13

This puzzle is inspired by a puzzle by @Andrew, where the goal is to make 1998 from a minimal amount of 8s. This is a variation of the puzzle, where the goal is to make 2019 using a minimal amount of single digits (0s, 1s, 2s, etc.)

More formal description (to remove any disambiguations and loopholes):
You need to make a (so-called) valid $i$-expression ($i$ is a digit from 0 to 9) which evaluates to 2019 under the common sense of the characters being included into it, and has the minimum number of digits (not counting all other characters) in it (the count is separate for each $i$), according to the following rules (which determine what an $i$-expression is):

  • A string of one or more identical digits $ii\dots i$ is a valid $i$-expression (i.e. $2$, $22$ and $22222$ are valid $2$-expressions).
  • If $A$ and $B$ are valid $i$-expressions, so are $A+B$, $A-B$, $A\times B$ , $\frac A B$, $A^B$ and $\sqrt[A]B$.
  • If A is valid $i$-expression, so are $(A)$, $(A)!$ (factorial - note the parentheses, to explicitly disallow the multifactorials) and $\sqrt A$ (here, 2 is implied).

Example (a "5-expression"): $$(5\times(5+5)-5)^{\frac{5+5}{5}}-5-\frac55=(5\times10-5)^\frac{10}5-5-1=45^2-6=2025-6=2019\ (10\ \mathrm{fives})$$

Clarification: The ultimate aim is to find a minimum number for each kind of digit (i.e. a solution with minumum number of zeroes, ones, etc.), with a total of 10 different solutions. Of course, this does not invalidate the existing answers - providing a single solution is already great.

trolley813
  • 11,282
  • 2
  • 25
  • 57
  • 4
    Should make this an annual puzzle! – Adam Sep 12 '19 at 09:42
  • @Adam Yes, it's a great idea. – trolley813 Sep 12 '19 at 10:38
  • According to this consensus on meta, these types of questions must have a provably optimized answer. It may be the case that the only valid answers are brute-force computer searches. (It may also be too broad, because it asks ten distinct questions.) To answerers: Answers must have a proof of optimality. An answer without a proof of optimality is a comment, not an answer. – Deusovi Sep 13 '19 at 03:54
  • If you'd like to discuss this puzzle without knowing whether your solution is optimal, you can do that in the comments. You can also make a chat room for this particular puzzle, and propose solutions there. – Deusovi Sep 13 '19 at 04:06
  • @Deusovi is the issue with unproven answers not in the question? If a provable answer is required, I argue that OP should have a proven answer ready, or at least know it is provable. Proving the questions asked in this post is non-trivial and I think fits better in the Mathematics stack exchange. I acknowledge the meta post and I don't want to open that discussion again, but how I read it, the question should not be open ended (rather provable). That forces the answers to come with a proof. – P1storius Sep 13 '19 at 06:59
  • @P1storius Yes, a question is likely not good (or at least, not a good fit for this site) if it does not allow for those types of answers easily - just like a cipher question would be a bad question for this site if it could only be solved through brute-forcing keys. – Deusovi Sep 13 '19 at 07:41
  • 10 sevens: $7\left(\dfrac{7!}{7+7}-77+7\right)-\dfrac{77}7$ – ə̷̶̸͇̘̜́̍͗̂̄︣͟ Sep 13 '19 at 08:08
  • I've temporarily locked this question due to concerns about answer validity and how the question falls under this meta post - see this discussion in The Sphinx's Lair for more details. – Deusovi Sep 13 '19 at 08:32
  • @TheSimpliFire fun fact: 2019 is nearly $7^4-7^3-7^2+7$ – Mr Pie Sep 30 '19 at 10:41

16 Answers16

20

This is an agglomeration of all existing answers.

If you have a more optimal solution than presented above, please replace the answer with your one, change the number of digits used, and remember to change the link to your profile's.

If your solution alters the status of concatenating, please change that too (i.e. with -> without, or without -> with).

Anyone can edit as this is now a community wiki.


Zero

$(0! + 0! + 0!)!^{0! + 0! + 0! + 0!} + ((0! + 0! + 0!)!)! + 0! + 0! + 0!$

One

$(1+1)^{11}-(1+1+1)(11-1)+1$

Two

$\sqrt{2^{22}}-2^2!-2-2-\dfrac22$

Three

$3!\sqrt{3!^{3!}}+(3!)!+3$

Four

$\dfrac{\sqrt{\sqrt{4^{4!}}}}{\sqrt{4}} - 4! - 4 - \dfrac{4}{4}$

Five

  • Using eight 5s with concatenation (@steffen)

$5 - \frac{5555}{5} + 5^5$

Six

$6\sqrt{6^6}+6!+\dfrac{6+6+6}6$

Seven

$7\left(\dfrac{7!}{7+7}-77+7\right)-\dfrac{77}7$

Eight

$\dfrac{8!}{(\sqrt{8+8})!-\sqrt{8+8}}+\dfrac{(\sqrt{8+8})!}{8}$

Nine

$(\sqrt9\,!)\sqrt{(\sqrt9\,!)^{\sqrt9\,!}}+(\sqrt9\,!)!+\sqrt9$

  • 4
    Your $7$s solution betters mine! – Weather Vane Sep 12 '19 at 18:39
  • 2
    Based on the rules established by this meta post and the consensus around it, an answer must have justification for why the solution is optimal. Without that, this is a comment, not an answer. – Deusovi Sep 13 '19 at 04:01
  • If... without concatenation is better than with, should we replace the without then? – athin Sep 13 '19 at 13:29
  • I mean, without concatenation is the subset of with right? Because without concatenation is more constrained thus with concatenation will be always better than without. (Hence, I suggest if without concatenation is better then we can replace with by without, cmiiw) – athin Sep 13 '19 at 13:34
  • @athin Is this edit what you mean? I selected the lowest number from each digit – ə̷̶̸͇̘̜́̍͗̂̄︣͟ Sep 13 '19 at 13:38
  • Ah, actually just if without concatenation is better than with, we can replace them. In the case of without concatenation is not better (for example, one or seven) we may leave them as one can still try to find the upperbound for both cases :) – athin Sep 13 '19 at 13:44
  • But anyway, we can leave it like this (like your edit). It's also fine as OP doesn't mention we need the version of not needing concatenation haha. – athin Sep 13 '19 at 13:46
  • 1
    You can use the factorial of the cardinality of an empty set |{}|! = 1, and from that generate any number using no digits! –  Sep 13 '19 at 16:27
  • If it is allowed to interprete the brackets ( ) as binomial coefficient, I can offer a solution using 8 fives. – ThomasL Sep 13 '19 at 16:55
  • 3 can be simplified to $3! * 3!^3 + 3!! + 3$ – Florian F Sep 14 '19 at 20:54
  • @FlorianF Still five digits... – ə̷̶̸͇̘̜́̍͗̂̄︣͟ Sep 14 '19 at 20:55
  • Yes. Just more elegant. Same for the 9's. – Florian F Sep 14 '19 at 20:57
11

My own (very trivial) solution with 0s (using 21 of them - so expecting to be beaten by somebody):

$$(0!+0!)^{0!+0!+0!+0!+0!+0!+0!+0!+0!+0!+0!}-(0!+0!+0!)^{0!+0!+0!}-0!-0!\\=2^{11}-3^3-1-1\\=2048-27-1-1\\=2019$$

trolley813
  • 11,282
  • 2
  • 25
  • 57
  • The long chain of 0! in the first power could be simplified to (0!+0!+0!)^(0!+0!)+0!+0! . What a monster this will end out being! – Adam Sep 12 '19 at 11:08
  • Please enlighten me and anyone else who doesn't have a degree in mathematics: how the hell do you make a string of 0s mean anything else than 0? I'd really like to understand this one. – Elmy Sep 12 '19 at 17:33
  • @Elmy by the magic of factorials! The definition of 0! is 0! = 1 – Adam Sep 12 '19 at 17:46
  • @Adam Thanks. But couldn't you shorten the first part to (0!+0!)^(0!0!) instead of adding 0! eleven times? The rules say that ii is a valid 2 digit number, equaling 11 in your case. – Elmy Sep 12 '19 at 20:04
  • @Elmy no we cannot. It works perfectly for when $i=1$ however this particular answer is assessing $i=0$. Annoyingly since $i=0$ it is impossible to take advantage of concatenation like that. Although if you look at this answer since $i=1$, they did exactly that! – Adam Sep 12 '19 at 20:09
  • 1
    Based on the rules established by this meta post and the consensus around it, an answer must have justification for why the solution is optimal. Without that, this is a comment, not an answer. – Deusovi Sep 13 '19 at 04:00
10

After an exhaustive search (by hand) I've found a really good solution for 0 using 14 digits!

$(0!+0!+0!)\times(((0!+0!+0!)!)!-((0!+0!+0!)!+0!))-((0!+0!+0!)!-0!)!$
$=3\times((3!)!-(3!+1))-(3!-1)!$
$=3\times(6!-(6+1))-(6-1)!$
$=3\times(720-7)-5!$
$=3\times(713)-120$
$=2139-120$
$=2019$

Proof for optimisation


To make this proof less convoluted (and to help in future similar problems) this is a dictionary of the construction of all numbers from 1-10. [Number ($i$ count): Expression]

1 (1): $0!$
2 (2): $1+1$
3 (3): $2+1$
4 (4): $3+1$ or $2^2$
5 (4): $3!-1$
6 (3): $3!$
7 (4): $6+1$
8 (5): $2^3$ or $6+2$
9 (5): $3^2$
10 (6): $9+1$ or $2\times5$

Proof that it isn't possible to reach 2019 using two (1-10) number $i$-expressions, so at least more than 8 digits are required

The only way to reach 2019 using two $i$-expressions is to use an expression either of the form $A \times B$, $A^B$ or $A-B$ ($\frac{A}{B}$ and $A+B$ are redundant). The only factors of 2019 are 3 and 673 and no factorial can equal 673 so $A \times B$ is ruled out and $A^B$ is ruled out by default. For $A-B$ to bear a solution, $A \ge 2019$ however no large number $B$ obtainable by a 1-10 number expression exists to constrain possible large $A$.

Following off of that proof, we now know the approach to take to get 2019

$A^B$ and $\frac{A}{B}$ are completely out of the question. From this it is safe to say that either $3\times673=2019$, $A+B$, $A-B$ or $A=2019^B$

On the subject of factorials...

No perfect power factorial exists above $1!$. Also no expression containing only factorials (above $1!$) can equal a prime number due to common factors.

This helps to drastically rule out many expressions

$A!=2019^B$ is out. $673$ is a prime number so the only way to reach it is by $A \pm B$. In the case of $673$, $A \pm B$ is disprovable by exhaustion for 1-10 numbers since nested factorials are disproved. In a more general case, $A=2019^B$ cannot be done if $C$ is added either. For it to even bear results it would have to be $A \pm B = 2019^C$ or $A \times B = 2019^C$. Since $2019^C=3^C673^C$ it clearly isn't possible using nested factorials so this can be disproved by trivial exhaustion. A common factor of $3^C$ can be brought out of $A!\pm B!$ for $A!,B! > 3!$ leaving two other factorials which must sum or difference to $673^C$ which isn't possible because it is many of the same prime. This is now trivially disprovable by exhaustion! Thus this case generally requires at least 12+ digits since more than three (1-10) number $i$-expressions are required. I doubt this will lead anywhere so it is fair to say that it has been disproved! {of course this isn't 100% irrefutable proof since it is possible for $D$ to enter the equation however the numbers (1-10) would be highly restrained and I don't want to write a full paper on this}

On the subject of $3\times673=2019$ we already know that $3$ takes 3 digits so all I have to do is show that $673$ cannot be made in a reasonable amount of digits. We know that $A,B$ expressions that contain only factorials cannot be prime and the other solutions are trivially disprovable by exhaustion ($A \pm B$). This brings it up to 11+ digits. Adding in $C$, the cases to be examined are $A \times B \pm C$, $A-B \times C$ and $A^B-C$. The unlisted cases are either clearly logically wrong, already listed and/or trivially disproved by exhaustion. $A \times B \pm C$ requires a nested factorial somewhere in order for the numbers to be large enough for a solution. If $A$ or $B$ is a factorial $>1!$ then $C$ must be an odd number for the result to be prime. Therefore either all numbers are nested factorials (disproved by previous logic) or only one of them are a nested factorial. Since only one of them can be a nested factorial, this is disprovable by exhaustion (gee this is getting exhausting!). This is extended to $A-B \times C$ (which is now very obvious without proof by "..."). For $A^B-C$, $A$ and $C$ cannot both be nested factorials so logically only $B,C$ should be examined. $A$ must then be odd. Looking at $A^{B!}-C!$, the cases where $A \le C$ - have a common factor thus disproved. $A>C$ restricts nested factorial $C$ to $A$'s upper bound thus is disprovable by "...". All of this proof comfortably disproves $3\times673=2019$ since $673$ requires more than three (1-10) numbers and at worse case scenario (unlike my $A=2019^B$ proof) is +12 digits but should generally require +15 digits.

Adam
  • 2,907
  • 1
  • 13
  • 36
8

The best solution I found so far (painful to say with concatenation...):
With 6 copies of 3

$( (3!)! - 3!3! ) \times 3 - 33$

$ = ( 6! - 6*6 ) \times 3 - 33$
$ = ( 720 - 36 ) \times 3 - 33$
$ = ( 684 ) \times 3 - 33$
$ = 2052 - 33$
$ = 2019$

P1storius
  • 2,341
  • 7
  • 24
  • 1
    Based on the rules established by this meta post and the consensus around it, an answer must have justification for why the solution is optimal. Without that, this is a comment, not an answer. – Deusovi Sep 13 '19 at 04:01
7

Update:

The first offering for (sixteen) $7$s based on the $1$s solution

$(\frac{7777 - 777 + 77 - 7}{7}) (\frac{7 + 7}{7}) - (\frac{7}{7}) = 1010 \times 2 - 1$

I have fourteen $1$s

$(1 + 1)(1111 - 111 + 11) - 1 - 1 - 1 = 2 \times 1011 - 3$

I have eight $4$s

$44(44 + \sqrt{4}) - 4 - \frac{4}{4} = 44 \times 46 - 5$

Weather Vane
  • 14,420
  • 1
  • 22
  • 54
6

With six nines:

$((\sqrt 9)!)!\times\sqrt 9 -((9+(\sqrt 9)!)\times9) -(\sqrt 9)!$
$=720\times3-(15\times9)-6$
$=2160-135-6$
$=2019$

Jan Ivan
  • 5,196
  • 14
  • 31
6

For $4$, we can use seven digits:

$\sqrt{\sqrt{\sqrt{4^{44}}}} - ((4-4)! + 4! + 4)$

athin
  • 34,177
  • 4
  • 70
  • 221
6

Ten 5s (different from @TheSimpliFire and OP)

$(5*5-5)*((5*5-5)*5+\frac{5}5)-\frac{5}5=2019$

Eight 6s (thanks @TheSimpliFire)

$6\sqrt{6^6}+6!+\frac{6+6+6}6=2019$

Ten 6s

$(666+6+\frac{6}{6})\times\frac{6+6+6}{6}=2019$

Ten 8s

$\frac{\sqrt{8^8}}{\sqrt{\sqrt{8+8}}} - \sqrt{(8+8)}! - \sqrt{8+8} - \frac{8}{8} = 2019 $
$\sqrt{8^8}=4096$
$\sqrt{\sqrt{8+8}}=2$
$\sqrt{(8+8)}!=24$

Omega Krypton
  • 21,640
  • 2
  • 25
  • 125
6

Seven 4's no concatenation

\begin{align}\frac{\sqrt{\sqrt{4^{4!}}}}{\sqrt{4}} - 4! - 4 - \frac{4}{4}&= \frac{4096}{2} - 4! - 4 - \frac{4}{4}\\&= 2048 - 24 - 4 - 1\\&= 2019\end{align}

SP812
  • 161
  • 4
  • 1
    Based on the rules established by this meta post and the consensus around it, an answer must have justification for why the solution is optimal. Without that, this is a comment, not an answer. – Deusovi Sep 13 '19 at 04:01
5

I have found a solution with 7 (threes)

$$3!^3*3!+(3*3)^3-3! = 6^3*6*(9)^3-6=1296 + 729 - 6 = 2019$$

53RT
  • 181
  • 2
  • 1
    Based on the rules established by this meta post and the consensus around it, an answer must have justification for why the solution is optimal. Without that, this is a comment, not an answer. – Deusovi Sep 13 '19 at 04:01
5

All solutions without concatenation.

A solution for $0$ using 13 digits

$(0! + 0! + 0!)!^{0! + 0! + 0! + 0!} + ((0! + 0! + 0!)!)! + 0! + 0! + 0! = 1296 + 720 + 3 = 2019$

A solution for $7$ using 10 digits

$\frac{(\frac{7! + 7 + 7}{7} - (7\times7)) \times (7+7+7)}{7} = \frac{673 * 21}{7} = 2019$

A solution for $8$ using 8 digits

$\dfrac{8!}{(\sqrt{8+8})!-\sqrt{8+8}}+\dfrac{(\sqrt{8+8})!}{8} = 2016 + 3 = 2019$

Old solutions:

A solution for $8$ using 9 digits:

$ \frac{8! + 8*8 - \sqrt{8+8}}{8+8+\sqrt{8+8}} = \frac{40320 + 64 - 4}{16 + 4} = 2019 $

Bananenkopp
  • 2,506
  • 9
  • 15
4

a "3-expression" instead of 5 in example:

$((3!^3 \times 3)+((3! \times (3+ \frac 33))+ \frac 33)) \times 3$
$((6^3 \times 3) + ((6 \times 4)+1)) \times 3$
$((216 \times 3) + (24+1)) \times 3$
$(648+25) \times 3$
$673\times3 = 2019$
10 threes

Shinjo
  • 412
  • 2
  • 10
  • 1
    Based on the rules established by this meta post and the consensus around it, an answer must have justification for why the solution is optimal. Without that, this is a comment, not an answer. – Deusovi Sep 13 '19 at 04:01
4

A solution for 7 using 18 7's

$7\times7\times7\times7 - \frac{777}{7} - \frac{777}{7} - \frac{777}{7} - 7\times7 = 2019$
$2401 - 111 -111 -111 -49$

Adam
  • 2,907
  • 1
  • 13
  • 36
Eye
  • 41
  • 2
4

12 1s:

$(1+1)^{11} - (1+1+1)^{1+1+1} - (1+1)$

Adam
  • 2,907
  • 1
  • 13
  • 36
1

Nine 3s without concatenation.

$3(3\times3)^3-3!\left(3^3+\dfrac{3}{3}\right)$

Adam
  • 2,907
  • 1
  • 13
  • 36
1

Nine ones with concatenation and decimal point:

$(1+1)*\frac{1111}{1.1}-1$

Eight fives with concatenation:

$5 - \frac{5555}{5} + 5^5$

Steffen
  • 13
  • 4