Sometimes when writing brainfuck code, you feel the need to make it longer than needed to encourage debugging. You could do it by just plopping a >< in there, but what fun is that? You'll need something longer and less NOPey to confuse anybody reading your code.

Quick introduction to Brainfuck

Brainfuck is an esoteric programming language created in 1993 by Urban Müller, and notable for its extreme minimalism. (Wikipedia)

Brainfuck is a language based on eight commands: +-><,.[]. The code is run on something like a Turing machine: an infinite tape on which values can be changed. In this challenge, we'll focus on the first four:

+    increment the value at the pointer
-    decrement the value at the pointer
>    move the pointer right
<    move the pointer left

Brainfuck NOPs

A brainfuck NOP is a sequence of brainfuck characters that, when executed from any state, leads to no change in the state. They consist of the four characters mentioned above.

The Challenge

The challenge is to write a program or function that, when executed, generates a random brainfuck NOP of the given length.


You will receive as input a nonnegative even integer n. (NOPs are impossible for odd n.)


You will output a random brainfuck NOP of the length n.


  • The definition of NOP: when the output of the program is inserted at any point in a brainfuck program, the behavior of said program must not change in any way. In other words, it must not change the state of the interpreter.
    • Note that for example +>-< is incorrect, since it changes the values of the two cells without changing them back. Please test your solution for these before posting.
    • Also note that +>-<->+< is a NOP that can't be reduced to nothing just by removing >< <> +- -+. Thus, you can't use an algorithm that just inserts these inside each other.
  • Every valid NOP of the length n must have a nonzero chance of appearing in the output. The distribution does not have to be uniform, though.
  • The brainfuck interpreter in question has a doubly infinite tape of arbitrary precision cells. That is, you can go infinitely to the both directions, and increment/decrement each cell indefinitely.
  • The program must finish within 1 minute for n = 100 on my machine, so no generating all possible NOPs and picking one.
  • If given invalid input (non-integer, negative, odd, etc.) you may do anything you'd like, including crash.


This is , so the shortest answer in bytes wins.


Here are all valid outputs for n = 4:

++--    +-+-    +--+    --++    -+-+    -++-
>><<    ><><    ><<>    <<>>    <><>    <>><
><+-    ><-+    <>+-    <>-+
>+-<    >-+<    <+->    <-+>
+><-    -><+    +<>-    -<>+
+-><    -+><    +-<>    -+<>

Here are a few possible outputs for n = 20:

  • 16
    here's a brainfuck NOP that doesn't use +-<> like you asked for: a – undergroundmonorail Dec 7 '15 at 13:46
  • 1
    I don't think non-simple NOPs exist, so you can probably remove that qualification. . has a side-effect, , overwrites a value which cannot be recovered without the use of []. But [] will end up setting a value to zero. This also overwrites a value (so we'd need another [] to recover it) unless we can be sure that the affected cell was zero to begin with. However, we'd have to search for such a cell with something like [>], and it's impossible to reliably return to the position we came from. – Martin Ender Dec 7 '15 at 14:52
  • 4
    @Eumel "The brainfuck interpreter in question has a doubly infinite tape of arbitrary precision cells." – Martin Ender Dec 7 '15 at 15:50
  • 2
    Please note that "Brainfuck" is no longer allowed in question titles on the system level. It appears you were able to circumvent the restriction by using non-ASCII characters. In the future, please abide by this restriction. – Alex A. Dec 7 '15 at 19:10
  • 1
    I'd love to see someone post an answer to this in brainfuck – Patrick Roberts Dec 8 '15 at 4:59
up vote 13 down vote accepted

CJam, 62 59 bytes

Thanks to nhahtdh for saving 3 bytes.

Because there is no requirement for any particular distribution as long as each no-op appears with finite probability, we can simplify this a lot by simply generating string containing a balanced number of -+ and <>, respectively, testing if it's a NOP and sorting it if it isn't.

Of course, for longer inputs, this will almost always result in sorted output, but you can test the code with some input like 8 to see that it can in principle produce any NOP of the given length.


Try it online.

  • 1
    Yes... the arbitrary limit should have been n=1000 under 10 seconds. Computers are just way to fast today ^^ Because the algorithmic answer solves it in under a second even for n = 1000 – Falco Dec 8 '15 at 11:50
  • For even larger n, I think it's possible to just sort the output if the balanced string is not NOP. The distribution is terribly skewed, but it's allowed by the question. – n̴̖̋h̷͉̃a̷̭̿h̸̡̅ẗ̵̨́d̷̰̀ĥ̷̳ Dec 9 '15 at 9:02
  • @n̴̖̋h̷͉̃a̷̭̿h̸̡̅ẗ̵̨́d̷̰̀ĥ̷̳ that's a neat idea. – Martin Ender Dec 9 '15 at 9:08
  • @n̴̖̋h̷͉̃a̷̭̿h̸̡̅ẗ̵̨́d̷̰̀ĥ̷̳ Thanks, that actually saves three bytes here as well. – Martin Ender Dec 9 '15 at 14:49

CJam, 118 116 bytes

This got slightly out of hand... especially the second half seems like it should be very golfable.


Test it here.

This handles N = 100 pretty much instantly. I don't have time to write the full breakdown of the code now, so here is the algorithm:

  • Generate a random balanced string of < and > with random (even) length between 0 and N inclusive.
  • Riffle the tape head positions into this array. E.g. "<>><" becomes [0 '< -1 '> 0 '> 1 '< 0].
  • Get a list of all positions reached in the process.
  • For each such position initialise an empty string. Also determine how many pairs of characters are left to reach a string of length N.
  • For each remaining pair append +- to the string of a random position.
  • Shuffle all of those strings.
  • For each position determine how often that position occurs in the riffled array, and split the corresponding string into that many (random-length) chunks.
  • In the riffled array, replace the occurrences of the position with its random chunks.

Done. This is based on the observation that:

  • Any NOP must have an equal amount of < and > to return the tape head to the original position.
  • The code will be a NOP as long as each tape cell is incremented as often as decremented.

By distributing random but balanced amounts of +s and -s between all the places where the tape head is on a given cell, we ensure that we find every possible NOP.

Mathematica, 350 bytes


Way too long? Yes. Do I even care? Not until someone else posts a valid answer.

  • 4
    Would you mind adding an explanation, so people can actually convince themselves this is valid? :) – Martin Ender Dec 7 '15 at 15:53
  • How exactly does this work? If I call the function with a number it only returns +. – Martin Ender Dec 7 '15 at 19:57
  • @MartinBüttner Fixed... Currently, it just generates random programs with an equal number of +-- and <-> pairs until one happens to be a NOP. Half of it is taken by a simple BF interpreter. – LegionMammal978 Dec 7 '15 at 22:00
  • does that actually generate a valid no-op of length 100 in under a minute? – Martin Ender Dec 7 '15 at 22:07
  • @MartinBüttner Yes. On average, I would say that it takes about 5 seconds. At first, I tried completely random programs, but it never terminated for length 100. – LegionMammal978 Dec 7 '15 at 22:14

Python 3, 177 bytes

from random import*
for c in s:p+=~c%2*(c-61);r[p]+=c%2*(44-c)
if any(r+[p]):s=a*(n//2)

Try it online!

I used code from Bubbler's answer for the BF simulation.

Python 3, 163 bytes

from random import*
for c in a:d[p]+=c%2*(44-c);p+=~c%2*(c-61)
if p|any(d):a=n//2*b'+-'

Try it online!

Full program that prints results to STDOUT. The line that runs BF code might be golfable.

Adopted Tyilo's approach; if the generated BF code is not a NOP, discard it altogether and fall back to '+-' repeated.

  • Timeout for n=100 – l4m2 Oct 24 at 12:51
  • @l4m2 Didn't notice that requirement. Fixed. – Bubbler Oct 24 at 23:44

JavaScript (Node.js), 160 bytes


Try it online!

Wolfram Language (Mathematica), 224 bytes


Try it online!

Here is the un-golfed (or rather, pre-golfed) version:

 k = RandomInteger[n/2];
 s = RandomSample[## & @@@ Table["<" ">", k]];
 While[Length[s] < n,
   s = Insert[s, "+", i = RandomInteger[Length[s]] + 1];
   p = GatherBy[Range[0, Length[s]], 
     Count[#, "<"] - Count[#, ">"]& @ Take[s, #]&];
   j = RandomChoice @@ Select[p, ! FreeQ[#, i] &]];
   s = Insert[s, "-", j + 1];

We first pick a random number of <'s and >'s to use, and generate a random list with an equal number of each.

To fill in the rest of the characters, we pick a position in which to add a +, then find a position where the pointer points to the same location and add a - there.

Repeat until the list has length n, and stringify the result.

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.