take It and Run Thursday.

Geobits

today i am reminded that

trichoplax

Ah right

Dennis

I like mine better.

Geobits

Me too, tbh.

trichoplax

6 days early...

its time will come

Nathan Merrill

17:09

I thought it was "Today I already knew that", but know that I type that out, I realize it doesn't match the initials

TimmyD

Today I Already RKnew That ... the "R" is silent.

Dennis

With a silent R.

Geobits

I definitely did not go through seven acronyms before deciding on that one >_>

TimmyD

Well, that's good. Wouldn't want you to waste time.

Geobits

Heaven forbid.

Morgan Thrapp

17:13

No one here ever wastes time.

Cyoce

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Pattern finding! No OEIS. 0, 1, 1, 3, 5, 9, 13, 27, 36, 73, 115

Cyoce

trichoplax

On my screen "Great minds waste time alike" is all on the second line so it doesn't split

Geobits

Wait... nvm

On mine it's all on the first except the dash and name.

trichoplax

17:22

Show off :P

TimmyD

The 13 -> 27 jump keeps messing up my pattern.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

:3

Geobits

The 36 -> 73 is no better.

But they're both *2+1

trichoplax

Yeah it's awful - @CᴏɴᴏʀO'Bʀɪᴇɴ you need to change some of these numbers for us

TimmyD

It's not something egregiously esoteric, like "Number of cuboid solids that can be combined in n-dimensional space by drawing lines connecting their vertices" or something

Right?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

17:24

@TimmyD It's worse.

TimmyD

(ₒ_ₒ)ʷʰᵃᵗ

Cᴏɴᴏʀ O'Bʀɪᴇɴ

wait I may have copied it wrong

Geobits

>_>

Cᴏɴᴏʀ O'Bʀɪᴇɴ

oh yup sure enough

Nathan Merrill

-.-

trichoplax

17:25

You mean I've been wasting my time working on the wrong time wasting??

4

Nathan Merrill

vote as unclear

Cyoce

Cᴏɴᴏʀ O'Bʀɪᴇɴ

uhh sorry guys I concatenated two sequences from my sheet >_>

TimmyD

@Cyoce ==?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

0, 1, 1, 3, 5, 9, 12, 20, 29, 43

^ that is the correct one

Cyoce

17:27

@TimmyD that's what it thinks I'm trying to do, but it isn't

stupid jslint

trichoplax

Does that return n or 2?

Cyoce

2

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Why are you assigning something in a return statement?

trichoplax

Hyperoptimisation

Cyoce

to avoid multiple lines

TimmyD

17:28

Encapsulate in parens?

Cyoce

nope

I just changed it

TimmyD

@CᴏɴᴏʀO'Bʀɪᴇɴ This isn't better ... ._.

Geobits

I looked it up. It is indeed not better.

Nathan Merrill

a sequence of non-decreasing numbers

I won!

Geobits

Not strictly increasing, though.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

17:30

@TimmyD It may be a tad esoteric, I'll give you a hint: σ

Nathan Merrill

ninjedit-d

Cᴏɴᴏʀ O'Bʀɪᴇɴ

σ_σ

TimmyD

@CᴏɴᴏʀO'Bʀɪᴇɴ With this group, I can't tell if that's supposed to be an emoticon or a sigma for deviation. :D

Geobits

@CᴏɴᴏʀO'Bʀɪᴇɴ I wouldn't call it "a tad esoteric". IMO it's more along the lines of "nobody is going to stumble upon this, ever".

Martin Büttner

does anyone know if we've done oeis.org/A052409 or something closely related?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

17:32

@Geobits Oh. :I

Martin Büttner

I think it would make a nice and fairly simple challenge that probably isn't killed directly by built-ins

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@TimmyD Both σ_σ

TimmyD

Bad Martin. Conor said no OEIS

Martin Büttner

this is not related to Conor's sequence

Cᴏɴᴏʀ O'Bʀɪᴇɴ

σ_σ

Nathan Merrill

17:32

@MartinBüttner I'm not smart enough to understand that sequence, so I'm looking forward to your post

Cᴏɴᴏʀ O'Bʀɪᴇɴ

"Sad disapproval"

Geobits

@MartinBüttner Most languages probably don't have builtins for it, but I'm guessing some golfing ones do. It's basically gcd(prime_factors(n)), right?

Martin Büttner

@NathanMerrill are we talking about the same sequence?

TimmyD

a(n)>1 iff n is a perfect power ... Interesting.

Geobits

Oh wait it's the GCD of exponents. Hmm.

Martin Büttner

17:34

@Geobits uh, prime factor counts, but yes

Nathan Merrill

@MartinBüttner I'm no good at reading math

Geobits

Yea, that's better. I'm not sure how gcd of primes would work well anyway ^^;

Cᴏɴᴏʀ O'Bʀɪᴇɴ

o[σ_σ]o

Martin Büttner

@NathanMerrill it's really just asking what is the largest power in n... perfect squares get 2 (provided they aren't also a higher power), perfect cubes get 3 (provided...) and so on.

trichoplax

Does offset 1,4 mean it is contained as a subset of itself by taking every fourth item?

Nathan Merrill

17:36

@MartinBüttner ah, ok

TimmyD

@Geobits return 1

Nathan Merrill

wait, but 3*2^2 would get 2?

trichoplax

Apparently not. Ignore me...

Martin Büttner

@trichoplax no

it means the first term is a(1) and the fourth term is the first non-zero one

@NathanMerrill no it's not a perfect square

Geobits

@TimmyD Golf that sucker down! I'm sure there's got to be some language that spits a 1 for the null program :P

Martin Büttner

17:38

@NathanMerrill you only have a perfect square if you can write it as m^2 for integer m

Nathan Merrill

ok

I was mixing up your statements and Geobits :P

if I take a number, and represent it as a list of prime powers, isn't there a name for that representation?

Martin Büttner

prime factorisation?

Nathan Merrill

so, 24 = 2^3 * 3^1 -> [3,1]

Martin Büttner

oh right

Nathan Merrill

no, there was some guy whose name is attached to that list

trichoplax

17:40

So you'd have zeros in there for primes not present?

Nathan Merrill

yeah

Cᴏɴᴏʀ O'Bʀɪᴇɴ

2D OEIS anyone?

:3

Martin Büttner

huh?

OEIS has loads of tables and triangles in it

if that's what you're suggesting

Cᴏɴᴏʀ O'Bʀɪᴇɴ

no, like f(x,y) => (x,y)'

instead of a(x) => x'

Geobits

Aww, I thought you meant a grid where each row/column was an entry. Like a crossword puzzle.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

17:46

^ that would be cool, too

Martin Büttner

@CᴏɴᴏʀO'Bʀɪᴇɴ oh okay. how would you visualise that though if the values are also pairs?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

hi @primo!

primo

golf.shinh.org/p.rb?99+shinichiroes+of+hamaji#ranking

:)

Just discovered auto-complete in Vim

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@MartinBüttner Maybe with a graph?

uhh

Nathan Merrill

bah, I hate it when I need to refactor all my code

Cᴏɴᴏʀ O'Bʀɪᴇɴ

17:48

ugh about to do a bio test >_>

Cyoce

@CᴏɴᴏʀO'Bʀɪᴇɴ Just had one of those they're awful

Geobits

Good thing I've been dropping bio facts in here today. Good luck!

primo

hmm, i thought the majority of mammal species were bats

trichoplax

More bats than everything else put together?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@Cyoce D:

primo

17:51

eh, only 20%

trichoplax

That's still quite a lot

Geobits

Just think of bats as rodents with wings then >_>

primo

rodents are indeed double that

Nathan Merrill

the number of species is fairly arbitrary: just take all of the bats and subdivide all of their species in two

primo

i suppose it would depend heavily on the stance taken

can mate, but don't, ring species, etc.

trichoplax

17:53

TIL we are more closely related to bats than to rodents

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Fun fact: [dno] matches republican candidates but not democrat ones.

Martin Büttner

Hey if anyone has been wondering how beeswax works (the second hexagonal 2D language), M L added a fairly comprehensive explanation to his FizzBuzz solution: codegolf.stackexchange.com/a/67675/8478

Nathan Merrill

@CᴏɴᴏʀO'Bʀɪᴇɴ in all of history? what about ones that have dropped out?

Martin Büttner

@Geobits That's kinda funny, because the German word for "bat" is "Fledermaus" which basically means "flutter mouse".

Nathan Merrill

if your regex only includes 7 strings, then it's fairly easy :P

Cᴏɴᴏʀ O'Bʀɪᴇɴ

17:55

@NathanMerrill current 2016 candiates.

It's still kind of interesting/

Roujo

@MartinBüttner In French it's "chauve-souris", which is literally "bald-mouse"

Martin Büttner

That makes even less sense.

Roujo

¯\_(ツ)_/¯

Nathan Merrill

ok, another Tournament question: Most of the time elimination tournaments only allow games of 2 teams

how would an elimination tournament work if 3 teams played each other?

Roujo

I'd see it being the same way, only with three branches leading up to the finals

Nathan Merrill

17:59

well, it gets confusing with double elimination

Roujo

It'd be harder to represent on paper, but in concept I think it works

trichoplax

Unless you want to only eliminate last place each time

Nathan Merrill

or that too

trichoplax

That way you don't need a power of 3 players

Although I suppose you don't anyway

Roujo

@NathanMerrill Right. Are the games "someone wins and the others lose", or is it more like "someone wins, someone comes in second and someone comes in third"?

trichoplax

18:00

Group everyone into 3s. Eliminate last place. Repeat with the remaining 2 thirds in new groups

Nathan Merrill

@Roujo yes, there are rankings

VoteToClose

5000 rep! :D

And almost the silver CG badge!

Nathan Merrill

ok, so lets say there's a double elimination bracket that pits 11 teams against each other

primo

how about 12 instead?

Nathan Merrill

top two go on

no, 11

I want to break this

trichoplax

18:02

What does double elimination mean here?

Nathan Merrill

if you are a loser, then you move to a different bracket

primo

lose again, you're out?

Nathan Merrill

yes

trichoplax

Ah - so you get an overall ordering?

So would you start with 3 teams of 3 and one team of 2? Or would you just have 3 teams of 3 and the spare 2 don't play in the first round (safe from elimination)?

Nathan Merrill

I'd do a team of 2 if the game allowed it

so I have to consider both possibilities

[1,2,3], [4,5,6], [7,8,9], 10, 11

Cyoce

18:05

Guys! You give me n, I give you f(n), you find f!

Nathan Merrill

say that lower numbers always win

next round:

primo

if f(n) polynomial?

trichoplax

If the game does allow it, I guess you just need to decide how to score the team of 2. 1st and 2nd place? 1st and 3rd place? 2nd and 3rd place?

Nathan Merrill

oooh, good point

trichoplax

I'm not sure which way would avoid giving an advantage/disadvantage for being in the team of 2

Nathan Merrill

18:07

I'd probably do 1/3

it'd mean higher stakes, but I think I'm ok with that

primo

or just say the two top seeded get a by on the first round

trichoplax

@Cyoce is f deterministic?

@Cyoce f(0)

Cyoce

@trichoplax 0

trichoplax

f(1)

Martin Büttner

what's the domain of f?

Cyoce

18:09

@trichoplax 0

trichoplax

f(f(f(f(2))))

RikerW

@CᴏɴᴏʀO'Bʀɪᴇɴ if you don't link python, why is oration written in it? :P

Nathan Merrill

so, [1,2,3], [4,5,6], [7,8,9], [10, 11] -> [1,4,7], [10,2,5], 8 and [3,6,9], 11

Cyoce

@MartinBüttner ℝ (limited by JS floating-point precision)

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@RikerW link? :/

Martin Büttner

18:10

@Cyoce f(-1)

Cᴏɴᴏʀ O'Bʀɪᴇɴ

f(Math.PI)

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

The greatest power... code-golfnumber-theorysequence Yeah, I'm still trying to come up with a better title... A positive integer n is a perfect power of order k if it can be written in the form mk for some integer m. The greatest power of n is the largest number k for which n is a perfect powe...

Cyoce

@trichoplax 0

@MartinBüttner 0

Martin Büttner

f(0.5)

Nathan Merrill

bah, then the loser bracket gets really confusing

Cᴏɴᴏʀ O'Bʀɪᴇɴ

18:12

f(Infinity)

f(NaN)

f(123.3456)

Lynn

Oh nooo

I finally got a gold badge...

Cyoce

0
Infinity
NaN
94

trichoplax

@NathanMerrill Is your final result going to be a strict ordering of all the players?

Cyoce

(respectively with the last 4 n)

Geobits

@Lynn Oh nooo? Most people like getting those ;)

Cᴏɴᴏʀ O'Bʀɪᴇɴ

18:13

f(Math.PI)

primo

94 exactly?

Nathan Merrill

@trichoplax hmmm. typically KoTHs do order players, so, yes?

Lynn

I was previously the highest-rep user without any gold badges, which was funny.

Cyoce

@CᴏɴᴏʀO'Bʀɪᴇɴ 0

Lynn

Now I'm just not special at all...

Geobits

18:14

Ah, I see. Well, them's the breaks, kid.

trichoplax

@NathanMerrill Just trying to take a step back to see what's required in case that helps

Cᴏɴᴏʀ O'Bʀɪᴇɴ

f(10000)

g2g bai

taking bio test

Geobits

Now it looks like @flawr is the highest rep user with no golds.

Nathan Merrill

@trichoplax I'm trying to design something that's flexible. So far, I don't know of any KoTHs that have elimination tournaments, but its certainly not unfeasible

Lynn

@Cyoce Array(20).fill().map((_, i) => f(i))

trichoplax

18:16

@NathanMerrill I'm interested in this myself, particularly for the general case of more than 2 players per game. I'd like to make KotHs with an arena with many players, but if there are more players than can reasonably fit in one arena, it would need to be a tournament

Lynn

Also, Array(20).fill().map((_, i) => f(0.1*i))

primo

small values seem to be zero

trichoplax

@Lynn The domain is the reals

Cyoce

@Lynn [0, 0, 1, 1, 2, 4, 0, 2, 5, 4, 4, 7, 6, 5, 4, 0, 0, 8, 15, 17]

primo

i is the counting var?

Lynn

18:18

That’s [f(0), f(1), ... f(19)]

Geobits

f(123.3456) = 94 seems so weird.

trichoplax

@Lynn Ignore me - I just realised that isn't the imaginary i - I need more sleep :)

primo

i imagine it's something to do with the digital sum

Nathan Merrill

@trichoplax well, are you planning on making it an elimination challenge?

Cyoce

@Geobits Almost... random ? ( ͡° ͜ʖ ͡°)

primo

18:19

f(123456)

Nathan Merrill

or more of a "take a random sample of X players N times, and sum their scores"?

primo

94, right?

Lynn

f(9999999999)

primo

oops, i meant f(1233456)

trichoplax

@NathanMerrill I'm guessing elimination would take fewer games to give a reasonable result, but I'm not sure

primo

18:20

double 3 <_<

Cyoce

@primo 302944

@Lynn 6492494190

primo

/me is stumped

trichoplax

@NathanMerrill Ideally I'd run more than one tournament, with the groups shuffled, but within one tournament it would be elimination

f(phi)

(golden ratio)

Nathan Merrill

well, player order is always shuffled, so you don't need to worry about that

len(str(n^n))

or something like that

trichoplax

I mean so that a player finding themself in a tough group has another chance next tournament

Nathan Merrill

18:26

right, but running multiple tournaments means that each time the player is in a different group

trichoplax

yes - that's all I was looking for

Is there anything simpler about 3 than about the general case for groups of n?

Nathan Merrill

no, I was doing 3 because its easy to communicate groups that way

but I hope to support groups of N

and M-elimination

and top K move on

trichoplax

With K=N-M?

Nathan Merrill

no

M is single elimination, double eliminiation, so on

K simply needs to be less than N

Cyoce

@trichoplax 0

trichoplax

18:30

f(phi)

Cyoce

@trichoplax still 0

trichoplax

Just checking - you didn't confirm that it was deterministic :P

Nathan Merrill

@trichoplax you still don't know

trichoplax

Indeed.

Cyoce

@trichoplax I think.

trichoplax

18:32

That's even more confusing... :)

Nathan Merrill

ok, how about this: order players, make as many groups of N as possible, make a group of remaining players, if possible.

ordering orders by (num 1st places, num 2nd places, num 3rd places) and so on

Cyoce

Nathan Merrill

ties are determined randomly

Cyoce

if that answers your question

trichoplax

Do you always want to favour N as the group size, with only one small group, or would you consider bringing the small group up to N-1 but reducing some of the other groups to N-1?

Nathan Merrill

18:34

@trichoplax favor N

that's what all of my other tournaments have been doing, so I'll keep it that way

however, I think I am going to favor the small group, and if the group is of size 4, and 5 people move on, then all 4 will move on

in places 1,2,3,4

f([])

@HelkaHomba welcome stranger

Dennis

@Cyoce That reminds me of a definition of insanity.

Nathan Merrill

@Dennis people who think everything is deterministic are likely insane

:P

Dennis

Looks like you're a scientist.

How could you choose avoiding a little pain over understanding a magic lightning machine?

TimmyD

@NathanMerrill Doing something like 3 points for a win, 1 point for 2nd place, and 0 points for 3rd place seems like it would work. You wouldn't need to do a double-elimination, just play two rounds of shuffled teams and take the top X points to move on to the next stage.

Adnan

Hahaha

Nathan Merrill

18:42

@TimmyD that's not an elimination tournament then

TimmyD

Sure it is.

You're just counting points rather than W/L

Which makes the calculations easier.

PhiNotPi

I want in on this KOTH discussion.

Nathan Merrill

the structure is dramatically different than an elimination tournament

so, if somebody wanted a tournament in that structure, they can implement the interface

TimmyD

I'm failing to see how. If someone wins twice, they'll have 6 points and will move on. If someone loses twice, they'll have 0 points and be eliminated.

PhiNotPi

We are trying to figure out how to make a 3-way elimination tournament, or just how to rank players in general?

TimmyD

18:45

X group of players split into Y games. Each game has 3 players. Double elimination.

Nathan Merrill

and elimination tournament with matches of an arbitrary size

trichoplax

So the tournament structure we're working on involves dividing into groups of N, and no one competes against anyone outside their own group, until next round when new groups of N are chosen from the survivors?

TimmyD

Right, 3's just an example.

Liam

http://meta.codegolf.stackexchange.com/a/8573/45459
Final thoughts before I post anyone?

Nathan Merrill

@TimmyD and if they tie twice?

regardless, with elimination tournaments, if you lose once, you don't play the winners

TimmyD

18:46

That would depend upon the ratio of 1st/2nd/3rd place teams

Nathan Merrill

(unless we are talking about end-tournament scenario)

TimmyD

@trichoplax That sounds really reminiscent of the FIFA schedule for World Cup matches that I brought up earlier.

Nathan Merrill

@PhiNotPi A n-elimination tournament with games that support J to K players, and X players get removed each time

TimmyD

But, again, that FIFA schedule is predicated on games of two players, with one winner and one loser.

Nathan Merrill

(n means single, double, triple, so on)

trichoplax

18:48

I suppose we could do with two approaches, depending on whether or not ties are possible

TimmyD

@NathanMerrill What's defined as a loss here? 2nd place?

Nathan Merrill

@TimmyD depends on how many players move on

which is a parameter to the constructor

so it can be "top 3 move on", or "top 1 moves on"

as long as it is less than the maximum game size

TimmyD

So, that's the same as the points system, then. You're just changing the barrier of how many points are required to advance.

If you say 1st=3, 2nd=1, and 3rd=0 points, and set the advancement at 3, you're saying only the top players from each game move on.

Nathan Merrill

the points system allows a player who lost to play a winner.

TimmyD

I'm not really understanding how? The wouldn't have advanced to the next round, so how would they still be playing?

Nathan Merrill

18:52

you initially said two rounds

TimmyD

Sure. Or five. Or one.

Or twenty.

Nathan Merrill

if its one, then its no different than what I'm saying

TimmyD

Yes. That's why I think we were arguing past each other.

Nathan Merrill

if its two or more, then its dramatically different

TimmyD

We were agreeing without agreeing.

Nathan Merrill

18:54

but if it is just one, then there are no reason to keep points

TimmyD

I wouldn't say "dramatically different" ... it's just a case of "Number of games per round" = 1

No, meaning it could be the same system.

You wouldn't need to implement it twice.

Nathan Merrill

oh, you're wanting additional functionality to the n-elimination system

TimmyD

It's a small change to allow for bigger flexibility.

trichoplax

@TimmyD So yours is the even more general case, of which this is a specific instance?

TimmyD

Right.

PhiNotPi

18:55

TBH, I think elimination systems in general are a terrible way to pick a winner.

4

trichoplax

If it's equivalent for fixed size groups, how does the points system apply to a smaller group? Does it help decide what to do there?

TimmyD

You could even have a toggle for "Shuffle players between games" ... if that's FALSE, then you're at the FIFA style system.

Nathan Merrill

so, with a point system, I would need to assign a point value to each ranking?

TimmyD

@trichoplax It goes back to how you want to count the oddballs. For three-player groups with a two-player leftover, is it 1st/3rd? 1st/2nd? Etc.

trichoplax

@PhiNotPi What would you recommend for a game played on an arena which can only handle N players, when there are more than N entries? Ideally I'd like something that gives an accurate order, without taking huge amounts of time.

Nathan Merrill

18:57

@trichoplax I like the sample N, average your scores

I've already implemented that

trichoplax

@TimmyD Yes I'm still undecided on this. I'm trying to think of a way of telling whether there's a single best way, or whether it depends on the type of game.

Nathan Merrill

and it makes the number of games played as even as possible

TimmyD

@NathanMerrill Could be as simple as "point = inverse of rank" so 1st place gets X points, where X is the number of players in the game.

2nd gets X-1

etc.

If that's too linear, you could weight it slightly, such as 1st place gets X+2, 2nd gets X, 3rd gets X-2, and so on

So the "better" players stand out more.

PhiNotPi

@trichoplax I would recommend something like this: meta.codegolf.stackexchange.com/a/4769/2867

Transcript for