Fixed by not using string interpolation, seems the F# version here doesn't support it

Not a kata issue.

This doesn't seem like a kata issue, because the errors are coming from your code. I don't know F#, but reading the error messages leads me to believe that you're using a different version of F# locally, hence the errors.

Has F# been tested? I can solve the solution locally but when testing it I get this error:
/home/codewarrior/program.fsx(43,29): error FS1160: This token is reserved for future use

/home/codewarrior/program.fsx(65,76): error FS0035: This construct is deprecated: ':' is not permitted as a character in operator names and is reserved for future use

ok after looking at the other solutions I completely overthought this problem, spent a night on the red and found it particularly difficult for no reason - of course I should have just calculated the fib sequence in reverse. Sometimes a simple solution is missed

Not an issue, everything works correctly.

Please, NASM Translation

Nice simple explanation with unit testing. Good Kata, well done.

Approved

Fixed

Well I spent more time trying to understand the question than coding, hacking it together piece by piece as my understanding of the question evolved. I notice a lot of problems on this site are like that. Briefer, simpler to understand questions but with complicated solutions would be a far nicer thing to encounter.

What a terribly worded Kata. "An array of numbers" is incorrect, it's an array of strings. For something like computer programming where precision is everything this is absurd.

Why are slow solutions often the #1 voted just because they contain the fewest lines of code? In terms of efficiency for the end user I'd be interested in seeing a category that solves a problem with the minimal execution time. An example of this is in C# where nearly every solution uses LINQ, which is slower than non-LINQ based solutions despite it being shorter.

i feel like since the data is so small for these tests, that run time is not super relevant. Run time for small sets of input is not a good indicator, it is when we talk about "sufficently" large enough data sets.

Works for me