This "sandbox" is a place where Code Golf users can get feedback on prospective challenges they wish to post to main. This is useful because writing a clear and fully specified challenge on your first try can be difficult, and there is a much better chance of your challenge being well received if you post it in the sandbox first.

Sandbox FAQ


To post to the sandbox, scroll to the bottom of this page and click "Answer This Question". Click "OK" when it asks if you really want to add another answer.

Write your challenge just as you would when actually posting it, though you can optionally add a title at the top. You may also add some notes about specific things you would like to clarify before posting it. Other users will help you improve your challenge by rating and discussing it.

When you think your challenge is ready for the public, go ahead and post it, and replace the post here with a link to the challenge and delete the sandbox post.


The purpose of the sandbox is to give and receive feedback on posts. If you want to, feel free to give feedback to any posts you see here. Important things to comment about can include:

  • Parts of the challenge you found unclear
  • Comments addressing specific points mentioned in the proposal
  • Problems that could make the challenge uninteresting or unfit for the site

You don't need any qualifications to review sandbox posts. The target audience of most of these challenges is code golfers like you, so anything you find unclear will probably be unclear to others.

If you think one of your posts requires more feedback, but it's been ignored, you can ask for feedback in The Nineteenth Byte. It's not only allowed, but highly recommended! Be patient and try not to nag people though, you might have to ask multiple times.

It is recommended to leave your posts in the sandbox for at least several days, and until it receives upvotes and any feedback has been addressed.


Search the sandbox / Browse your pending proposals

The sandbox works best if you sort posts by active.

To add an inline tag to a proposal, use shortcut link syntax with a prefix: [tag:king-of-the-hill]. To search for posts with a certain tag, include the name in quotes: "king-of-the-hill".

Get the Sandbox Viewer to view the sandbox more easily!


4125 Answers 4125


Output a unique sign sequence

  • 3
    \$\begingroup\$ Is shortness of the answer not supposed to be a factor at all? That is, we don't try to golf our answers? Even more, are voters supposed to consider just the sequence, or the code producing it? I'm not quite sure if "How clever is the implementation" is meant to apply to the code itself or just the general strategy. \$\endgroup\$
    – xnor
    Sep 22, 2020 at 23:14
  • \$\begingroup\$ @xnor No, golfing your answer is unimportant, but perfectly acceptable. Voters should consider both the code and the sequence, but primarily the sequence. The code should mainly be considered for how "elegant/clever/good" the vote perceives it to be. "How clever is the implementation" kinda applies to both the code and the strategy. Using language-specific tricks are a reason to consider code "clever" as is implemented a fully generalised method that any language to pick up and use with no trouble \$\endgroup\$ Sep 22, 2020 at 23:34
  • 2
    \$\begingroup\$ I think this works much better as answer-chaining than as pop-con. As evidence I'd say that almost half of the challenge being devoted to explaining the voting rather than the challenge isn't a good sign. I think looking at recent successful pop-cons shows that pop-cons are generally only good if the task just won't lend itself at all to another scoring criterion. This, at least to me, seems like a great idea for answer-chaining. \$\endgroup\$ Sep 24, 2020 at 19:08
  • 2
    \$\begingroup\$ @FryAmTheEggman I do dislike the fact that this is a pop-con, and the only reason it is, is because I couldn't find any other criteria which would inspire the creativity in sequences I'm looking for (e.g. the reductive attitude of code golf would make this very boring). The issue I see with making it answer chaining is that the winning criteria is a bit iffy. "Second-to-last" isn't the best winning criteria as already discussed on meta, but it gets even worse when new answers aren't more difficult than previous ones, as the challenge will continue ad infinitum until (cont) \$\endgroup\$ Sep 24, 2020 at 19:19
  • 1
    \$\begingroup\$ (cont) people get bored and stop posting, rather than new answers are genuinely too difficult to make. There are an infinite number of sign sequences, so "running" out isn't an issue, and each new answer isn't more difficult because there are so many to choose from. Furthermore, I don't think it'd inspire the creativity I'm looking for. I don't think it'd be as bad as, say code golf, but I feel like people will resort to posting basic sequences in order to keep the chain from dying, thus extending it way beyond being fun \$\endgroup\$ Sep 24, 2020 at 19:20
  • \$\begingroup\$ I had initially thought that preventing the obvious abuse cases was relatively simple (no prefix/suffixes and not periodic) but I've realised that it would actually be much more difficult than that. I still think this probably works better if massaged into something that works for answer-chaining, but you are probably right that it isn't as good a fit as my first impression was. \$\endgroup\$ Sep 24, 2020 at 20:36
  • 1
    \$\begingroup\$ This might be interesting as a cops-and-robbers challenge, though it would have to be a finite slice of an infinite sequence in that case. \$\endgroup\$
    – Beefster
    Oct 2, 2020 at 19:21
  • \$\begingroup\$ @Beefster I think the only (or maybe best) way this could work as a CnR is where the cops reveal their code+language (and maybe a few terms of the sequence) and the robbers have to find the generating function for that sequence, but that would almost definitely lead to obfuscated code, rather than „how is the sequence generated“ being the focal point for answerers, which pivots away from the main idea of the challenge being interesting sequences, not interesting code \$\endgroup\$ Oct 2, 2020 at 19:51

Calculate longest Stack Exchange streak

  • \$\begingroup\$ I'd say the best input format would be nested maps/dictionaries/whatever in the language's most normal data structure with the schema given in the input format, or alternatively (i.e. leave it up to the answerer), the exact text format you gave above (minus the var visited = ). \$\endgroup\$
    – pxeger
    Mar 24, 2021 at 14:05
  • \$\begingroup\$ I think for a internet challenge in general it is fine to ask for the input data to be parsed after collecting it from the internet. I'd expect most of the variance to come from trying to find shortcuts in parsing in this case anyway. Finding users who are good test cases might be tricky - so perhaps including some made-up results which test basic stuff (skipped months/years, February) could be nice. I don't know that CGSE is new enough to not have any weird calendar stuff (missing days, etc.) that should be specified. \$\endgroup\$ Mar 24, 2021 at 22:15
  • \$\begingroup\$ Not sure why, but if I go to https://codegolf.stackexchange.com/users/daily-site-access/66833 (my user id is 66833), I just get a blank page? Challenge-wise, I think that this is pretty clear, and I agree with the other commenters that the default format is perfectly fine. You are missing an objective winning criteria however \$\endgroup\$ Mar 30, 2021 at 21:39
  • \$\begingroup\$ Thanks for your input! @ChartZBelatedly it's a blank page indeed, but viewing the source helps. It was intended to be a [code-golf] challenge, I assumed that was the default. I'm missing test cases too, of course. \$\endgroup\$
    – Glorfindel
    Mar 31, 2021 at 9:01
  • \$\begingroup\$ I've edited this down to a stub now that it's been posted to save space \$\endgroup\$ Apr 6, 2021 at 1:04

Did I die or not?

  • 1
    \$\begingroup\$ Perhaps add a few cases to make sure solutions check for exactly 12 wrong guesses? Like "ABCD...Z", "L" -> True and "ABCD...Z", "M" -> False \$\endgroup\$ Apr 20, 2021 at 19:38
  • 1
    \$\begingroup\$ @ManishKundu Added \$\endgroup\$ Apr 20, 2021 at 23:32

It almost works (C&R)

Cops thread

Robbers thread

  • \$\begingroup\$ Looks really fun! \$\endgroup\$
    – mathcat
    Jul 12, 2021 at 14:00

Laggy text editor

  • \$\begingroup\$ Your example only submit an input after hitting Enter. Am i required to follow this behavior or may I consider each key press as an input? \$\endgroup\$
    – tsh
    Mar 1 at 7:50
  • \$\begingroup\$ @tsh I'm not quite sure how to deal with "prompting for input" in general, but I guess you can take each key press as an input if you want. \$\endgroup\$
    – emanresu A
    Mar 1 at 7:56
  • \$\begingroup\$ A fiddle which delay every input 5 second \$\endgroup\$
    – tsh
    Mar 1 at 8:38
  • \$\begingroup\$ @tsh Thanks! (filler) \$\endgroup\$
    – emanresu A
    Mar 1 at 8:49

Fill in the next numbers

  • 1
    \$\begingroup\$ Perhaps a better title would be "Fill in the gaps with previews" (previews because the gaps contain the next number to be seen, so the gaps "preview" the sequence) but it also sounds rather strange. \$\endgroup\$
    – ophact
    Feb 26 at 14:59
  • 2
    \$\begingroup\$ I would like to exclude the "prime" part in the example. It makes me confuse this challenge is about doing something to test a number is prime or not. \$\endgroup\$
    – tsh
    Mar 3 at 2:11
  • 1
    \$\begingroup\$ Also, instead of show gaps as numbers, underscore could be less confusion, as _ 2 3 _ 5 _ 7 _ _ _ 11. And say any gaps (underscore) should be filled by right hand values. \$\endgroup\$
    – tsh
    Mar 3 at 2:12
  • 1
    \$\begingroup\$ I agree with @tsh that I wouldn't mention 'primes' since it might lead to confusion. The example could still be 2,3,5,7,11 and people can see for themselves they're primes, but I just wouldn't mention this explicitly. So the sentence "For example, if the input is the first 5 prime numbers:" would be: "For example, if the input is:" and # primes in the code block would be # input. Apart from that it's a great challenge, and I've already prepared a solution for when it goes to main in the near future. :) \$\endgroup\$ Mar 3 at 13:41

Eh, codegolf shmodegolf

  • 2
    \$\begingroup\$ Nice idea! You probably want a general procedure people can follow to do this, and I don't think the initialisms is a good idea. \$\endgroup\$
    – emanresu A
    May 18 at 10:34
  • \$\begingroup\$ @emanresuA You're probably right about the initialisms. I added an example of an ungolfed python function that solves the challenge, and some restrictions on the input. Do you think more things should be specified? \$\endgroup\$
    – JSorngard
    May 19 at 10:34
  • 1
    \$\begingroup\$ This looks fine! Leave it for another day or two though just in case. \$\endgroup\$
    – emanresu A
    May 19 at 11:04
  • 1
    \$\begingroup\$ I suggest adding some more test-cases (especially starting with more than one consonant). Also, re: last sentence - definitely specify what to expect as input (only letters? at least one vowel? will the input start with a vowel?) and add some test-cases accordingly if necessary. In the end, please clean up the post (one can view edit history if interested in the evolvement of the question), so we can see what you intend to post and maybe give some feedback on that. \$\endgroup\$
    – pajonk
    May 19 at 12:02
  • 2
    \$\begingroup\$ I recommend stating the rules of shmification used in this challenge explicitly - I guess this will be three points (to present briefly): 1. consonants up to first vowel replaced with shm; 2. prepend shm to words beginning with vowel; 3. no vowels = no shm. Also, I suggest altering the example program to actually solve the challenge or removing it altogether to avoid confusion. \$\endgroup\$
    – pajonk
    May 19 at 17:43
  • \$\begingroup\$ @pajonk Thank you, excellent points. I have added rules and updated the code to solve the challenge as it is stated. \$\endgroup\$
    – JSorngard
    May 20 at 10:48
  • \$\begingroup\$ You should specify that vowels are specifically aeiou (and not y). Also, can we assume the input only contains lowercase letters? \$\endgroup\$
    – pxeger
    May 20 at 14:54
  • \$\begingroup\$ @pxeger The input being lowercase letters is already specified, but I moved it and added a line about what the vowels are, thanks :D \$\endgroup\$
    – JSorngard
    May 20 at 19:33
  • \$\begingroup\$ Maybe some words on output formats? People will wonder whether they can output as a list with two elements, a function that returns two words, or whether it has to be a single string joined with a space. \$\endgroup\$
    – chunes
    May 22 at 7:28
  • \$\begingroup\$ @chunes I changed the output to be a single string containing the reduplicated version of the input, that way it is clearer what format the output should be. \$\endgroup\$
    – JSorngard
    May 23 at 8:55
  • \$\begingroup\$ After the change, rule 3 no longer makes sense. Instead of saying "only repeated," maybe "unchanged" or "leave it alone"? \$\endgroup\$
    – chunes
    May 23 at 9:01
  • \$\begingroup\$ @chunes You are completely correct, thanks! I have changed rule 3 to say that words without vowels should be returned unchanged. \$\endgroup\$
    – JSorngard
    May 23 at 9:15
  • 1
    \$\begingroup\$ I've edited this down to a stub now that it's been posted to save space \$\endgroup\$
    – Steffan
    May 24 at 18:02
  • \$\begingroup\$ @Steffan thanks! \$\endgroup\$
    – JSorngard
    May 24 at 20:11

Collatz Encoding

  • \$\begingroup\$ What should the input 0 output? \$\endgroup\$
    – Alan Bagel
    Apr 29 at 14:08
  • \$\begingroup\$ @BgilMidol "Your challenge is to write a function or program which takes an integer (n>0) as input" i.e. 0 will never be an input \$\endgroup\$ Apr 29 at 14:08
  • \$\begingroup\$ oh, i didn't see that \$\endgroup\$
    – Alan Bagel
    Apr 29 at 14:16

Make Spanish from Latin

Note: In the final challenge \$N\$ will be a concrete number (I am thinking about 100), but while this is in the sandbox it is subject to change so I have left it as \$N\$. Currently \$N=133\$

This challenge is based off of a list of \$N\$ Castilian Spanish words and the words they originate from.

You are to write a program or function which takes the origin word as input and outputs as close as possible the Castilian derivative. Your program should be no longer than \$N\$ bytes.


To calculate your score run your program on every origin word and calculate the distance between your output and the correct answer. Your score is the sum of all these distances.

The distance here is a modified version of Levenshtein distance. It is the same as Levenshtein distance except replacement steps that add or remove a diacritic cost only 1 of a step as opposed to their normal 2. This is because if you guessed á instead of a that's not as bad as if you guessed f instead of a.

You can use this code to calculate the distance between two strings.

The goal is to have as low a score as possible.

About the list

All of the origin words, spare 2, are Latin words (Late or Classical depending on the word). The two exception is ezkerra (the origin for izquierda) which is of Basque origin and brixta (the origin for bruxa) which is of Proto-celtic origin. It has been added as an extra curve-ball in case you can get all the others with a little space to spare.

Verbs are always in the infinitive form and nouns in the nominative singular.

The words are not chosen randomly but rather I have focused on choosing words that follow a number of simple rules. The list is also organized so that words that undergo similar transformations are grouped together. This is for your ease of use, nothing more.

The list

imperatrix, emperatriz
cicatrix, cicatriz
actrix, actriz
matrix, matriz
carex, carrizo
radix, raíz
falx, hoz
fovea, hoyo
formica, hormiga
fodiare, hozar
folia, hoja
filum, hilo
filius, hijo
filia, hija
ficum, higo
filare, hilar
ficcare, hincar
afflare, hallar
fundus, hondo
profundus, profundo
fungus, hongo
fabulare, hablar
furnus, horno
bufus, búho
ferrum, hierro
fetere, heder
fagea, haya
faba, haba
facienda, hacienda
facere, hacer
fastidium, hastío
fastidiare, hastiar
factor, hechor
factum, hecho
tectum, techo
octo, ocho
octavus, ochavo
noctu, noche
lacte, leche
iactare, echar
coctus, cocho
dictatum, dechado
phalanga, palanca
capere, caber
sapere, saber
lupus, lobo
lacrima, lágrima
lacuna, laguna
eruca, oruga
pater, padre
mater, madre
liber, libro
thema, tema
theatrum, teatro
thesaurus, tesoro
thesis, tesis
thorax, tórax
aether, éter
anthropologia, antropología
orthographia, ortografía
sapphirus, zafiro
philosophia, filosofía
echo, eco
chalare, callar
chamaeleon, camaleón
chaos, caos
materia, materia
resistentia, resistencia
aurum, oro
taurus, toro
autumnus, otoño
canna, caña
annus, año
ungla, uña
plangere, plañir
stringere, estreñir
signa, seña
scribere, escribir
stare, estar
stabulare, estabular
stipare, estibar
stria, estría
strix, estrige
studiare, estudiar
stimulare, estimular
stillare, estilar
sternutare, estornudar
statuere, estatuir
spirare, espirar
sputare, esputar
spina, espina
speclum, espejo
spectrum, espectro
squama, escama
squamosus, escamoso
squalidus, escuálido
scalare, escalar
scandere, escandir
scutum, escudo
scutella, escudilla
scriptor, escritor
sobrina, sobrina
sobrinus, sobrino
secta, secta
sector, sector
sibilare, silbar
subire, subir
subito, súbito
suspirum, suspiro
subiugare, subyugar
corvus, cuervo
fossa, huesa
socrus, suegro
lupus, lobo
acutus, agudo
tecula, teja
apicula, abeja
flamma, llama
flammare, llamear
planus, llano
plagare, llagar
plorare, llorar
plovere, llover
plicare, llegar
clavis, llave
clamare, llamar
gladius, gladio
gleba, gleba
globus, globo
blandus, blando
blancus, blanco
brixta, bruxa
ezkerra, izquierda
  • \$\begingroup\$ I think it's interesting in that it should be near impossible to get a perfect score without built-ins. As a suggestion I'd remove the non-ASCII words, or at least normalise them, and perhaps not let \$N\$ be too high. Also, I wonder what the default cat program would be. \$\endgroup\$
    – Jo King Mod
    Mar 3, 2020 at 12:31
  • 2
    \$\begingroup\$ @JoKing I am looking to somewhat twart perfect scores, I feel there should always be some room for improvement, It just is a little hard to balance this with golfing-languages ability for expressiveness. I am interested to hear what ranges for \$N\$ you think are too high. I started out by avoiding any non-ASCII characters, but it was really hard to build up a representative corpus of words. Plus the accents and eñe really are a feature of the language. I may adjust the scoring so that i and í for example are only half away from each other so that the penalty is small. \$\endgroup\$
    – Wheat Wizard Mod
    Mar 3, 2020 at 13:05
  • \$\begingroup\$ Suggested title: Hispanize these words. Also mention that words should be in lowercase. Finally I think it makes more sense to restrict programs to \$N\$ characters rather than \$N\$ bytes. \$\endgroup\$
    – SunnyMoon
    Nov 20, 2020 at 15:24
  • 7
    \$\begingroup\$ I'm not a fan of "Your program should be no longer than N bytes". I think a scoring rule which incorporates both the Levenshtein distance and the byte count would be better, as in the Moby Dick challenge: it would allow more languages to compete, and would allow for more creativity. \$\endgroup\$ Nov 22, 2020 at 23:43
  • 1
    \$\begingroup\$ @RobinRyder I think that metrics that combine two non-obviously related factors such as in the Moby Dick challenge or as you are suggested rarely work. In fact I can't even think of an example I feel is good. They simply require fine tuning that can only really be done in retrospect. I also don't know how this would allow more languages to compete and I certainly don't know how it would "allow for more creativity". If you have a specific metric in mind and a compelling reason why that metric would not be broken I would be happy to try it out and see. \$\endgroup\$
    – Wheat Wizard Mod
    Nov 23, 2020 at 1:38
  • 2
    \$\begingroup\$ Sorry, I'll expand: with a low, hard limit on the number of characters, non-golfing languages don't stand a chance (hence the "more languages" part). If instead you set a high hard limit, it will be possible to reach a (near) perfect score, with no incentive to golf or search for trade-offs (hence the "creativity" part). There surely exists a sweet spot for the character limit where neither of these issues arise, but (a) it will be hard to find and (b) it will be highly language-dependent. \$\endgroup\$ Nov 23, 2020 at 7:53
  • 1
    \$\begingroup\$ @RobinRyder I think that 100 characters is plenty of characters for nearly any language to implement something a bit more complex than cat (e.g. replace ^f with ^h). I also think that it would require quite a few characters any language to acheive a perfect score. More than half of the words have seemly random vowel mutations that are not covered by any general rule. The only way I can see a perfect score is a compression decompression method. It seems to me to write the de-compressor and make up for the loss would require a deal of room. \$\endgroup\$
    – Wheat Wizard Mod
    Nov 24, 2020 at 15:56
  • \$\begingroup\$ What I'm saying is that I think that sweet spot is actually very large. And I think that adding more dimensions to the problem only increase the risk of missing the sweet spot. \$\endgroup\$
    – Wheat Wizard Mod
    Nov 24, 2020 at 15:57
  • \$\begingroup\$ Some possible typos (I'm a native Castillian Spanish speaker and I'm familiar with Latin): afflare, hallar; factor, factor; chalare, callar \$\endgroup\$
    – Luis Mendo
    Dec 20, 2020 at 18:00
  • \$\begingroup\$ @LuisMendo hallar and chalare are definitely typos. But hechor was intentional. I suspect your hang up is that factor and hechor are doublets but factor is more similar to the Latin. There are plenty of words here for which the Latin is not a direct translation or the Spanish or vice versa (even factor is not). I chose to include hechor because it is a good example of some of the changes that I want to highlight, (f -> h and ct -> ch). \$\endgroup\$
    – Wheat Wizard Mod
    Dec 21, 2020 at 23:26
  • 1
    \$\begingroup\$ The thing is, I never heard hechor. But I just checked in the dictionary and it is included, albeit as an old form, not in current use \$\endgroup\$
    – Luis Mendo
    Dec 21, 2020 at 23:46


  • \$\begingroup\$ I quite like this challenge, but why are there four possible outputs for xxxx, X? Was it not specified that the first occurrence should be capitalized and only that one? \$\endgroup\$
    – user100690
    May 27, 2021 at 19:31
  • \$\begingroup\$ I'll change it so that it's always the first option. \$\endgroup\$
    – pxeger
    May 28, 2021 at 7:35
  • \$\begingroup\$ May we take a list of words as input? A list of characters? Or a list of lists of characters/character codes? (I think it's not covered by the I/O link you provided) \$\endgroup\$
    – pajonk
    Jul 5 at 8:38
  • \$\begingroup\$ @pajonk I think it's generally understood that a string and list of characters are interchangeable. But I'll add notes about the other two. \$\endgroup\$
    – pxeger
    Jul 5 at 9:05

When will the DVD logo hit the corner?


Numbers vs. Strings: Language fitness challenge

  • \$\begingroup\$ The thing is, lenguage is quite easy to golf down due to how its encoding works and it'd require a lot of diligence to create one that's almost certainly optimal. The longer you make a solution, the more potential there is for it to be smaller. \$\endgroup\$
    – emanresu A
    Sep 13 at 5:24
  • \$\begingroup\$ May I take input as a list of characters for the string task? \$\endgroup\$
    – alephalpha
    Sep 13 at 9:41
  • \$\begingroup\$ @alephalpha Yes. \$\endgroup\$
    – Bubbler
    Sep 13 at 9:53
  • 2
    \$\begingroup\$ One concern I have about the string task is that it seems relatively easy in a language like BQN, which treats strings as arrays of characters and is good at manipulating arrays. I wouldn't think of BQN as being a good language for string-handling, though. Maybe the string task should involve more of the standard string-manipulation builtins like case conversion, search-and-replace, etc. \$\endgroup\$
    – DLosc
    Sep 16 at 19:20
  • \$\begingroup\$ @DLosc I changed the string task to one that involves string replacement. Does it look better? \$\endgroup\$
    – Bubbler
    Sep 19 at 4:17
  • \$\begingroup\$ @pxeger "Given a nonempty string s consisting of uppercase letters..." \$\endgroup\$
    – Bubbler
    Sep 19 at 10:30
  • \$\begingroup\$ @Bubbler Wow, I'm not good at reading. \$\endgroup\$
    – pxeger
    Sep 19 at 10:31


Many Unicode composed characters have two forms, one being a precomposed character, while the other is an ASCII character with a combining diacritic. For instance, é has Unicode code point U+00E9, but , which looks identical, is actually the ASCII character 0x65 with the combining diacritic U+0301.

Your task is to write two programs or functions. The first will take a string of Unicode containing some composed characters and a string of printable ASCII. It will then output an identical-looking Unicode string which the second program can then decode to recover the printable ASCII.

By identical-looking, it must be the case that performing your choice of NFC or NFD normalisation on the input and output Unicode strings returns an identical string. If necessary, you may require that the input Unicode string be already NFC or NFD normalised (please specify).

If the input string does not contain sufficient composed characters for you to use, you should repeat it until its length is sufficient. You may also join the repeats with newlines should you so wish.

Your score is the sum of the byte lengths of your two programs divided by the number of composed characters you support in the input Unicode string. For instance, you may only wish to support those composed characters that expand to two characters under NFD. In this case your code's behaviour for other composed characters must be consistent (i.e. one of always unchanged, always NFC, always NFD). (Note that the Unicode string may also contain characters that do not change under NFC or NFD. You obviously can't use these to encode the ASCII string but they must still appear in the output.)

If your language supports arbitrary-precision integers, you may also elect to provide a demonstration of your algorithm on integers rather than printable ASCII strings.

  • 1
    \$\begingroup\$ "you may repeat it until its length is sufficient" - Do we have to, or can we assume that there are enough? If it's the first, "should" or "must" would be clearer. \$\endgroup\$ Oct 15 at 18:36
  • 1
    \$\begingroup\$ +1 just because Sʨɠɠan \$\endgroup\$
    – Seggan
    Oct 17 at 14:45
  • \$\begingroup\$ I feel so proud :p \$\endgroup\$
    – Ginger
    Oct 22 at 1:02

PETSCII banner

PETSCII on CodGolf.StackExchange

In an other world... I was using a PET 2001 who used some particular PETSCII charset.

The screen green on black, with 40 columns and 25 lines, was only able to display characters from this charset. No way to draw dots or lines...

But in the chaset, there is some and , which, ( by the use of reverse video in order to obtain 16 chars: ' ','▖','▗','▘','▝','▀','▄','▐','▌','▞','▚','▟','▛','▜','▙','█' ) make us able to draw graphics on a 80x50 dots plan.

Using an internal clock triggering IRQ, I've done a animated prompter like this:

Animated display sample Hello world!

The goal of this is to make a similar banner, with same charset, (but using UTF-8 characters: ' ','▖','▗','▘','▝','▀','▄','▐','▌','▞','▚','▟','▛','▜','▙','█'). Warn, this charset use inverted lower/upper cases.

  • This imply the use of PETSCII charset, I will post them there as a json string, before getting this out of the sandbox if some interest...

  • The tool have to change his position 20 time per second.

  • The tool must accept as argument, the string to display.

  • The tool must add date and time in the form - WDay MDay Mnth Year, HH:MM:SS -

  • Scrolling have to be done bit per bit: I.E.: by half character!

  • Shortest code...
    • -3 if size of console is not limited to 40 columns
    • -5 if cpu usage stay less than 90% (On my poor Core(TM)2 Duo CPU E8400 @ 3.00GHz, with 4G ram)
    • -5+ if cpu usage stay less than 50%
    • -5+ if cpu usage stay under 5%


  • 3
    \$\begingroup\$ as for the CPU bonuses - what is the target environment, what is the smoothing factor, and what processes count against this measure? \$\endgroup\$ Dec 15, 2013 at 6:19
  • \$\begingroup\$ This sandbox post has had little activity in a while and little positive reception from the community. Please improve / edit it or delete it to help us clean up the sandbox. \$\endgroup\$
    – user58826
    Jun 9, 2017 at 15:32

Do-nothing Polyglot

This challenge requires you to write a polyglot which contains a comment in as many languages as possible.

The comment must say This is a polyglot comment.

The program must do nothing at all.

Besides comments, your code may contain no-ops from the language(s) you are using, if it helps incorporate more languages into your answers. For example in bash we could use :. Or in the family, something like the following may be used, as long as it compiles:

if (0) {
    statements with side effects

Since this is a challenge, the answer with the most languages wins, so I think this should also be tagged

  • \$\begingroup\$ I think it is funny that the content of the comment makes it so that H9+ cannot be included. Yes, tag it [code-challenge]. \$\endgroup\$
    – Justin
    Mar 16, 2014 at 2:27
  • \$\begingroup\$ Many languages use # or // to signify comments. Perhaps restricting the counting so that each type of comment is counted only once. In other words, //This is a polyglot comment would only count for 1 language, not fo C, C++, Java, ... \$\endgroup\$
    – Justin
    Mar 16, 2014 at 5:13
  • 5
    \$\begingroup\$ What is a "program" for the purposes of this question? In particular, are compiled languages going to be excluded because the compiler won't be able to find an entry point? \$\endgroup\$ Mar 16, 2014 at 17:33
  • \$\begingroup\$ What does no-ops mean? \$\endgroup\$
    – Mohammad
    Mar 18, 2014 at 17:37
  • 3
    \$\begingroup\$ I think that in some esolangs it's not entirely decided what counts as a comment. Maybe you should say that the code must contain the text this is a polyglot comment. \$\endgroup\$ Mar 16, 2017 at 4:24
  • \$\begingroup\$ Thanks! Posting this (after I edit it) \$\endgroup\$ May 2, 2017 at 20:40
  • \$\begingroup\$ Yeah I removed the post as it was horrible and not well recived \$\endgroup\$ May 16, 2017 at 15:17

You say goodbye, and I say hello

In this challenge, you have to make a program that outputs Hello, World!. That's not all though! When the program is reversed, it has to output Goodbye, World!.

Unfortunately, there are a few boring rules:

  • No using comments, i.e., the Python code print"Hello, World!"#"!dlroW ,eybdooG"tnirp is illegal

This is , so try to do it in the least number of bytes.

(Should the forwards/backwards tasks be changed? I think they are probably too easy. Let me know if you have any suggestions.)

  • 2
    \$\begingroup\$ You're missing a golden opportunity to make a Beatles reference. \$\endgroup\$ Nov 4, 2015 at 18:11
  • 2
    \$\begingroup\$ This is a good point. Typically in good palindrome challenges at least comments should be banned; you can take a look at this for some examples of good rules in a palindrome challenge. \$\endgroup\$ Nov 5, 2015 at 21:06
  • 1
    \$\begingroup\$ @quartata Thank you for the example! This is what the sandbox is good for. \$\endgroup\$
    – DanTheMan
    Nov 6, 2015 at 17:35
  • 1
    \$\begingroup\$ Your answer may...? :) \$\endgroup\$ Nov 8, 2015 at 1:07
  • \$\begingroup\$ Do you think I can just go ahead and post this as a challenge? \$\endgroup\$
    – DanTheMan
    Nov 9, 2015 at 18:05
  • 2
    \$\begingroup\$ What counts as a comment? Unexecuted code? Strings which are pushed and never used? etc etc \$\endgroup\$
    – Sp3000
    Nov 10, 2015 at 10:06
  • \$\begingroup\$ I don't think it's too easy. Easy questions get lots of answers and therefore plenty of upvotes for you. That said, there's probably a duplicate or near duplicate out there somewhere. \$\endgroup\$ Nov 16, 2015 at 10:43
  • \$\begingroup\$ it makes lots of language impossible. Maybe comment have penalty \$\endgroup\$
    – l4m2
    Dec 31, 2017 at 14:43

What shape is that polygon?

Given an ASCII figure made of /, \, |, and _, write a program that determines (1) if the figure is a closed, non-self-intersecting polygon (meaning it has at least three sides that connect to form a loop), and if so, (2) how many sides the polygon has and (3) if the polygon is convex (all internal angles are less than 180˚) or concave (not convex, note that no polygons have 180˚ angles).

The multiline input is the ASCII figure. It can be assumed that the input only contains the four characters listed above.

If the figure is not closed or self-intersects (or has more than 2 sides connecting at the same point), the program should print Not a polygon.

The number of sides the polygon has is determined by the following chart.

Sides   Name
3       Triangle
4       Quadrilateral
5       Pentagon
6       Hexagon
7       Heptagon
8       Octagon
9       Nonagon
10      Decagon
11      Undecagon
12      Dodecagon
>12     n-gon (where n is the number of sides)

The output should be in the format <concavity> <polygon name> if (1) is satisfied.

Test cases

 /  \__

==> Concave hexagon


==> Concave 17-gon


==> Not a polygon

/ \
| |

==> Convex octagon

I need a little assistance on getting stricter definitions of points intersecting, so help would be appreciated. Thanks!

  • 1
    \$\begingroup\$ I like the concept of this challenge. A good test case would be including a line that goes inside the polygon as well as lines that doesn't even touch the polygon (unless you are going to guarantee that they will touch) \$\endgroup\$ Dec 6, 2015 at 19:45
  • \$\begingroup\$ Your first two examples look wrong: those polygons are clearly concave. \$\endgroup\$ Dec 7, 2015 at 9:45

Capsa, a card game KotH!

Capsa, known by many times, including the name Big Two in English, is a popular card game in East Asia and South East Asia, especially throughout China, Hong Kong, Indonesia, Macau, Malaysia, Singapore and Taiwan. There are many variations and house rules. The rules of this particular variation are thus:

  • This game will be played between exactly four bots. The cards are dealt between everyone, so that everyone has 13 cards.
  • Rank is ordered with 2s before As as follows: 2, A, K, Q, J, 10, 9, 8, 7, 6, 5, 4, 3, hence, the name Big Two.
  • Suits are ordered as follows: Spades, Hearts, Clubs, Diamonds
  • Ordering runs rank first then suit. Thus, 7 of Hearts > 7 of Diamonds > 6 of Hearts, the lowest card is the 3 of Diamonds, and the highest, the 2 of Spades

You can play cards in sets of one, two or five cards, (singles, pairs, or five-card poker hands). Each set must always be bigger than the one before.

With singles, play may proceed in this way: 1 3H, 2 4H, 3 5D, 4 5C, 1 10S, 2 Pass, 3 JD, 4 Pass, 1 2S, 3 Pass

Pairs are ordered by the higher suit in the pair. 6C6H is of a lower rank than 6D6S. Play may proceed in this way: 3 3D3C, 4 6D6H, 1 6D6S, 2 JHJS, 3 Pass, 4 ADAC, 1 Pass, 2 Pass

Poker hands are ordered in the following way, from lowest in rank to highest in rank:

  • Straight: Five cards that are consecutive in rank, e.g. 6H 7S 8D 9H 10C or JS QH KD AS 2C. Rank is determined by the highest card, with suit used as a tie-breaker.
  • Flush: Five cards with the same suit, e.g. 5H 7H 10H QH AH Rank is determined first by suit, then by highest card.
  • Full house: A three of a kind with a two of a kind, e.g. JD JS JH 3D 3S. Rank is determined by the triple, without regard for the pair.
  • Straight flush: Five cards that are consecutive in rank and are all of the same suit, e.g. 6H 7H 8H 9H 10H. Ranked the same as straights, with suit as a tie-breaker
  • Four-of-a-kind: Four cards of the same rank, with any 5th card, e.g. 9D 9C 9H 9S 4S This hand is known as the bomb. Wins any round of poker hands it is played in, unless someone else plays their own bomb.

Any five-card hand that is higher than the previous five-card hand played is eligible. For example, you can play a full house on a straight.

Rules of play:

  • At the beginning of the game, the cards are dealt between everyone, so that everyone has 13 cards. The first player is the one that holds the 3 of Diamonds and they must play this card first, whether singly, with another 3 in a pair, or in a poker hand.
  • A round begins with the first player playing a single card, a pair or a poker hand. Every other player either respond with the same number of cards (you must play singles on singles, never a pair or a poker hand), or that player passes for that round.
  • A player may pass even when they have a playable card, but they must pass if none of their cards are high enough in rank, that is, if they have no playable cards. For example, in a round of poker hands, if you only have a flush as a five-card set, but another player has already played a full house, you must pass, as your flush is too low.
  • The round ends when all but one player passes. That last player wins the round and starts the next round.
  • The game ends when one player has played all of their cards.

Possible scoring systems:

  • Bots are judged by the number of games they win over (TBD) games that they play. So if we have seven bots to test, we'll play them until every bot has played more than (TBD) games.
  • Bots are judged by the number of cards they have left at the end of the game. (The winner will obviously have 0 cards at the end of the game).

Here is the link to a preliminary controller and two bots to test against.

  • 1
    \$\begingroup\$ "The cards are dealt between everyone, so that everyone has 13 cards" belongs in the "rules of play" section, even if that means having to repeat it. \$\endgroup\$ Dec 30, 2015 at 11:09
  • \$\begingroup\$ @PeterTaylor Done. If you have any comments on the controller linked at the bottom of this post, I'd love to hear them. Thanks \$\endgroup\$
    – Sherlock9
    Dec 30, 2015 at 11:27

Let's Play Unikong

In honor of April Fool's day, we shall have an epic battle to see who can play Unikong best. Or, rather, whose program can play it best.


Write a program in any language to play the game Unikong. It should seek to try and score as high as possible.


Whomever's program has the highest average score, over 10 games wins.


Standard rules apply. Additionally, your program must actually play the game, not change the score variable by some other means, and you can read the variables from the game to avoid trolls and downvotes, but not modify any variables to make it easier. I will run the tests myself and will use the first 10 runs to calculate the score. Please include any specific instructions needed to run your program.

Notes: This would be my first question, so hopefully I got the format right. Are the rules clear enough? Do I need to clarify anything?

  • 2
    \$\begingroup\$ Do not require a video, run the programs yourself and see how far they get without your input. No pressing continue. \$\endgroup\$ Apr 1, 2016 at 17:09
  • \$\begingroup\$ Uh, will that link exist beyond today? If not then I think you'd need to find a way to replicate it (not sure about copyright here...). Anyway, I'm not really sure what counts as cheating. I think it will be hard to ban everything that's bad, so it might be better to write a controller that only allows looking at certain variables and only allows the basic player input. \$\endgroup\$ Apr 1, 2016 at 17:09
  • \$\begingroup\$ @FryAmTheEggman I've changed the link to something that will remain after April Fool's. \$\endgroup\$ Apr 1, 2016 at 22:02

Reverse the lengths


In this challenge, your task is to reverse the lengths of an array-of-arrays, while keeping its concatenation intact.

More explicitly, your input is an array of arrays of nonnegative integers, which you may assume to fit in the native int type of your language. The input may be an empty array or an array of empty arrays, or it may contain arrays of different lengths. You can take the input in any reasonable format.

Your output shall be another array of arrays, again in any reasonable format. The concatenation of the output shall be equal to the concatenation of the input, so it contains the same integers in the same order. However, the sequence of lengths in the output shall be the reverse of that of the input.


Consider the input array

A = [[4,10],[0],[],[3,3,2],[1]]

The concatenation of A is

B = [4,10,0,3,3,2,1]

and its length sequence is

C = [2,1,0,3,1]

The correct output is


since it's the unique array with concatenation B and length sequence reverse(C).

Rules and scoring

You con write a full program or a function. The lowest byte count wins, and standard loopholes are disallowed.

Test cases

TODO: make more

[] -> []
[[]] -> [[]]
[[],[],[1]] -> [[1],[],[]]
[[1,2],[4,5,6]] -> [[1,2,4],[5,6]]
[[4,10],[0],[],[3,3,2],[1]] -> [[4],[10,0,3],[],[3],[2,1]]

Sandbox comments

I'm debating whether I should guarantee that the input is non-empty, and/or only contains non-empty arrays. In some languages (like J), empty arrays make the challenge significantly harder, but on the other hand, I don't want it to be too easy either.

  • \$\begingroup\$ Just an FYI "reshape" operations pretty heavily trivialise this. I don't think it would be crazy to ban them, but it's probably fine with them also. \$\endgroup\$ Jun 7, 2016 at 17:12
  • \$\begingroup\$ I'd prefer the more general challenge of splitting an array the same way as another array of arrays with equally many elements. \$\endgroup\$
    – xnor
    Jun 8, 2016 at 0:28
  • \$\begingroup\$ @FryAmTheEggman I'd prefer not to explicitly ban any builtins. If that leads to 3-byte Jelly answers, so be it. :P I also suspect Jelly would win anyway... \$\endgroup\$
    – Zgarb
    Jun 8, 2016 at 14:35
  • \$\begingroup\$ @xnor That would be essentially be this challenge, but with only arrays-of-arrays. Do you think they would be different enough not to be duplicates? \$\endgroup\$
    – Zgarb
    Jun 8, 2016 at 14:39
  • \$\begingroup\$ Even if the depth is fixed at 2, the shortest approach would most likely still be the be the same in many languages. It's a borderline dupe, and since it takes only one gold badge user to close a duplicate, it will probably end up closed. \$\endgroup\$
    – Dennis
    Jun 8, 2016 at 17:23
  • \$\begingroup\$ @Zgarb I'm not sure if I'd count it as a dupe, but I think reversed input vs general array wouldn't make much difference there. I expect most solutions would just work with the reversed input as if it were a general array. \$\endgroup\$
    – xnor
    Jun 8, 2016 at 20:50
  • \$\begingroup\$ @xnor Reversed input has the added complexity that the same array must be used for content and shape. That's not a big problem for some languages, but I'd expect another approach to outgolf the pop/map approach with the added overhead of creating a second reference for the reversed array in, e.g., Python. I admit I haven't tried it yet though. \$\endgroup\$
    – Dennis
    Jun 9, 2016 at 0:24
  • \$\begingroup\$ I think what you call "concatenation" of the array is called "flattening". \$\endgroup\$ Oct 13, 2016 at 18:24

The Secret Handshake

This is based off of this sandbox comment.

This is a challenge of secrecy.

The goal of this KOTH is to write a program that is capable of identifying itself amongst a crowd of other programs. In order to do this, you must develop a secret handshake which will be recognized only by other copies of the same program.

A single game involves every program competing at once. There will be five instances of each program in the arena, and the winner will be the first program to correctly identify the four other copies. After a large number of games, the submission with the most victories will be the overall winner.

The Gameflow:

  1. Each program receives a number which tells the number of bots in the arena, which is five times the number of submissions. The bots are arranged in a circle, and they are each considered ID #0 from their own perspective, with N-1 being the maximum ID number.
  2. Then the game cycle starts. At this point, each bot will be awaiting input.
    1. Your bot will receive input consisting of an ID number and an optional message.
      • For example, 7 hi means that bot #7 said hi to you.
      • If the ID number is 0, then there will not be a message. This would occur if it is your turn but there is no message to receive.
    2. Now, your bot is allowed to output a guess consisting of four ID numbers.
      • If those four numbers are the IDs of your teammates, then your team will win that game.
      • You will receive no confirmation of an incorrect guess.
    3. Next, you must output a message to send. The message will be a destination ID number followed by up to 3* characters. Example messages: 7 4 w 12 #?Q.

*This number is subject to change. Larger messages make it harder to fake a secret handshake. I hope that a very short message forces people to use multi-step handshakes. It might also be interesting to limit it to 1-character messages.

Sandbox Notes

Something that I haven't quite figured out is how the controller program will determine which bot gets to move each turn. I suppose it would be simplest go in order: Each bot has an "unread message queue" and receives/sends one message each turn. The bot immediately after you (ID #1) then moves next.

Alternatively, there could be a set turn order which is unrelated to the ID order, simply to make it impossible for one bot to tell which bot moves next.

  • \$\begingroup\$ 1. I don't think it needs a turn order: the only problem with adding simultaneous moves would be the need to account for ties. (You are going to run it more than once to pick a winner, right?) 2. I would keep the guess per turn without penalty for guessing wrong. A good strategy will allow a team to win before anyone wins by blind guessing. Too harsh a penalty for bad guesses will bias the game too much in favour of the last person to update their bot so that it fakes the responses of other teams. \$\endgroup\$ Nov 11, 2014 at 10:43
  • 3
    \$\begingroup\$ The idea is quite amusing, but something tells me there must exist a not too hard to find optimal strategy to maximize the chances of winning. The fight would likely occur around sending fake messages to the competitors. Second problem I see is, the first contestants will be at a huge disadvantage since the new players will see exactly how their code works. So much for secrecy. I think the challenge would be more interesting if the code was kept secret, but this is hardly compatible with the spirit of this site. \$\endgroup\$
    – user16991
    Jan 24, 2015 at 6:48
  • \$\begingroup\$ @PhiNotPi are you still interested in this challenge? Would you be willing to allow me to implement this? \$\endgroup\$ Jul 29, 2016 at 20:23
  • \$\begingroup\$ @RohanJhunjhunwala Sure, you can implement this. I don't think I ever started on a controller for it, mainly because I don't think this challenge will be as fun in practice as in theory. \$\endgroup\$
    – PhiNotPi
    Jul 30, 2016 at 13:17

Solve the Nonogram!

It is time to embark on a perilous quest to defeat the British Intelligence. The aim of this challenge is to write the shortest code that will solve a Nonogram.

What is a Nonogram?

Nonogram Puzzle

The rules are simple. You have a grid of squares, which must be either filled in black or left blank. Beside each row of the grid are listed the lengths of the runs of black squares on that row. Above each column are listed the lengths of the runs of black squares in that column. Your aim is to find all black squares. In this puzzle type, the numbers are a form of discrete tomography that measures how many unbroken lines of filled-in squares there are in any given row or column. For example, a clue of "4 8 3" would mean there are sets of four, eight, and three filled squares, in that order, with at least one blank square between successive groups. [1][2]

So the solution to the above Nonogram would be:

Solved nonogram

Implementation Details

You can chose to represent the Nonogram however you would like and take it as an input in whatever way you deem fit for your language. Same goes for output. The aim of this challenge is to literally just get the job done; if you can solve the monogram with whatever output your program gives, that is valid. One caveat is you can't use an online solver :)

You are, of course, free to use any language you want and since this is code golf, the entries will be sorted in the order: accuracy -> length of code -> speed.

This problem is very algorithmically challenging in that there is no completely efficient solution to it and as such, you won't be penalized for not being able to solve larger ones, although your answer will be heavily rewarded if it is able to handle big cases (see bonus). As a benchmark, my solution works for up to roughly 50x50 within 5-10 mins.


I actually learnt about Nonograms from a cryptographic Christmas card released by the British Intelligence here. The first part was basically a massive 25x25 Nonogram. If your solution is able to solve this, you will get kudos :)

To make your life easier in terms of data entry, I have provided how I represented the data for this specific puzzle for your free use. The first 25 lines are the row clues, followed by a '-' separator line, followed by 25 lines of the col clues, followed by a '#' separator line, and then a representation of the grid with the square clues filled in.

7 3 1 1 7
1 1 2 2 1 1
1 3 1 3 1 1 3 1
1 3 1 1 6 1 3 1
1 3 1 5 2 1 3 1
1 1 2 1 1
7 1 1 1 1 1 7
3 3
1 2 3 1 1 3 1 1 2
1 1 3 2 1 1
4 1 4 2 1 2
1 1 1 1 1 4 1 3
2 1 1 1 2 5
3 2 2 6 3 1
1 9 1 1 2 1
2 1 2 2 3 1
3 1 1 1 1 5 1
1 2 2 5
7 1 2 1 1 1 3
1 1 2 1 2 2 1
1 3 1 4 5 1
1 3 1 3 10 2
1 3 1 1 6 6
1 1 2 1 1 2
7 2 1 2 5
7 2 1 1 7
1 1 2 2 1 1
1 3 1 3 1 3 1 3 1
1 3 1 1 5 1 3 1
1 3 1 1 4 1 3 1
1 1 1 2 1 1
7 1 1 1 1 1 7
1 1 3
2 1 2 1 8 2 1
2 2 1 2 1 1 1 2
1 7 3 2 1
1 2 3 1 1 1 1 1
4 1 1 2 6
3 3 1 1 1 3 1
1 2 5 2 2
2 2 1 1 1 1 1 2 1
1 3 3 2 1 8 1
6 2 1
7 1 4 1 1 3
1 1 1 1 4
1 3 1 3 7 1
1 3 1 1 1 2 1 1 4
1 3 1 4 3 3
1 1 2 2 2 6 1
7 1 3 2 1 1
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 1 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 1 1 0 0 1 0 0 0 1 1 0 0 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 1 0 0 0 0 1 1 0 0 0 0 1 0 0 0 0 1 1 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

There can be only 1!

Your task is, given a positive integer n, to generate an expression that equals to the number n.

The catch is: you're only allowed the number 1 in the output.

The operators at your disposal are:

  • +, -, * and /
  • sqrt (as s)
  • ceil and floor (as c and f respectively)
  • ! (factorial)
    • The factorial, in this case, only works for positive integers.

You are also allowed to stack 1's together, so something like 11 is acceptable in the output. However, they count as the same amount of 1's that's in the number (so 11 counts as 2 1's).

You must also include brackets in the output, so that the expression in the output, when executed through the order of operations, will result in the input.


  • Input = 24, one possible output = (1+1+1+1)!
  • Input = 11, one possible output = 11
  • Input = 5, one possible output = c(s((1+1+1+1)!))
    • The ceiling of the square root of 24 is 5.


  • You are guaranteed that the input is a positive integer from 1 to 2^31-1.
  • Your program must work for any positive integer up to 2^31-1, even if they are not tested.
  • Your program must finish processing all outputs for all numbers in the set in 1 hour.
  • The results for every run of the program must be exactly the same - also, no seeds.
  • You are not allowed to have imaginary numbers anywhere in the output (so no s(some negative number)).
  • You are also not allowed to have numbers larger than 2^31-1 anywhere in the output, even when they are sqrted or /ed (so no (((1+1+1)!)!)! or ((1+1+1+1)!)!).

Set of Numbers:

945536, 16878234, 32608778, 42017515, 48950830, 51483452, 52970263, 54278649, 63636656, 78817406, 89918907, 90757642, 95364861, 102706605, 113965374, 122448605, 126594161, 148064959, 150735075, 154382918, 172057472, 192280850, 194713795, 207721209, 220946392, 225230299, 227043979, 241011012, 248906099, 249796314, 250546528, 258452706, 276862988, 277140688, 280158490, 286074562, 308946627, 310972897, 322612091, 324445400, 336060042, 346729632, 349428326, 352769482, 363039453, 363851029, 392168304, 401975104, 407890409, 407971913, 425780757, 459441559, 465592122, 475898732, 482826596, 484263150, 506235403, 548951531, 554295842, 580536366, 587051904, 588265985, 588298051, 590968352, 601194306, 607771869, 618578932, 626776380, 667919873, 681786366, 689854904, 692055400, 697665495, 711608194, 734027104, 750869335, 757710567, 759967747, 777616154, 830071127, 833809927, 835873060, 836438554, 836945593, 863728236, 864158514, 871273503, 881615667, 891619600, 897181691, 918159061, 920521050, 924502226, 929983535, 943162304, 950210939, 950214176, 962610357, 974842859, 988572832

(These are 100 random numbers from 1 to 1 billion.)

Scoring System:

Your score is determined like so:

  • Your program will be tested against the random numbers in the set.
    • You must provide the output generated using the numbers random numbers in the set (either inside your answer or as a pastebin link).
  • Your then have two "scores": A primary score and a secondary score.
    • Your primary score is (no. of 1's in output)*(no. of operators in output). If your primary score is the lowest, you win.
    • Your secondary score is your byte-count, and is only used in the case of a tie-breaker - the person with the lowest byte-count wins.


  • Anything that I need to clear up?
  • Is this challenge a dupe?
  • \$\begingroup\$ Is implicit multiplication allowed? \$\endgroup\$ Sep 11, 2016 at 1:23
  • \$\begingroup\$ @LegionMammal978 No, it is not allowed. You must use *. \$\endgroup\$
    – clismique
    Sep 11, 2016 at 1:43
  • \$\begingroup\$ For the tiebreak, do parentheses contribute to the operator count? \$\endgroup\$ Sep 11, 2016 at 1:49
  • \$\begingroup\$ Does the code need to be deterministic (that is, do random algorithms need to set the PRNG seed to ensure the same results each time)? \$\endgroup\$ Sep 11, 2016 at 1:51
  • \$\begingroup\$ Surely at least one output must be hardcoded? \$\endgroup\$ Sep 11, 2016 at 4:45
  • \$\begingroup\$ Related, though I don't believe it's a dupe. \$\endgroup\$ Sep 11, 2016 at 8:47
  • \$\begingroup\$ code-challenge can't be used along with other scoring tags. \$\endgroup\$ Sep 11, 2016 at 11:18

Telegraphy Golf: Decode Baudot Code


In 1870 Émile Baudot invented Baudot Code, a fixed-length character encoding for telegraphy. He designed the code to be entered from a manual keyboard with just five keys; two operated with the left hand and three with the right:

Baudot 5-key keyboard

The right index, middle and ring fingers operate the I, II, and III keys, respectively, and the left index and middle fingers operate IV and . (Henceforth I'll use their Western Arabic numerals, i.e. 1, through 5.) Characters are entered as chords. To enter the letter "C," for example, the operator presses the 1, 3, and 4 keys simultaneously, whereupon a rotating brush arm reads each key in sequence and transmits a current or, for keys not depressed, no current. The result is, in modern terms, a 5-bit least-significant-bit-first binary encoding, in which our example, "C," is encoded as 10110.

5 bits??

You might be thinking that 5 bits, which can express at most 32 unique symbols, isn't enough for even all of the English letters and numerals, to say nothing of punctuation. Baudot had a trick up his sleeve, though: His character set is actually two distinct sets: Letters and Figures, and he defined two special codes to switch between them. Letter Shift, which switches to Letters mode, is activated by pressing the 5 key alone (00001), and Figure Shift is activated with the 4 key (00010).


Your challenge is to write a program or function that decodes Baudot Code transmissions.

A real transmission would begin with some initialization bits, plus a start and stop bit before and after each character, but we're going to skip those and only worry about the 5 unique bits for each character. Input and output formats are discussed below.

Baudot's Code

There are two different versions of Baudot Code: Continental and U.K. We're going use the U.K. version, which doesn't include characters like "É" from Baudot's native French. We're also going to leave out all of the symbols in the U.K. version that aren't among the printable ASCII characters. You will only have to decode the characters in the table below, all of which are printable ASCII characters except the final three control characters that are explained below the table.

The "Ltr" column shows the characters in Letter mode and "Fig" shows the Figure mode characters:

        Encoding             Encoding
Ltr Fig  12345       Ltr Fig  12345
--- --- --------     --- --- --------
 A   1   10000        P   +   11111
 B   8   00110        Q   /   10111
 C   9   10110        R   -   00111
 D   0   11110        S       00101
 E   2   01000        T       10101
 F       01110        U   4   10100
 G   7   01010        V   '   11101
 H       11010        W   ?   01101
 I       01100        X       01001
 J   6   10010        Y   3   00100
 K   (   10011        Z   :   11001
 L   =   11011        -   .   10001
 M   )   01011        ER  ER  00011
 N       01111        SP  FS  00010
 O   5   11100        LS  SP  00001
 /       11000

The last three rows in the right column are control characters:

  • ER is Erasure. Baudot's telegraphy machines would print an asterisk-like symbol for this character to tell the reader that the preceding character should be ignored, but we're going to be even nicer to the reader and actually omit (do not print) the preceding character. It acts the same in both Letter and Figure mode.

  • FS is Figure Shift. This switches the character set from Letters to Figures. If the decoder is already in Figure mode, FS is treated as a Space (ergo SP in the "Ltr" column). When the decoder is in Figure mode it stays in Figure mode until an LS character is received.

  • LS is Letter Shift. It switches the character set from Figures to Letters. If the decoder is already in Letter mode, LS is treated as a Space. When in Letter mode the decoder stays in Letter mode until an FS character is received.

The decoder always starts in Letter mode.

Here's an example with Figure Shift, Letter Shift, and Space:

01011 10000 00100 00001 00010 10000 11100 00001 10101 11010
  M     A     Y   LS/SP FS/SP   1     5   LS/SP   T     H

This yields the message MAY 15TH. As you can see, the first 00001 (Letter Shift/Space) character acts as a space, because the decoder is already in Letter mode. The next character, 00010 (Figure Shift/Space) switches the decoder to Figure mode to print 15. Then 00001 appears again, but this time it acts as Letter Shift to put the decoder back in Letter mode.

For your convenience, here are the characters in a format that's perhaps easier to digest in an editor, sorted by code:



Input will be a string, array, or list of bits in least-significant-bit-first order. Each character will be represented by a quintet of 5 bits. Bits may be in any reasonable format, e.g. a binary string, an array of 0s and 1s, a string of "0" and "1" characters, a single very large number, etc., as long as it maps directly to the bits of the transmission.

Every transmission will have at least one printable quintet and at most 255 quintets (printable or otherwise), i.e. 5–1,275 bits inclusive.

The input can contain only the bits of the transmission, with two allowed exceptions: Any number of leading or trailing 0 bits and/or, for string input, a single trailing newline may be added to the transmission. Leading or trailing bits or characters cannot be added before or after each quintet, i.e. you cannot pad each quintet to 8 bits or separate quintets with any additional bits, e.g. "01111\n11100".

Notes & edge cases
  1. The transmission will contain only the characters in the "Ltr" and "Fig" columns in the table above. You will never receive e.g. 01110 in Figure mode, because it is absent from the "Fig" column.

  2. It is assumed that the decoder will always be in Letter mode at the beginning of a transmission. However, the first character may be an FS character to switch to Figure mode immediately.

  3. When the decoder is in Letter mode, it may receive an LS character, and when it is in Figure mode it may receive an FS character. In either event a Space character must be printed (see Output).

  4. The ER character will never be the first character in a transmission, nor will it ever immediately follow an LS, FS, or another ER.

  5. An FS character may immediately follow an LS character and vice versa.

  6. Neither the LS nor FS character will be the last character in any transmission.

  7. The / and - characters may be received in either Letter mode (codes 11000 and 10001, respectively) or Figure mode (10111 and 00111).


Output may be in any reasonable format, the most reasonable being ASCII (or UTF-8, for which all of the represented characters are the same as ASCII). Please indicate in your answer if your output is in another encoding or format.

  • The space character (see 3. above) should be an ASCII space (0x20) or your encoding's equivalent, i.e. what you get when you press the space bar.


This is code golf. The shortest code in bytes wins.


  • Standard loopholes are forbidden.

  • Trailing spaces and/or a single trailing newline are allowed. Leading spaces or other characters (that are not part of the transmission) are disallowed.

  • You may not use any built-in or library functions that decode Baudot Code (or any of its descendants, e.g. Murray Code, ITA-1, etc.).

Test Cases

Input: 001101000010100111101110010101
Output: BAUDOT
Input: 11010010001001100011110111101111100
Output: HELLO
Input: 01011100000010000001000101000011100000011010111010
Output: MAY 15TH
Input: 00010001000001000001011101110011100101010010110101010001111100101
Output: 32 FOOTSTEPS
Input: 10110000110101011100111100001111011010000001101110
Output: GOLF
Input: 000100011000001111100000100010110111001100010110010000111111
Output: 8D =( :P
Input: 0000100001000010000100010001111011111011000011100010001
Output (4 leading spaces):     -/=/-
  • \$\begingroup\$ This is quite similar to the Morse decoding question. The main thing it adds is the three control characters. It might be worth adding a cross-reference in a comment after posting the question. \$\endgroup\$ Sep 19, 2016 at 21:56
  • 1
    \$\begingroup\$ Can the transmission start with erasure? \$\endgroup\$
    – Sp3000
    Sep 20, 2016 at 1:00
  • \$\begingroup\$ @Sp3000 No. Good catch. I had a note to that effect but apparently accidentally deleted it. \$\endgroup\$
    – Jordan
    Sep 20, 2016 at 1:07
  • \$\begingroup\$ Can an erasure follow a shift that is acting as a space? \$\endgroup\$
    – Phlarx
    Sep 20, 2016 at 14:16
  • \$\begingroup\$ @Phlarx Nope. ER will never follow LS or FS. \$\endgroup\$
    – Jordan
    Sep 20, 2016 at 14:17
  • \$\begingroup\$ Thank you for your feedback, all. I've posted the challenge: codegolf.stackexchange.com/questions/94056/… \$\endgroup\$
    – Jordan
    Sep 21, 2016 at 13:23
  • \$\begingroup\$ Don't forget to take of this answer in the Sandbox. " When you think your challenge is ready for the public, go ahead and post it, replace the post here with a link to the challenge and delete it. " \$\endgroup\$
    – Linus
    Sep 22, 2016 at 19:29

my first post on here, be gentle ;)

Find all anagrams within a text

Somehow I stumbled upon an implementation of a school assignment from about a year ago, and after having seen many amazing and mindblowing code-golf solutions on here, I thought it's time I bring my own challenge and see how much you guys can blow my mind again ;)

The assignment

Given a text, find all words that have at least one other word in the text as an anagram (case insensitive). Multiple occurrences of the same word are not counted.

The output shall be grouped by words that are an anagram of each other.


  • How you handle input/output is up to you. Function-parameters, file-io, standard in/out, whatever works the best for you.
  • You must be able handle any non-empty input as long as you don't run into language or memory limitations.
  • The output does not have any fixed formatting. That means you may put them each group at a line, or put them all at one line but use different delimiters, a 2d array, some other exotic data-structure your language of choice happens to have, as long as it makes reasonable sense, it is considered correct. (This means that for example if you are just writing a function, that function does not need to display the output, it could just provide it as a return-value.) Just keep in mind the requirement that the words that are an anagram of each other should be grouped together.
  • The order in which the output appears does not matter. That applies to the order of the groups as well as the order of the words within the group.
  • A group of only one word is invalid, since that fails the "have at least one other word in the text as an anagram" requirement. (just omit them from your output ^^)
  • Each word should only appear once in the output
  • All interpunction characters are stripped away from the word before checking for anagrams. That means that "it's" and "its" are the same word (and thus both are an anagram of "sit"). My sample program at the bottom uses http://en.cppreference.com/w/cpp/string/byte/ispunct as check if a character is an interpuntion character. If your language has such a method, you may use it. Otherwise take the characters from the default C locale as specified on there:


  • All other characters are part of the word and treated as is, so "a" and "á" are not the same.


Given the following input text (the actual text I was given as example by school :P, no idea where this text is coming from...)

Parts of the world have sunlight for close to 24 hours during summer. Dan went to the north pole to lead an expedition during summer. He had a strap on his head to identify himself as the leader. Dan had to deal with the sun never going down for 42 consecutive days and his leadership strap soon became a blindfold. He wondered what kind of traps lay ahead of him.

the following output would be correct:

  • 24, 42
  • deal, lead
  • and, dan
  • parts, strap, traps

Or this would also be correct:

24, 42 | deal, lead | and, dan | parts, strap, traps

This one would not

24, 42 , deal, lead , and, dan , parts, strap, traps

(since the groups are not obvious)

My own (non golfed) version to check

The is the exact program I submitted to school back then. You may use it to check your own results.

Added bonus: If it happens to be that this program has a bug (I haven't found them yet) your submission is allowed to have it as well, since it is used to check the result. (In that case you are of course not required to have said bugs)

#include <string>
#include <iostream>
#include <fstream>
#include <map>
#include <set>
#include <algorithm>
#include <cstring>

void stringRemoveInterpunction(std::string& string);
void stringToLower(std::string& string);
std::string stringToAnagramIdentifier(std::string word);

 * due to use of std::ispunct and std::tolower it may not work for text with non-ascii characters??!
int main(int argc, char* argv[]) {
    if (argc < 2) {
        std::cerr << "usage: " << argv[0] << " <filename>" << std::endl;
        return 1;

    std::ifstream fileStream(argv[1]);
    if (!fileStream) {
        std::cerr << "Could not open file " << argv[1] << std::endl;
        return 2;

    // map to store the anagrams, key is so called "anagram identifier", value is a list of the words.
    std::map<std::string, std::set<std::string>> anagrams;

    // read words separated by whitespace from the file
    for (std::string word; fileStream >> word;) {
        // remove interpunction & convert to lowercase, since casing should be ignored
        stringRemoveInterpunction(word); stringToLower(word);

        // add to anagrams-store

    // display all the anagrams
    for (auto anagram : anagrams) {
        // skip entries which contains only one item, no anagrams found
        if (anagram.second.size() <= 1) {

        // output a comma-separated list of the anagrams
        auto anagramIterator = anagram.second.begin();
        std::cout << *anagramIterator++;
        while (anagramIterator != anagram.second.end()) {
            std::cout << ", " << *anagramIterator++;
        std::cout << std::endl;

    return 0;

void stringRemoveInterpunction(std::string& string) {
    string.erase(std::remove_if(string.begin(), string.end(), std::ptr_fun<int, int>(&ispunct)), string.end());

void stringToLower(std::string& string) {
    std::transform(string.begin(), string.end(), string.begin(), std::ptr_fun<int, int>(&std::tolower));

std::string stringToAnagramIdentifier(std::string word) {
    // sort the characters
    std::sort(word.begin(), word.end());
    return word;

Sandbox Questions

  • Do i need to add other tags, or is just code-golf enough?
  • I'm not completely sure about the upper-limit of the input text. My idea was that the code should be able to handle any size input as long as its within the memory-limits of the language. Like you don't have to write "memory optimal code" or something, but also shouldn't asume it is smaller than X. I could also just pick an upper limit of "1 kilobyte" or something to avoid any uncertainty about the requirements I think this is fine now as it is.
  • Someone in the comments below asked how to handle special characters like $?() so I took a look at how my "check program" handled that and it strips them away before doing the anagram check. So I added a rule for that, but while writing that I felt it makes it needlessly complicated and I'm considering ditching that rule and altering my check-program to reflect that (but then I cant claim its the exact same anymore :( )
  • Any other parts that are not clear?
  • \$\begingroup\$ What happens if there are 3 that are anagrams, such as eat ate tea? Are they all printed in one line / group? \$\endgroup\$
    – FlipTack
    Dec 19, 2016 at 17:27
  • \$\begingroup\$ @Flp.Tkc yes, like "parts, strap, traps" in the example \$\endgroup\$ Dec 19, 2016 at 17:29
  • \$\begingroup\$ I suggest you remove the grouping rule, since it's not really the interesting part of the challenge. It just adds code and limits the solutions. It's apparent which of the words that are together anyway. I do suggest they have to be grouped though, but without the need for delimiters. \$\endgroup\$ Dec 20, 2016 at 23:24
  • \$\begingroup\$ I think the memory rule is fine. I'm quite sure people will write a script that in theory would work for any input length if it wasn't for language or memory limitations. I also suggest you guarantee at least one character in the input. Otherwise people would need to add code just to handle empty input and that's not the interesting part of the challenge. \$\endgroup\$ Dec 20, 2016 at 23:29
  • \$\begingroup\$ You should add rules regarding special characters. Are it's, sit and its anagrams? What about hyphens? Can there be any special characters such as $?() etc? How are they treated? \$\endgroup\$ Dec 20, 2016 at 23:31
  • \$\begingroup\$ @StewieGriffin The grouping thing is a natural result of my solution to the problem back then, and imo its a fun part of the challenge, so I'm not totally sure about removing it. About the memory limit: I'm totally fine with the guarantee of at least one character if you think empty input needs special handling (i didnt think it would, but I dont really care :P) And yes, you are correct about special characters. I'd say "it's" is two words: "it" and "s". All special characters are ignored, so essentially regarded as whitespace between words. I'll add that \$\endgroup\$ Dec 20, 2016 at 23:35
  • \$\begingroup\$ welp, turns out I handled special characters differently, I strip them away from the word, so "it's" becomes "its", guess that'll be the rule then since I want to keep as close as possible to the original program \$\endgroup\$ Dec 20, 2016 at 23:48
  • \$\begingroup\$ @StewieGriffin I added a rule about special character handling, but I'm not completely sure about it, thoughts? (see my added "sandbox question") \$\endgroup\$ Dec 21, 2016 at 0:05
  • \$\begingroup\$ parsing and anagrams are relevant tags, maybe strings too, but I'm not sure about that one. \$\endgroup\$ Dec 21, 2016 at 7:43
  • \$\begingroup\$ !"#$%&'()*+,-./:;<=>?@[]^_{|}~ are all the non-alphanumeric ASCII-characters, except \` . I suggest you include the missing two symbols in the list to ignore, and say: "All non-alphanumeric characters (except spaces and newlines) must be trimmed away. So, it's and its are the same word." (You might want to rephrase that since my English isn't perfect, but something along those lines. If this is the rule then R2D2 and dr.22 will be anagrams, I'm not sure if that the desired behavior..? \$\endgroup\$ Dec 21, 2016 at 10:17
  • \$\begingroup\$ PS! I don't mean to be difficult, it's just that in my experience you'll get these questions sooner or later. So it's a good thing to sort it out while it's still in the Sandbox :) \$\endgroup\$ Dec 21, 2016 at 10:20
  • \$\begingroup\$ @StewieGriffin those two characters should have been included, look at the list from en.cppreference.com/w/cpp/string/byte/ispunct where I got it from, somehow I messed up the copy-paste :( And about "what is desired behaviour?" Im honestly not sure, I never really thought about it before you mentioned it and then I looked at how my sample program handled it. I remember putting in the ispunct-trim for stuff like commas after a word, im not sure about characters within the word... \$\endgroup\$ Dec 21, 2016 at 10:26
  • \$\begingroup\$ You can say that the input will not have any special characters except ' and -. Those two must be trimmed away. I think you must include those two, since these can be found in many texts. \$\endgroup\$ Dec 21, 2016 at 10:34

PPCG Handwriting OCR

(insert logo here once I make it)

Given an image consisting of handwritten text, output the text that is written. The image of the handwritten text will be generated by taking characters from one or more handwriting samples given by PPCG users.


  • You may not hardcode your program to only recognize the samples in the corpus.
  • There will be sufficient spacing between characters to avoid ambiguity.
  • Only ASCII alphanumeric characters (those matching the regex [A-Za-z0-9], i.e. uppercase and lowercase English letters and digits) will be present in the input.
  • Inputs will be formed by concatenating individual characters from the handwriting samples.
  • The test cases used for scoring may be modified at any point if I feel it is necessary to do so. Reasons may include but are not limited to: needing more test cases to have a single winner, removing problematic test cases, and fixing errors in test cases.

Aside from the above considerations, there are no guarantees on the appearance of the handwriting, as these are actual handwriting samples and thus can have drastic variances.


Your score will be the number of test cases that are correctly recognized, divided by the total number of test cases. The highest score wins. In the event of a tie, the first submission to reach the high score wins. Additional test cases may be added to break ties.

Handwriting Samples

This Imgur album contains the handwriting samples, as well as the names of the users who contributed them.

I've made a chat room for submitting handwriting samples. The more samples I get, the better this challenge will be, so please take a few minutes and submit a sample!

  • \$\begingroup\$ You may not hardcode your program to only recognize the samples in the corpus.: Can we tailor our code to be better for the samples than other inputs though? \$\endgroup\$ Aug 2, 2017 at 16:08
  • \$\begingroup\$ The test cases used for scoring may be modified at any point: Will current answers be modified? \$\endgroup\$ Aug 2, 2017 at 16:09
  • 1
    \$\begingroup\$ @TheLethalCoder 1) No. 2) Answers will be run on the test cases any time the answer or the test cases change, and the scores will be updated accordingly. \$\endgroup\$
    – user45941
    Aug 2, 2017 at 16:11
  • \$\begingroup\$ Hadn't seen that loophole before :) \$\endgroup\$ Aug 2, 2017 at 16:13
  • \$\begingroup\$ Can I nitpick again and say handwriting OCR is called ICR? :P \$\endgroup\$ Aug 3, 2017 at 9:06
  • \$\begingroup\$ Can you add an example input as well? \$\endgroup\$ Aug 3, 2017 at 9:07
  • 2
    \$\begingroup\$ @TheLethalCoder 1) ICR is when a program tries to learn what handwriting looks like via machine learning. OCR is just parsing written/text input into data. This is OCR, not ICR. 2) I'll add some examples once I get more samples and finish writing my generator. \$\endgroup\$
    – user45941
    Aug 3, 2017 at 9:08
  • \$\begingroup\$ Can you assume a minimum height for the characters? \$\endgroup\$ Aug 3, 2017 at 9:28
  • \$\begingroup\$ @TheLethalCoder Aside from the above considerations, there are no guarantees on the appearance of the handwriting, as these are actual handwriting samples and thus can have drastic variances. \$\endgroup\$
    – user45941
    Aug 3, 2017 at 9:29
  • \$\begingroup\$ To be honest with no restrictions this is going to be very hard, Do you lose points for returning extra information? Assume the input Hello if I return H.e.l.l.o. is that 100% for that test case or do I lose and get something like 50%? \$\endgroup\$ Aug 3, 2017 at 9:31
  • \$\begingroup\$ And does the output have to be in the correct order? \$\endgroup\$ Aug 3, 2017 at 9:58
  • \$\begingroup\$ @TheLethalCoder It's all or nothing. Getting 100% on a test battery challenge should be hard. \$\endgroup\$
    – user45941
    Aug 3, 2017 at 9:58
  • \$\begingroup\$ Please update imgur album \$\endgroup\$
    – Pavel
    Sep 18, 2017 at 4:28
  • \$\begingroup\$ @Pavel I will once I finish cutting up the images. My free time has been limited these past few weeks. \$\endgroup\$
    – user45941
    Sep 18, 2017 at 4:29

2 Spooky 4 Me

In terms of halloween, some things are just too spooky for me... Feel like we need some serious doots from skeletons to fuel our hallowed weens. So, in the spirit of that end, print the following, exactly as it is shown, if and only if the input does not equal "DOOT" (in all caps ONLY):

             .'       '.
             :)       (:
             \ (@) (@) /
              \   A   /
               )     (
        / ':-(_.-: :-._)-:' \
       / /' (__.-: :-.__) '\ \
      / /  (___.-' '-.___)  \ \
     / /   (___.-'^'-.___)   \ \
    / /    (___.-'='-.___)    \ \
   / /     (____.'='.____)     \ \
  / /       (___.'='.___)       \ \
 (_.:       '---'.=.'---'       :._)
 :||        __  _.=._  __        ||:
 :||       (  '.-.=.-.'  )       ||:
 :||       \    '.=.'    /       ||:
 :||        \    .=.    /        ||:
 :||       .-'.'-._.-'.'-.       ||:
.:::\      ( ,): O O :(, )      /:::.
|||| '     / /''--'--''\ \     ' ||||
''''      / /           \ \      ''''
         / /             \ \
        / /               \ \
       / /                 \ \
      / /                   \ \
     / /                     \ \
    /.'                       '.\
   (_)'                       '(_)
    \\.                       .//
     \\.                     .//
      \\.                   .//
       \\.                 .//
        \\.               .//
         \\.             .//
          \\.           .//
          ///)         (\\\
        ,///'           '\\\,
       ///'               '\\\
      ""'                   '""

However, if the input DOES equal "DOOT", in all caps only, print this instead:

             .'       '.
             :)       (:
             \ (@) (@) /
              \   A   /
               )     (
                '._.'          ' ''''    _''|
                 .=.       @=====***===::_  |
         .---._.-.=.-._.---. (( \-@|_) )) `.|
        / ':-(_.-: :-._)-:' \ ]--------'"
       / /' (__.-: :-.__) '\ \   ||:
      / /  (___.-' '-.___)  \ \  ||:
     / /   (___.-'^'-.___)   \ \ ||:
    / /    (___.-'='-.___)    \ \||:
   / /     (____.'='.____)     \ ||:
  / /       (___.'='.___)       \||:
 (_.:       '---'.=.'---'       :._)
 :||        __  _.=._  __       
 :||       (  '.-.=.-.'  )      
 :||       \    '.=.'    /       
 :||        \    .=.    /      
 :||       .-'.'-._.-'.'-.       
.:::\      ( ,): O O :(, )   
|||| '     / /''--'--''\ \    
''''      / /           \ \
         / /             \ \
        / /               \ \
       / /                 \ \
      / /                   \ \
     / /                     \ \
    /.'                       '.\
   (_)'                       '(_)
    \\.                       .//
     \\.                     .//
      \\.                   .//
       \\.                 .//
        \\.               .//
         \\.             .//
          \\.           .//
          ///)         (\\\
        ,///'           '\\\,
       ///'               '\\\
      ""'                   '""


  • Trailing newlines and spaces are allowed.
  • The design is horizontally symmetric, if you find inconsistencies let me know.

Doot it up, and enjoy!

(Yes, I'm going to make it more official when posting on the actual SE)

  • 4
    \$\begingroup\$ What about this challenge ensures that the same old techniques won't be the best ones (i.e. that it adds value to the site)? \$\endgroup\$ Oct 13, 2017 at 23:01
  • 2
    \$\begingroup\$ @PeterTaylor worse? Better? 10x worse? 10x better? I don't really know what makes the challenge unique beyond a formal proof that it is, but if the users enjoy it; why not allow it... current event challenges attract new users. \$\endgroup\$ Oct 16, 2017 at 23:03

Check equation proofs in a ring

The recent to prove that (-a)×(-a) = a×a attracted a number of faulty submissions, because there wasn't an easy way to verify the proofs. So, let's write some proof checkers.


Your program should take a sequence of strings representing expressions in a ring. Valid expressions consist of:

  • Single-lowercase-letter variables (a to z)
  • Two constants: the additive identity 0 and multiplicative identity 1
  • Compound expressions: (X+Y), (X*Y) and (-X), where X and Y stand for subexpressions. (The parentheses must always be present, and there must be no whitespace.)


Your program should check whether:

  • All strings except the first and last represent valid expressions.
  • Each expression (after the first) can be obtained from the preceding expression, by substituting one of the ring axioms in the expression exactly once.

Output truthy if these conditions are met. Otherwise, output falsey.

You may assume that the first and last strings in the input are valid expressions. But your program must check the intermediate strings.

The ring axioms

For this challenge, use the following substitution rules (and do not use any others). Substitutions can go both left-to-right and right-to-left.

  1. (X+(Y+Z)) = ((X+Y)+Z)

  2. (X+0) = X

  3. (X+(-X)) = 0

  4. (X+Y) = (Y+X)

  5. (X*(Y*Z)) = ((X*Y)*Z)

  6. (X*1) = X

  7. (1*X) = X

  8. (X*(Y+Z)) = ((X*Y)+(X*Z))

  9. ((X+Y)*Z) = ((X*Z)+(Y*Z))


Proof checkers are traditionally small, so that people can review them easily. Therefore, your program should be written in as few bytes as possible.

Meta comments

Is the input format fair for most languages and approaches?

Usually, code-golf problems should not require input validation. However, I thought it would be appropriate behaviour for a proof checker. I think the current spec still accommodates regex-based solutions.

The format for the original challenge also listed which axiom was used for each step. I could include this but I doubt that it improves the challenge much.

Test cases

Valid proofs



Invalid proofs

These proofs are missing intermediate steps.




This is simply untrue, so no proof should ever be accepted.


Invalid expressions

Your proof checker should reject if these appear partway through a proof.



(a + b)


  • 1
    \$\begingroup\$ You should include the ring axioms to make the post self-contained. \$\endgroup\$
    – Laikoni
    Oct 18, 2017 at 16:23

ASCII Stock Exchange

At the ASCII Stock Exchange, each character has a price. If a character is used more often, its price rises, otherwise the price decreases over time.

Initially, each character has price 10. [Meta: Is this too low/high?] After each answer, the prices change as follows:

  • Each character that is not used in the answer's code has its price decreased by one, except when the price is already one, in which case it stays one.
  • If a character is used n-times, then its price increases by n.

We define the score of a piece of code as the sum of the prices of its characters.


For the sake of simplicity, we only consider characters A, B and C for this example. The challenge itself works with all bytes from \0 to \255. Initially, we have the following prices:

A -> 10, B -> 10, C -> 10

If the code of the first answer is BAAA, then is has a score of 40 (computed by taking the previous character prices into account) and the prices change to

A -> 13, B -> 11, C -> 9

If the next answer is CC, it has a score of 18 and the prices are updated to

A -> 12, B -> 10, C -> 11

The Task

Your objective is to write a program or function which calculates the score of a given piece of code in dependence of a list of previous answers which all influence the initial prices in the way described above.

The goal is to do so while minimizing the submission's score itself in the context of this challenge, that is your submission should be able to calculate its own score by taking a list of all previous submissions and its own source code as input.

The answer with the lowest score in each language wins.


  • You may take a list of strings and a string as input, or require that the string to be scored is the first/last element of the list, or any other reasonable input format.
  • You may not answer twice in a row.
  • If an answer in language X has already been posted, you may only post another answer in language X if your submission achieves a lower score than the previous answer and the code is not identical.
  • For this challenge only major releases of languages are considered their own language (e.g. Java 7 vs. Java 8). If there already is an answer in version A of a language and you have an answer in version B of the language and are in doubt whether version B is different enough from version A to be treated as different language, make sure that your code is not valid in version A.

Answer Format

To avoid having to copy all previous answers in order to calculate your submission's score, the chain will maintain a score calculator on TIO. Click on the link to the calculator in the previous answer and enter your code into the input field to calculate its score. Then add your code as a command-line argument, generate a new link and include it in your answer for the next submission.

If you wrote answer number 42 in Haskell with a score of 100, please format it as

42. Haskell, score 100


TIO-Link, explanation, ...

Score calculator for next answer

Test Cases

These test cases are in the format list of strings, string to score -> result.

[], "BAAA" -> 40
["BAAA"], "CC" -> 18
["abc"], "abc" -> 33
["ab12", "aa22", "31a"], "ac23" -> 42
["a","b","c","d","e","f","g","h","i","j"], "123" -> 3


  • Any idea what could be a good initial price? I picked the number 10 rather arbitrarily.
  • Letting the cost of unused characters decrease until 0 might lead to score 0 answers. Do you think this is a problem and the minimum cost should be 1? Minimum changed to 1.
  • I'm unsure what range of characters is a sensible choice. Limiting answers to printable ASCII plus white space would make things easier but also exclude a lot of languages. Another possibility would be to allow the whole byte range from \0 to \255. Then also golfing languages could participate, albeit to score them they would need to be converted to their byte form which usually contains a lot of unprintable characters. The score calculator is able to handle unprintables, but I don't know how to insert them into the text fields on TIO. All bytes from \0 to '\255' are allowed.
  • Is the winning criterion suitable for answer chaining?
  • A leader board snippet would be nice, but I don't know how to modify the existing ones. If someone could provide such a snippet, I would be very grateful.
  • Relevant tags?
  • 1
    \$\begingroup\$ I think decreasing the scores to 1 is more sensible, because otherwise I could write a submission using all (or even just one of) the zero-score characters a billion times (say, in a comment), keeping a "minimum" score, but then subsequent submissions would have completely absurd scores, and it'd be easy to outgolf later, leading to a boring answer chain. \$\endgroup\$
    – Giuseppe
    Oct 2, 2017 at 18:08
  • \$\begingroup\$ You may want to correct "enter your code into the input field to calculate its score. Then add your code as a command-line argument," \$\endgroup\$ Oct 2, 2017 at 19:55
  • 2
    \$\begingroup\$ I think allowing all 256 bytes is a good idea because languages like Jelly, 05AB1E, etc. will more than likely use more than the printable ASCII chars. \$\endgroup\$ Oct 2, 2017 at 20:30
  • 3
    \$\begingroup\$ "While in principal Python 2.7.14 and Python 2.7.13 are different languages according to the site rules, I recommend to avoid using the fact that it is technically allowed as an excuse to post boring answers." Recommendations do not work, especially when they're as imprecise as this. (Are you "recommending" not treating Python 2 and Python 3 as separate languages? I have no idea). If you want to ban boring exploitation of the convention on different interpreters, ban it outright, but give a clear definition of how lines should be drawn. \$\endgroup\$ Oct 2, 2017 at 21:21

Tower Builder

The king is coming to visit! Your city is competing with another neighboring city to attract his attention.

Each player owns their own tower within a city full of other players. Each turn you place 1 blocks on your tower or another tower in your city.

After 100 turns, the king will only visit the city with the highest tower. If he visits your city, then you will gain points equal to the number of blocks in your personal tower. In the case of a tie, neither city is visited.

Games include all players, and each game will randomly arrange players into cities. Your score is the total score across all games.

  • Player identifiers are randomly generated at the start of the tournament, but are consistent from game to game.
  • In addition, contrary to past KoTHs: I allow saving state from game to game (but not between tournaments)
  • You will always have complete information, including:
    • The size of everybody's towers (including the other city)
    • The actions players have taken
    • The current score of all players
  • \$\begingroup\$ How is always playing on the tallest tower in your city a bad strategy? Maybe the other city will do the same, but you won't lose. \$\endgroup\$
    – aschepler
    Jul 22, 2017 at 22:47
  • \$\begingroup\$ @aschepler because the number of points you get is height of your personal tower. If you only build on another's tower, and never your own, you'll never get any points. \$\endgroup\$ Jul 23, 2017 at 1:55
  • \$\begingroup\$ @NathanMerrill it took me a few minutes to understand what you even meant by that. you probably need to make this more clear \$\endgroup\$ Aug 7, 2017 at 0:24
  • \$\begingroup\$ If the points are only awarded to the player who is visited by the king, it would seem that putting 99 blocks on the city tower and 1 block on your personal tower guarantees victory, no? \$\endgroup\$ Jan 30, 2018 at 17:22
  • \$\begingroup\$ @AdmBorkBork there's no singular "city tower". Furthermore, you are still competing against the players in your city. Your score is your total score across many games. \$\endgroup\$ Jan 30, 2018 at 17:40
  • \$\begingroup\$ Oh, the part that I missed was that multiple players share a city. I think that should be made more clear because that will drastically change strategy. I read it as you have a personal tower and everyone also gets their own city with a separate tower, and it's only the city's tower that matters for the king. \$\endgroup\$ Jan 30, 2018 at 17:45
  • \$\begingroup\$ This needs some clarification about arrangement. The game will randomly arrange players into one of () cities, or a city with () others in it \$\endgroup\$
    – pfg
    Jan 30, 2018 at 19:10
  • \$\begingroup\$ @pfg It'll definitely be one of () cities, but I'm not sure how many. My intuition says "2", but I can't give a solid reason. Any suggestions? \$\endgroup\$ Jan 30, 2018 at 19:21
  • \$\begingroup\$ 2 or 3 per city or total cities? 3 seems good per city because with just two the strategy is much easier @Nathan Merrill \$\endgroup\$
    – pfg
    Jan 30, 2018 at 19:23
  • \$\begingroup\$ Hmmm...I'm thinking 100 players in 2 cities, and allow for duplicate entries. \$\endgroup\$ Jan 30, 2018 at 19:28
  • \$\begingroup\$ It seems to me that if you allow saving state you can simplify the last point to "You will be notified of the actions of every player" and let those who want to analyse it calculate the sizes of the towers, scores, etc. Normally I would favour supplying calculated information to save the players from re-implementing the game logic, but here the logic is simpler than the serialisation would be. \$\endgroup\$ Jan 31, 2018 at 8:43
  • \$\begingroup\$ @PeterTaylor I already provide complete information about the current game (including its history). The difference here is that you are allowed to persist information from game to game, meaning you can remember players who are antagonistic. \$\endgroup\$ Jan 31, 2018 at 15:04
  • \$\begingroup\$ My point is that the communication between the server and the bots would be simpler if you didn't provide that information, and you don't need to provide it because it can be calculated easily. \$\endgroup\$ Jan 31, 2018 at 15:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .